Support knowledge metadata filter (#15982)

api/core/rag/datasource/vdb/myscale/myscale_vector.py

@@ -133,6 +133,10 @@ class MyScaleVector(BaseVector):
             if self._metric.upper() == "COSINE" and order == SortOrder.ASC and score_threshold > 0.0
             else ""
         )
+        document_ids_filter = kwargs.get("document_ids_filter")
+        if document_ids_filter:
+            document_ids = ", ".join(f"'{id}'" for id in document_ids_filter)
+            where_str = f"{where_str} AND metadata['document_id'] in ({document_ids})"
         sql = f"""
             SELECT text, vector, metadata, {dist} as dist FROM {self._config.database}.{self._collection_name}
             {where_str} ORDER BY dist {order.value} LIMIT {top_k}