Update test suite

Author: jonavellecuerdo

tests/fixtures/cassettes/bulk_update_embeddings_success.yaml

@@ -0,0 +1,22 @@
+interactions:
+- request:
+    body: null
+    headers:
+      content-type:
+      - application/json
+      user-agent:
+      - opensearch-py/2.8.0 (Python 3.12.11)
+    method: GET
+    uri: http://localhost:9200/_cat/aliases?format=json
+  response:
+    body:
+      string: '[{"alias":"all-current","index":"libguides-2025-12-11t16-36-09","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},{"alias":"libguides","index":"libguides-2025-12-11t16-36-09","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},{"alias":"all-current","index":"test-index-2025-12-11t16-58-08","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},{"alias":"test-index","index":"test-index-2025-12-11t16-58-08","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"},{"alias":".kibana","index":".kibana_1","filter":"-","routing.index":"-","routing.search":"-","is_write_index":"-"}]'
+    headers:
+      content-length:
+      - '671'
+      content-type:
+      - application/json; charset=UTF-8
+    status:
+      code: 200
+      message: OK
+version: 1

tests/fixtures/cassettes/opensearch/bulk_update_raises_bulk_operation_error_if_record_not_found.yaml

@@ -0,0 +1,30 @@
+interactions:
+- request:
+    body: '{"update":{"_id":"i-am-not-found","_index":"test-index"}}
+
+      {"doc":{"timdex_record_id":"i-am-not-found","title":"Materials Science & Engineering
+      (UPDATED)"}}
+
+      '
+    headers:
+      Content-Length:
+      - '156'
+      content-type:
+      - application/json
+      user-agent:
+      - opensearch-py/2.8.0 (Python 3.12.11)
+    method: POST
+    uri: http://localhost:9200/_bulk
+  response:
+    body:
+      string: '{"took":9,"errors":true,"items":[{"update":{"_index":"test-index-2025-12-11t16-58-08","_id":"i-am-not-found","status":404,"error":{"type":"document_missing_exception","reason":"[i-am-not-found]:
+        document missing","index":"test-index-2025-12-11t16-58-08","shard":"0","index_uuid":"in04_JvQS5qqCvUXeZta_g"}}}]}'
+    headers:
+      content-length:
+      - '308'
+      content-type:
+      - application/json; charset=UTF-8
+    status:
+      code: 200
+      message: OK
+version: 1

tests/fixtures/cassettes/opensearch/bulk_update_updates_records.yaml

@@ -0,0 +1,29 @@
+interactions:
+- request:
+    body: '{"update":{"_id":"libguides:guides-175846","_index":"test-index"}}
+
+      {"doc":{"timdex_record_id":"libguides:guides-175846","title":"Materials Science
+      & Engineering (UPDATED)"}}
+
+      '
+    headers:
+      Content-Length:
+      - '174'
+      content-type:
+      - application/json
+      user-agent:
+      - opensearch-py/2.8.0 (Python 3.12.11)
+    method: POST
+    uri: http://localhost:9200/_bulk
+  response:
+    body:
+      string: '{"took":7,"errors":false,"items":[{"update":{"_index":"test-index-2025-12-11t16-58-08","_id":"libguides:guides-175846","_version":4,"result":"updated","_shards":{"total":2,"successful":1,"failed":0},"_seq_no":7,"_primary_term":1,"status":200}}]}'
+    headers:
+      content-length:
+      - '245'
+      content-type:
+      - application/json; charset=UTF-8
+    status:
+      code: 200
+      message: OK
+version: 1

tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=10/629d15f4-84e4-4b32-92c1-1b1debd377fb-0.parquet

@@ -5,7 +5,7 @@
 from freezegun import freeze_time
 
 from tim.cli import main
-from tim.errors import BulkIndexingError
+from tim.errors import BulkIndexingError, BulkOperationError
 
 from .conftest import EXIT_CODES, my_vcr
 
@@ -274,6 +274,93 @@ def test_bulk_update_with_source_raise_bulk_indexing_error(
     )
 
