feat: implement rebuild indexes functionality for documents

[iziang]

No description provided.

[cursor]

🚨 BugBot couldn't run

Something went wrong. Try again by commenting "bugbot run", or contact support (requestId: serverGenReqId_46150768-d3f6-4e9d-9f82-d06d2cdbc205).

[iziang]

bugbot run

[cursor]

Bug: API Contract Violation: Incorrect Response Code

The rebuild_indexes endpoint's OpenAPI specification defines a 204 No Content response, but the document_service.py implementation returns a 200 OK with a JSON body {"code": "200", "message": "..."}. This mismatch between the API contract and implementation can cause client issues.

aperag/api/paths/collections.yaml#L321-L324

ApeRAG/aperag/api/paths/collections.yaml

Lines 321 to 324 in e380ab0

	$ref: '../components/schemas/document.yaml#/rebuildIndexesRequest'
	responses:
	'204':
	description: Index rebuild initiated successfully

aperag/service/document_service.py#L457-L461

ApeRAG/aperag/service/document_service.py

Lines 457 to 461 in e380ab0

	return {
	"code": "200",
	"message": f"Index rebuild initiated for types: {', '.join(index_types)}"
	}

Fix in Cursor

---

Bug: Verification Queries Execute Outside Transaction

The rebuild_document_indexes method's transaction function (_rebuild_document_indexes_atomically) performs verification queries (self.db_ops.query_document, self.db_ops.query_collection) using a global db_ops instance instead of the transaction's session parameter. This causes these queries to execute outside the transaction, breaking isolation and potentially leading to race conditions or inconsistent reads. These verification queries should either be moved outside the transaction or rewritten to use the provided session.

aperag/service/document_service.py#L440-L451

ApeRAG/aperag/service/document_service.py

Lines 440 to 451 in e380ab0

	# Verify document exists and user has access
	document = await self.db_ops.query_document(user_id, collection_id, document_id)
	if not document:
	raise DocumentNotFoundException(f"Document {document_id} not found")
	if document.collection_id != collection_id:
	raise ResourceNotFoundException(f"Document {document_id} not found in collection {collection_id}")
	# Verify user has access to the collection
	collection = await self.db_ops.query_collection(user_id, collection_id)
	if not collection or collection.user != user_id:
	raise ResourceNotFoundException(f"Collection {collection_id} not found or access denied")

Fix in Cursor

---

Was this report helpful? Give feedback by reacting with 👍 or 👎