From 46a94d170708655bf24667fb392a08a0be532cbb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Marek=20Such=C3=A1nek?= <marek.suchanek@gmail.com>
Date: Thu, 15 Jan 2026 11:38:42 +0100
Subject: [PATCH] fix(docworker): Fix handling deleted document

---
 .../dsw/document_worker/exceptions.py         | 13 +++++++++-
 .../dsw/document_worker/worker.py             | 26 ++++++++++++++++++-
 2 files changed, 37 insertions(+), 2 deletions(-)

diff --git a/packages/dsw-document-worker/dsw/document_worker/exceptions.py b/packages/dsw-document-worker/dsw/document_worker/exceptions.py
index 3b7b56b2..1b5095c9 100644
--- a/packages/dsw-document-worker/dsw/document_worker/exceptions.py
+++ b/packages/dsw-document-worker/dsw/document_worker/exceptions.py
@@ -23,7 +23,18 @@ def db_message(self):
                f'{str(self.exc)}'
 
 
-def create_job_exception(job_id: str, message: str, exc=None):
+class DocumentNotFoundException(JobException):
+    pass
+
+
+def create_job_exception(job_id: str, message: str, document_found=True, exc=None):
+    if not document_found:
+        return DocumentNotFoundException(
+            job_id=job_id,
+            msg=message,
+            exc=exc,
+        )
+
     if isinstance(exc, JobException):
         return exc
 
diff --git a/packages/dsw-document-worker/dsw/document_worker/worker.py b/packages/dsw-document-worker/dsw/document_worker/worker.py
index 95adb253..efa8cc8e 100644
--- a/packages/dsw-document-worker/dsw/document_worker/worker.py
+++ b/packages/dsw-document-worker/dsw/document_worker/worker.py
@@ -19,7 +19,8 @@
     CMD_COMPONENT, CMD_CHANNEL, PROG_NAME
 from .context import Context
 from .documents import DocumentFile, DocumentNameGiver
-from .exceptions import create_job_exception, JobException
+from .exceptions import create_job_exception, JobException, \
+    DocumentNotFoundException
 from .limits import LimitsEnforcer
 from .templates import TemplateRegistry, Template, Format
 from .utils import byte_size_format, check_metamodel_version
@@ -84,6 +85,21 @@ def safe_format(self) -> Format:
             raise RuntimeError('Format is not set but it should')
         return self.format
 
+    def check_document_exists(self) -> bool:
+        LOG.debug('Checking if document "%s" exists in DB', self.doc_uuid)
+        try:
+            doc = self.ctx.app.db.fetch_document(
+                document_uuid=self.doc_uuid,
+                tenant_uuid=self.tenant_uuid,
+            )
+            exists = doc is not None
+            LOG.debug('Document "%s" exists: %s', self.doc_uuid, exists)
+            return exists
+        except Exception as e:
+            LOG.error('Failed to check if document "%s" exists: %s',
+                      self.doc_uuid, str(e))
+            return False
+
     @handle_job_step('Failed to get document from DB')
     def get_document(self):
         SentryReporter.set_tags(phase='fetch')
@@ -102,6 +118,7 @@ def get_document(self):
             raise create_job_exception(
                 job_id=self.doc_uuid,
                 message='Document record not found in database',
+                document_found=False,
             )
         self.doc.retrieved_at = datetime.datetime.now(tz=datetime.UTC)
         LOG.info('Job "%s" details received', self.doc_uuid)
@@ -312,6 +329,11 @@ def _run(self):
         self.finalize()
 
     def _set_failed(self, message: str):
+        document_exists = self.check_document_exists()
+        if not document_exists:
+            LOG.warning('Document %s does not exist, cannot set to FAILED',
+                        self.doc_uuid)
+            return
         if self.try_set_job_state(DocumentState.FAILED, message):
             LOG.info('Set state to FAILED')
         else:
@@ -323,6 +345,8 @@ def _set_failed(self, message: str):
     def run(self):
         try:
             self._run()
+        except DocumentNotFoundException as e:
+            LOG.warning('Document not found: %s', e.log_message())
         except JobException as e:
             LOG.warning('Handled job error: %s', e.log_message())
             SentryReporter.capture_exception(e)