From a8879057c0bcf1f12e3fecfd1515c1a715f440cd Mon Sep 17 00:00:00 2001
From: Jyong <76649700+JohnJyong@users.noreply.github.com>
Date: Wed, 19 Mar 2025 19:44:56 +0800
Subject: [PATCH] fix tidb metadata filter (#16237)

---
 .../tidb_on_qdrant/tidb_on_qdrant_vector.py    | 18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)

diff --git a/api/core/rag/datasource/vdb/tidb_on_qdrant/tidb_on_qdrant_vector.py b/api/core/rag/datasource/vdb/tidb_on_qdrant/tidb_on_qdrant_vector.py
index 2b0502833..ae4baeb17 100644
--- a/api/core/rag/datasource/vdb/tidb_on_qdrant/tidb_on_qdrant_vector.py
+++ b/api/core/rag/datasource/vdb/tidb_on_qdrant/tidb_on_qdrant_vector.py
@@ -328,16 +328,13 @@ class TidbOnQdrantVector(BaseVector):
         )
         document_ids_filter = kwargs.get("document_ids_filter")
         if document_ids_filter:
-            should_conditions = []
-            for document_id_filter in document_ids_filter:
-                should_conditions.append(
+            if filter.must:
+                filter.must.append(
                     models.FieldCondition(
                         key="metadata.document_id",
-                        match=models.MatchValue(value=document_id_filter),
+                        match=models.MatchAny(any=document_ids_filter),
                     )
                 )
-            if should_conditions:
-                filter.should = should_conditions  # type: ignore
         results = self._client.search(
             collection_name=self._collection_name,
             query_vector=query_vector,
@@ -382,16 +379,13 @@ class TidbOnQdrantVector(BaseVector):
         )
         document_ids_filter = kwargs.get("document_ids_filter")
         if document_ids_filter:
-            should_conditions = []
-            for document_id_filter in document_ids_filter:
-                should_conditions.append(
+            if scroll_filter.must:
+                scroll_filter.must.append(
                     models.FieldCondition(
                         key="metadata.document_id",
-                        match=models.MatchValue(value=document_id_filter),
+                        match=models.MatchAny(any=document_ids_filter),
                     )
                 )
-            if should_conditions:
-                scroll_filter.should = should_conditions  # type: ignore
         response = self._client.scroll(
             collection_name=self._collection_name,
             scroll_filter=scroll_filter,