[NOTICKET] minor fix in chat.py, add package for query, change schema used to hybrid (cosine+bm25)

pyproject.toml

@@ -79,6 +79,8 @@ dependencies = [
     "jsonpatch>=1.33",
     "pymongo>=4.14.0",
     "psycopg2>=2.9.11",
+    # --- SQL parsing / guardrails ---
+    "sqlglot>=25.0.0",
     # --- User-DB connectors (db_pipeline) ---
     "pymysql>=1.1.1",
     "pymssql>=2.3.0",

src/api/v1/chat.py

@@ -61,7 +61,7 @@ def _extract_sources(results: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
     seen = set()
     sources = []
     for result in results:
-        if "document_id" in result["metadata"]["data"]:
+        if "document_id" in result["metadata"].get("data", {}):
             meta = result["metadata"]
             key = (meta.get("data", {}).get("document_id"), meta.get("data", {}).get("page_label"))
             if key not in seen:
@@ -182,12 +182,20 @@ async def chat_stream(request: ChatRequest, db: AsyncSession = Depends(get_db)):
 
             if not intent_result.get("needs_search"):
                 retrieval_task.cancel()
+                try:
+                    await retrieval_task
+                except asyncio.CancelledError:
+                    pass
                 raw_results = []
             else:
                 search_query = intent_result.get("search_query", request.message)
                 logger.info(f"Searching for: {search_query}")
                 if search_query != request.message:
                     retrieval_task.cancel()
+                    try:
+                        await retrieval_task
+                    except asyncio.CancelledError:
+                        pass
                     raw_results = await retriever.retrieve(
                         query=search_query,
                         user_id=request.user_id,

src/rag/retrievers/schema.py

@@ -31,7 +31,7 @@ logger = get_logger("schema_retriever")
 _TABULAR_FILE_TYPES = ("csv", "xlsx")
 
 Strategy = Literal["dense_no_threshold", "dense_dot", "dense_l2", "hybrid", "hybrid_bm25"]
-ACTIVE_STRATEGY: Strategy = "dense_no_threshold"
+ACTIVE_STRATEGY: Strategy = "hybrid_bm25"
 
 
 class SchemaRetriever(BaseRetriever):