update

py/core/main/services/ingestion_service.py

@@ -468,9 +468,11 @@ class IngestionService:
             )
             if user.limits_overrides and "max_chunks" in user.limits_overrides:
                 max_chunks = user.limits_overrides["max_chunks"]
-
+            
             # Add to our local batch
             vector_batch.append(msg)
+            print(f"vector_batch: {msg}")
+            print(f"vector_batch: {vector_batch}")
             document_counts[msg.document_id] = (
                 document_counts.get(msg.document_id, 0) + 1
             )

py/core/providers/database/collections.py

@@ -222,8 +222,8 @@ class PostgresCollectionsHandler(Handler):
                 COUNT(DISTINCT u.id) FILTER (WHERE u.id IS NOT NULL) as user_count,
                 COUNT(DISTINCT d.id) FILTER (WHERE d.id IS NOT NULL) as document_count
             FROM updated_collection uc
-            LEFT JOIN {self._get_table_name("users")} u ON uc.id = ANY(u.collection_ids)
-            LEFT JOIN {self._get_table_name("documents")} d ON uc.id = ANY(d.collection_ids)
+            LEFT JOIN {self._get_table_name("users")} u ON u.collection_ids @> ARRAY[uc.id]::uuid[]
+            LEFT JOIN {self._get_table_name("documents")} d ON d.collection_ids @> ARRAY[uc.id]::uuid[]
             GROUP BY uc.id, uc.owner_id, uc.name, uc.description, uc.graph_sync_status, uc.graph_cluster_status, uc.created_at, uc.updated_at
         """
         try:
@@ -314,7 +314,7 @@ class PostgresCollectionsHandler(Handler):
                 d.size_in_bytes, d.ingestion_status, d.extraction_status, d.created_at, d.updated_at, d.summary,
                 COUNT(*) OVER() AS total_entries
             FROM {self._get_table_name("documents")} d
-            WHERE $1 = ANY(d.collection_ids)
+            WHERE d.collection_ids @> ARRAY[$1]::uuid[]
             ORDER BY d.created_at DESC
             OFFSET $2
         """

py/migrations/versions/c45a9cf6a8a4_add_user_and_document_count_to_.py

@@ -72,8 +72,8 @@ def upgrade():
                    COUNT(DISTINCT u.id) as user_count,
                    COUNT(DISTINCT d.id) as document_count
             FROM {project_name}.collections c
-            LEFT JOIN {project_name}.users u ON c.id = ANY(u.collection_ids)
-            LEFT JOIN {project_name}.documents d ON c.id = ANY(d.collection_ids)
+            LEFT JOIN {project_name}.users u ON ANY(u.collection_ids) @> ARRAY[c.id]::uuid[]
+            LEFT JOIN {project_name}.documents d ON ANY(d.collection_ids) @> ARRAY[c.id]::uuid[]
             GROUP BY c.id
         )
         UPDATE {project_name}.collections c