Proper node deletion

src/paperless_ai/indexing.py

@@ -31,7 +31,7 @@ RAG_CHUNK_OVERLAP = 200
 
 
 def _index_lock_path() -> Path:
-    """Return the path used as the file lock for FAISS index mutations.
+    """Return the path used as the file lock for LLM index mutations.
 
     The lock file lives in DATA_DIR/locks/ (not inside LLM_INDEX_DIR) so that a
     rebuild — which calls shutil.rmtree(LLM_INDEX_DIR) — cannot delete the lock
@@ -184,8 +184,7 @@ def load_or_build_index(nodes=None):
 
 def remove_document_docstore_nodes(document: Document, index: "VectorStoreIndex"):
     """
-    Removes existing documents from docstore for a given document from the index.
-    This is necessary because FAISS IndexFlatL2 is append-only.
+    Removes existing nodes for a given document from the vector index and docstore.
     """
     all_node_ids = list(index.docstore.docs.keys())
     existing_nodes = [
@@ -193,16 +192,10 @@ def remove_document_docstore_nodes(document: Document, index: "VectorStoreIndex"
         for node in index.docstore.get_nodes(all_node_ids)
         if node.metadata.get("document_id") == str(document.id)
     ]
+    if existing_nodes:
+        index.vector_store.delete_nodes(existing_nodes)
     for node_id in existing_nodes:
-        # Delete from docstore, FAISS IndexFlatL2 are append-only
         index.docstore.delete_document(node_id)
-        # Also purge the FAISS position -> UUID mapping so subsequent similarity
-        # queries don't raise KeyError on ghost vector positions.
-        stale_keys = [
-            k for k, v in index.index_struct.nodes_dict.items() if v == node_id
-        ]
-        for key in stale_keys:
-            del index.index_struct.nodes_dict[key]
     # Re-sync the mutated index_struct so persist() writes the updated nodes_dict.
     index.storage_context.index_store.add_index_struct(index.index_struct)
 
@@ -457,12 +450,18 @@ def query_similar_documents(
         )
         try:
             results = retriever.retrieve(query_text)
+        except Warning as e:
+            logger.debug(
+                "Skipping LLM similarity query for document %s because the "
+                "vector store returned no results: %s",
+                document.pk,
+                e,
+            )
+            return []
         except KeyError as e:
-            # Ghost FAISS positions remain after deletion because IndexFlatL2 is
-            # append-only. Treat them as absent and return no results.
             logger.debug(
                 "Skipping LLM similarity query for document %s due to a stale "
-                "FAISS position with no docstore node: %s",
+                "vector mapping with no docstore node: %s",
                 document.pk,
                 e,
             )

src/paperless_ai/tests/test_ai_indexing.py

@@ -330,6 +330,7 @@ def test_remove_document_deletes_node_from_docstore(
     indexing.llm_index_remove_document(real_document)
     index = indexing.load_or_build_index()
     assert len(index.docstore.docs) == 0
+    assert index.vector_store.table.count_rows() == 0
 
 
 @pytest.mark.django_db