clean up args

backend/onyx/background/celery/tasks/shared/RetryDocumentIndex.py

@@ -31,6 +31,7 @@ def __init__(self, index: DocumentIndex):
     def delete_single(
         self,
         doc_id: str,
+        *,
         tenant_id: str | None,
         chunk_count: int | None,
     ) -> int:
@@ -48,11 +49,11 @@ def delete_single(
     def update_single(
         self,
         doc_id: str,
+        *,
         tenant_id: str | None,
         chunk_count: int | None,
         fields: VespaDocumentFields,
     ) -> int:
-        print("Update single")
         return self.index.update_single(
             doc_id,
             tenant_id=tenant_id,

backend/onyx/document_index/interfaces.py

@@ -136,7 +136,7 @@ def __init__(
         index_name: str,
         secondary_index_name: str | None,
         *args: Any,
-        **kwargs: Any
+        **kwargs: Any,
     ) -> None:
         super().__init__(*args, **kwargs)
         self.index_name = index_name
@@ -221,6 +221,7 @@ class Deletable(abc.ABC):
     def delete_single(
         self,
         doc_id: str,
+        *,
         tenant_id: str | None,
         chunk_count: int | None,
     ) -> int:
@@ -247,8 +248,9 @@ class Updatable(abc.ABC):
     def update_single(
         self,
         doc_id: str,
-        chunk_count: int | None,
+        *,
         tenant_id: str | None,
+        chunk_count: int | None,
         fields: VespaDocumentFields,
     ) -> int:
         """
@@ -269,7 +271,7 @@ def update_single(
 
     @abc.abstractmethod
     def update(
-        self, update_requests: list[UpdateRequest], tenant_id: str | None
+        self, update_requests: list[UpdateRequest], *, tenant_id: str | None
     ) -> None:
         """
         Updates some set of chunks. The document and fields to update are specified in the update

backend/onyx/document_index/vespa/index.py

@@ -433,7 +433,7 @@ def _update_chunk(
                         raise requests.HTTPError(failure_msg) from e
 
     def update(
-        self, update_requests: list[UpdateRequest], tenant_id: str | None
+        self, update_requests: list[UpdateRequest], *, tenant_id: str | None
     ) -> None:
         logger.debug(f"Updating {len(update_requests)} documents in Vespa")
 
@@ -561,6 +561,7 @@ def update_single_chunk(
     def update_single(
         self,
         doc_id: str,
+        *,
         chunk_count: int | None,
         tenant_id: str | None,
         fields: VespaDocumentFields,
@@ -610,6 +611,7 @@ def update_single(
     def delete_single(
         self,
         doc_id: str,
+        *,
         tenant_id: str | None,
         chunk_count: int | None,
     ) -> int:
@@ -795,7 +797,12 @@ def enrich_basic_chunk_info(
         return enriched_doc_info
 
     @classmethod
-    def delete_entries_by_tenant_id(cls, tenant_id: str, index_name: str) -> None:
+    def delete_entries_by_tenant_id(
+        cls,
+        *,
+        tenant_id: str,
+        index_name: str,
+    ) -> None:
         """
         Deletes all entries in the specified index with the given tenant_id.
 

backend/scripts/force_delete_connector_by_id.py

@@ -5,6 +5,7 @@
 from sqlalchemy import delete
 from sqlalchemy.orm import Session
 
+from onyx.db.document import delete_documents_complete__no_commit
 from onyx.db.enums import ConnectorCredentialPairStatus
 
 # Modify sys.path
@@ -70,14 +71,17 @@ def _unsafe_deletion(
         if not documents:
             break
 
-        # document_ids = [document.id for document in documents]
-        # for doc_id in document_ids:
-        #     document_index.delete_single(doc_id)
+        for document in documents:
+            document_index.delete_single(
+                doc_id=document.id,
+                tenant_id=None,
+                chunk_count=document.chunk_count,
+            )
 
-        # delete_documents_complete__no_commit(
-        #     db_session=db_session,
-        #     document_ids=document_ids,
-        # )
+        delete_documents_complete__no_commit(
+            db_session=db_session,
+            document_ids=[document.id for document in documents],
+        )
 
         num_docs_deleted += len(documents)
 
@@ -215,6 +219,7 @@ def _delete_connector(cc_pair_id: int, db_session: Session) -> None:
     parser.add_argument(
         "connector_id", type=int, help="The ID of the connector to delete"
     )
+
     args = parser.parse_args()
     with get_session_context_manager() as db_session:
         _delete_connector(args.connector_id, db_session)

backend/scripts/orphan_doc_cleanup_script.py

@@ -15,6 +15,7 @@
 from onyx.db.document import delete_documents_complete__no_commit  # noqa: E402
 from onyx.db.search_settings import get_current_search_settings  # noqa: E402
 from onyx.document_index.vespa.index import VespaIndex  # noqa: E402
+from onyx.db.document import get_document  # noqa: E402
 
 BATCH_SIZE = 100
 
@@ -63,6 +64,10 @@ def main() -> None:
             with concurrent.futures.ThreadPoolExecutor(max_workers=100) as executor:
 
                 def process_doc(doc_id: str) -> str | None:
+                    document = get_document(doc_id, db_session)
+                    if not document:
+                        print(f"Document {doc_id} not found in Postgres")
+                        return None
                     # Check if document exists in Vespa first
                     try:
                         chunks = vespa_index.id_based_retrieval(
@@ -83,11 +88,13 @@ def process_doc(doc_id: str) -> str | None:
 
                     try:
                         print(f"Deleting document {doc_id} in Vespa")
-                        # chunks_deleted = vespa_index.delete_single(doc_id)
-                        # if chunks_deleted > 0:
-                        #     print(
-                        #         f"Deleted {chunks_deleted} chunks for document {doc_id}"
-                        #     )
+                        chunks_deleted = vespa_index.delete_single(
+                            doc_id, tenant_id=None, chunk_count=document.chunk_count
+                        )
+                        if chunks_deleted > 0:
+                            print(
+                                f"Deleted {chunks_deleted} chunks for document {doc_id}"
+                            )
                         return doc_id
                     except Exception as e:
                         print(