from typing import Annotated, TypedDict from langchain_core.messages import AIMessage, BaseMessage from langgraph.graph import StateGraph, END from langgraph.graph.message import add_messages from app.core.database import ChatMessageModel import logging logger = logging.getLogger(__name__) class ChatState(TypedDict): messages: Annotated[list[BaseMessage], add_messages] # Define a simple graph def agent_node(state: ChatState): # Mock LLM logic for now last_message = state["messages"][-1] response = AIMessage(content=f"Echo: {last_message.content}") return {"messages": [response]} workflow = StateGraph(ChatState) workflow.add_node("agent", agent_node) workflow.set_entry_point("agent") workflow.add_edge("agent", END) async def save_chat_history(session, session_id, human_msg, ai_msg): # session 是 SQLAlchemy 的 AsyncSession try: user_msg = ChatMessageModel( session_id=session_id, role="user", content=human_msg ) ai_msg = ChatMessageModel(session_id=session_id, role="ai", content=ai_msg) session.add(user_msg) session.add(ai_msg) await session.commit() # 提交事务 except Exception as e: await session.rollback() raise e