构建上下文感知对话机器人,传统聊天机器人常因缺乏上下文记忆导致对话割裂 本教程将使用LangChain框架，打造能记住对话历史的...

from langchain_openai import ChatOpenAI from langchain_core.runnables.history import RunnableWithMessageHistory from langchain_community.chat_message_histories import ChatMessageHistory from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder import os from dotenv import load_dotenv os.environ["OPENAI_API_KEY"] = '' # 创建内存存储字典 session_store = {} def get_chat_history(session_id: str): """智能会话存储器""" if session_id not in session_store: session_store[session_id] = ChatMessageHistory() return session_store[session_id] # 加载环境变量 load_dotenv() # 创建内存存储字典 session_store = {} def get_chat_history(session_id: str): """智能会话存储器""" if session_id not in session_store: session_store[session_id] = ChatMessageHistory() return session_store[session_id] # 构建三级对话模板 conversation_blueprint = ChatPromptTemplate.from_messages([ ("system", "你是一个专业的人工智能助手"), # 角色定义层 MessagesPlaceholder(variable_name="history"), # 历史记忆层 ("human", "{input}") # 用户输入层 ]) # 初始化语言引擎 #llm_engine = ChatOpenAI(model="gpt-4o-mini", max_tokens=1000, temperature=0) # 修改base_url并调整模型名称 llm_engine = ChatOpenAI( base_url="https://api.deepseek.com/v1", # DeepSeek API端点 model="deepseek-chat", # DeepSeek模型标识 openai_api_key="sk-xxxxxxxxxxxxxxxxxx", # 替换为DeepSeek密钥 max_tokens=1000, temperature=0 ) # 构建处理链 processing_pipeline = conversation_blueprint | llm_engine # 添加历史管理模块 smart_agent = RunnableWithMessageHistory( processing_pipeline, get_chat_history, input_messages_key="input", history_messages_key="history" ) # 实现对话 user_session = "user_007" # 首次对话 first_response = smart_agent.invoke( {"input": "你好！请问是先有鸡还是先有蛋？"}, config={"configurable": {"session_id": user_session}} ) print(f"AI回复: {first_response.content}") # 延续对话 followup_response = smart_agent.invoke( {"input": "我刚才问的第一个问题是什么？"}, config={"configurable": {"session_id": user_session}} ) print(f"AI回复: {followup_response.content}") print("\n完整对话记录:") for message in session_store[user_session].messages: print(f"[{message.type.upper()}] {message.content}") # 添加情感分析模块示例 from textblob import TextBlob def analyze_sentiment(text): analysis = TextBlob(text) return analysis.sentiment.polarity