 
+@patch("tim.helpers.validate_bulk_cli_options")
+@patch("tim.opensearch.bulk_update")
+def test_bulk_update_embeddings_success(
+    mock_bulk_update, mock_validate_bulk_cli_options, caplog, monkeypatch, runner
+):
+    monkeypatch.delenv("TIMDEX_OPENSEARCH_ENDPOINT", raising=False)
+    mock_bulk_update.return_value = {"updated": 1, "errors": 0, "total": 1}
+    mock_validate_bulk_cli_options.return_value = "libguides"
+
+    result = runner.invoke(
+        main,
+        [
+            "bulk-update-embeddings",
+            "--source",
+            "libguides",
+            "--run-date",
+            "2025-05-07",
+            "--run-id",
+            "85cfe316-089c-4639-a5af-c861a7321493",
+            "tests/fixtures/dataset",
+        ],
+    )
+
+    assert result.exit_code == EXIT_CODES["success"]
+    assert (
+        f"Bulk update with embeddings complete: {json.dumps(mock_bulk_update())}"
+        in caplog.text
+    )
+
+
+@patch("tim.helpers.validate_bulk_cli_options")
+@patch("tim.opensearch.bulk_update")
+def test_bulk_update_embeddings_raise_error(
+    mock_bulk_update, mock_validate_bulk_cli_options, caplog, monkeypatch, runner
+):
+    monkeypatch.delenv("TIMDEX_OPENSEARCH_ENDPOINT", raising=False)
+    mock_bulk_update.return_value = {"updated": 0, "errors": 1, "total": 1}
+    mock_validate_bulk_cli_options.return_value = "libguides"
+
+    result = runner.invoke(
+        main,
+        [
+            "bulk-update-embeddings",
+            "--source",
+            "libguides",
+            "--run-date",
+            "2025-05-07",
+            "--run-id",
+            "85cfe316-089c-4639-a5af-c861a7321493",
+            "tests/fixtures/dataset",
+        ],
+    )
+    assert result.exit_code == EXIT_CODES["success"]
+    assert (
+        f"Bulk update with embeddings complete: {json.dumps(mock_bulk_update())}"
+        in caplog.text
+    )
+
+
+@patch("tim.helpers.validate_bulk_cli_options")
+@patch("tim.opensearch.bulk_update")
+def test_bulk_update_embeddings_raise_bulk_operation_error(
+    mock_bulk_update, mock_validate_bulk_cli_options, caplog, monkeypatch, runner
+):
+    monkeypatch.delenv("TIMDEX_OPENSEARCH_ENDPOINT", raising=False)
+    mock_bulk_update.side_effect = BulkOperationError(
+        action="update", record="alma:0", index="index", error="exception"
+    )
+    mock_validate_bulk_cli_options.return_value = "libguides"
+
+    result = runner.invoke(
+        main,
+        [
+            "bulk-update-embeddings",
+            "--source",
+            "libguides",
+            "--run-date",
+            "2025-05-07",
+            "--run-id",
+            "85cfe316-089c-4639-a5af-c861a7321493",
+            "tests/fixtures/dataset",
+        ],
+    )
+    assert result.exit_code == EXIT_CODES["success"]
+    assert "Bulk update with embeddings failed" in caplog.text
+
+
 @patch("tim.opensearch.create_index")
 @patch("tim.opensearch.promote_index")
 @patch("tim.opensearch.get_index_aliases")

tests/test_cli.py

@@ -51,6 +51,17 @@ def test_generate_bulk_actions_delete():
     }
 
 
+def test_generate_bulk_actions_update():
+    records = [{"timdex_record_id": "12345"}]
+    actions = helpers.generate_bulk_actions("test-index", records, "update")
+    assert next(actions) == {
+        "_op_type": "update",
+        "_index": "test-index",
+        "_id": "12345",
+        "doc": {"timdex_record_id": "12345"},
+    }
+
+
 def test_generate_bulk_actions_invalid_action_raises_error():
     records = [{"timdex_record_id": "12345", "other_fields": "some_data"}]
     actions = helpers.generate_bulk_actions("test-index", records, "wrong")

tests/test_helpers.py

@@ -9,6 +9,7 @@
 from tim.config import PRIMARY_ALIAS
 from tim.errors import (
     AliasNotFoundError,
+    BulkOperationError,
     IndexExistsError,
     IndexNotFoundError,
 )
@@ -532,3 +533,38 @@ def test_bulk_delete_logs_error_if_record_not_found(
         "Record to delete 'i-am-not-found' was not found in index 'test-index'."
         in caplog.text
     )
+
+
+# what happens when you try to 'bulk_update_embeddings' a document that doesn't exist*
+#   - do you need all these mocks or VCRs to test that (probably no :o)
+# fixture that reindexes a source into the mock
+# if we run tda in fixtures/datasets -- what's in it??
+@my_vcr.use_cassette("opensearch/bulk_update_updates_records.yaml")
+def test_bulk_update_updates_records(test_opensearch_client):
+    updates = [
+        {
+            "timdex_record_id": "libguides:guides-175846",
+            "title": "Materials Science & Engineering (UPDATED)",
+        }
+    ]
+    assert tim_os.bulk_update(test_opensearch_client, "test-index", iter(updates)) == {
+        "updated": 1,
+        "errors": 0,
+        "total": 1,
+    }
+
+
+@my_vcr.use_cassette(
+    "opensearch/bulk_update_raises_bulk_operation_error_if_record_not_found.yaml"
+)
+def test_bulk_update_raises_bulk_operation_error_if_record_not_found(
+    test_opensearch_client,
+):
+    updates = [
+        {
+            "timdex_record_id": "i-am-not-found",
+            "title": "Materials Science & Engineering (UPDATED)",
+        }
+    ]
+    with pytest.raises(BulkOperationError):
+        tim_os.bulk_update(test_opensearch_client, "test-index", iter(updates))

tests/test_opensearch.py

@@ -10,7 +10,7 @@
 from tim import errors, helpers
 from tim import opensearch as tim_os
 from tim.config import PRIMARY_ALIAS, VALID_SOURCES, configure_logger, configure_sentry
-from tim.errors import BulkIndexingError
+from tim.errors import BulkIndexingError, BulkOperationError
 
 logger = logging.getLogger(__name__)
 
@@ -390,7 +390,7 @@ def bulk_update_embeddings(
 
     try:
         update_results.update(tim_os.bulk_update(client, index, embeddings_to_index))
-    except BulkIndexingError as exception:
+    except BulkOperationError as exception:
         logger.info(f"Bulk update with embeddings failed: {exception}")
 
     logger.info(f"Bulk update with embeddings complete: {json.dumps(update_results)}")