Merge pull request #585 from openchatai/feat/follow_up

move the followup question to another thread

llm-server/routes/chat/chat_controller.py

@@ -232,9 +232,7 @@ async def handle_chat_send_common(
         if result.error:
             logger.error("chat_conversation_error", message=result.error)
 
-        emit(session_id, "|im_end|") if is_streaming else jsonify(
-            {"type": "text", "response": {"text": result.message}}
-        )
+        return jsonify({"type": "text", "response": {"text": result.message}})
     except Exception as e:
         logger.error(
             "An exception occurred",
@@ -273,6 +271,7 @@ def m_called_actions_by_bot(bot_id: str):
     response = most_called_actions_by_bot(bot_id)
     return jsonify(response)
 
+
 @chat_workflow.route("/transcribe", methods=["POST"])
 async def transcribe_audio():
     # Check if the post request has the file part

llm-server/routes/chat/implementation/chain_strategy.py

@@ -22,6 +22,7 @@
 )
 from utils.llm_consts import VectorCollections
 from models.repository.action_call_repo import add_action_call
+from utils.llm_consts import enable_followup_questions
 
 
 class ChainStrategy(ChatRequestHandler):
@@ -114,15 +115,15 @@ async def handle_request(
                 session_id=session_id,
             )
 
+            emit(session_id, "|im_end|") if is_streaming else None
+
             # we only support follow_up question this in streaming mode
-            if is_streaming:
+
+            if enable_followup_questions:
                 followup_question_list = await generate_follow_up_questions(
-                    conversations_history, response.message or "", current_input=text
+                    conversations_history, response.message or "", text
                 )
-
-                print(followup_question_list.json())
                 if is_streaming:
                     emit(f"{session_id}_follow_qns", followup_question_list.json())
-
             response.knowledgebase_called = True
             return response

llm-server/utils/llm_consts.py

@@ -78,3 +78,7 @@ def get_mysql_uri():
 
 chat_strategy = os.getenv("CHAT_STRATEGY", ChatStrategy.chain)
 max_pages_to_crawl = int(os.getenv("MAX_PAGES_TO_CRAWL", "15"))
+
+enable_followup_questions = (
+    True if os.getenv("ENABLE_FOLLOWUP_QUESTIONS", "YES") == "YES" else False
+)