From 781b3f450e394b4f6096c8ac01987604a9814138 Mon Sep 17 00:00:00 2001
From: goulustis <cnimaiv@gmail.com>
Date: Fri, 30 Jan 2026 10:42:29 +0800
Subject: [PATCH] record conversations

---
 lang_agent/pipeline.py | 26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)

diff --git a/lang_agent/pipeline.py b/lang_agent/pipeline.py
index ca96d4a..28cca47 100644
--- a/lang_agent/pipeline.py
+++ b/lang_agent/pipeline.py
@@ -1,5 +1,5 @@
 from dataclasses import dataclass, field
-from typing import Type
+from typing import Type, List
 import tyro
 import asyncio
 import websockets
@@ -8,7 +8,7 @@ from loguru import logger
 import os
 
 from langchain.chat_models import init_chat_model
-from langchain_core.messages import SystemMessage, HumanMessage
+from langchain_core.messages import SystemMessage, HumanMessage, BaseMessage
 
 from langchain.agents import create_agent
 from langgraph.checkpoint.memory import MemorySaver
@@ -16,7 +16,7 @@ from langgraph.checkpoint.memory import MemorySaver
 from lang_agent.config import InstantiateConfig, KeyConfig
 from lang_agent.graphs import AnnotatedGraph, ReactGraphConfig, RoutingConfig
 from lang_agent.base import GraphBase
-
+from lang_agent.components.conv_store import CONV_STORE
 
 DEFAULT_PROMPT="""你是半盏新青年茶馆的服务员，擅长倾听、共情且主动回应。聊天时语气自然亲切，像朋友般轻松交流，不使用生硬术语。能接住各种话题，对疑问耐心解答，对情绪及时回应，避免冷场。保持积极正向，不传播负面信息，语言简洁易懂，让对话流畅舒适。与用户（User）交流时必须遵循[语气与格式]、[互动策略]、[安全与边界]、[输出要求]
 [角色设定]
@@ -108,9 +108,9 @@ class Pipeline:
     def invoke(self, *nargs, **kwargs)->str:
         out = self.graph.invoke(*nargs, **kwargs)
 
-        # If streaming, yield chunks from the generator
+        # If streaming, return the raw generator (let caller handle wrapping)
         if kwargs.get("as_stream"):
-            return self._stream_res(out)
+            return out
 
         # Non-streaming path
         if kwargs.get("as_raw"):
@@ -128,9 +128,13 @@ class Pipeline:
         assert 0, "something is wrong"
 
 
-    def _stream_res(self, out:list):
+    def _stream_res(self, out:List[str | List[BaseMessage]], conv_id:str=None):
         for chunk in out:
-            yield chunk
+            if isinstance(chunk, str):
+                yield chunk
+            else:
+                logger.info("logged message")
+                CONV_STORE.record_messages(conv_id, chunk)
 
     async def _astream_res(self, out):
         """Async version of _stream_res for async generators."""
@@ -142,8 +146,6 @@ class Pipeline:
         as_stream (bool): if true, enable the thing to be streamable
         as_raw (bool): return full dialoge of List[SystemMessage, HumanMessage, ToolMessage]
         """
-        # NOTE: this prompt will be overwritten by 'configs/route_sys_prompts/chat_prompt.txt' for route graph
-        u = DEFAULT_PROMPT
 
         device_id = "0"
         spl_ls = thread_id.split("_")
@@ -151,10 +153,6 @@ class Pipeline:
         if len(spl_ls) == 2:
             thread_id, device_id = spl_ls
 
-        # inp = {"messages":[SystemMessage(u),
-        #                         HumanMessage(inp)]}, {"configurable": {"thread_id": thread_id,
-        #                                                                "device_id":device_id}}
-
         inp = {"messages":[HumanMessage(inp)]}, {"configurable": {"thread_id": thread_id,
                                                                   "device_id":device_id}}
 
@@ -162,7 +160,7 @@ class Pipeline:
 
         if as_stream:
             # Yield chunks from the generator
-            return self._stream_res(out)
+            return self._stream_res(out, thread_id)
         else:
             return out