Update _parse_proto_output function to include the entity_id as part …

…of the metadata returned.

Currently similarity_search_by_vector() returns slightly different information when looking at a BQ store versus a Feature Store. When executed against a BQ store the nearest neighbor ID is returned, but when executed against the Feature Store no ID is returned. To bring both into alignment we should include the entity_id for the Online Feature Store as well.

libs/community/langchain_google_community/bq_storage_vectorstores/featurestore.py

@@ -406,6 +406,7 @@ def _parse_proto_output(
             documents.append(
                 [
                     Document(
+                        id=result.entity_id,
                         page_content=content,
                         metadata=metadata,
                     ),