Avoid unnecessary LLM response generation

lightrag/operate.py

@@ -1025,6 +1025,10 @@ async def _build_query_context(
             [hl_relations_context, ll_relations_context],
             [hl_text_units_context, ll_text_units_context],
         )
+    # not necessary to use LLM to generate a response
+    if not entities_context.strip() and not relations_context.strip():
+        return None
+        
     return f"""
 -----Entities-----
 ```csv