print only assitant

backend.py

@@ -165,13 +165,15 @@ def handle_query(query_str: str,
         ),
     )
         
+        retriever = index.as_retriever(similarity_top_k=3)
+
         # Let's test it out
-        """relevant_chunks = chat_engine.retrieve("mercato")
+        relevant_chunks = relevant_chunks = retriever.retrieve(query_str)
         print(f"Found: {len(relevant_chunks)} relevant chunks")
         for idx, chunk in enumerate(relevant_chunks):
             info_message += f"{idx + 1}) {chunk.text[:64]}...\n"
             print(info_message)
-            gr.Info(info_message)"""
+            gr.Info(info_message)
             
         
         #prompts_dict = chat_engine.get_prompts()
@@ -187,13 +189,10 @@ def handle_query(query_str: str,
         #response = chat_engine.chat(query_str)
 
         for token in response.response_gen:
-            #if not token.startswith("system:") and not token.startswith("user:"):
-                #if token.startswith("file_path:"):
-                    #sources.append(token)
-                
-                outputs.append(token)
+            if token.startswith("assistant:"):
+                # Remove the "assistant:" prefix
+                outputs.append(token[len("assistant:"):])
                 print(f"Generated token: {token}")
-                
                 yield "".join(outputs)
                 #yield CompletionResponse(text=''.join(outputs), delta=token)