基于LangChain和DeepSeek的Agent开发指南

# 创建虚拟环境
python -m venv agent-env
source agent-env/bin/activate  # Linux/Mac
agent-env\Scripts\activate     # Windows

# 安装核心依赖
pip install langchain langchain-core langchain-community
pip install deepseek-api python-dotenv

from langchain_community.chat_models import ChatDeepSeek
from langchain_core.prompts import ChatPromptTemplate
from dotenv import load_dotenv

load_dotenv()

# 初始化模型
model = ChatDeepSeek(
    model_name="deepseek-chat",
    temperature=0.5,
    max_tokens=1024
)

# 创建提示模板
prompt = ChatPromptTemplate.from_messages([
    ("system", "你是一个乐于助人的AI助手"),
    ("human", "{input}")
])

# 构建基础链
chain = prompt | model

response = chain.invoke({
    "input": "请用Python写一个快速排序算法"
})
print(response.content)

from langchain.memory import ConversationBufferMemory

memory = ConversationBufferMemory(
    memory_key="history",
    return_messages=True
)

# 更新提示模板
prompt = ChatPromptTemplate.from_messages([
    ("system", """你是一个专业的技术顾问，具有以下特点：
    1. 回答简洁明确
    2. 会主动追问细节
    3. 使用Markdown格式"""),
    MessagesPlaceholder(variable_name="history"),
    ("human", "{input}")
])

# 构建带记忆的链
conversation_chain = (
    RunnablePassthrough.assign(
        history=RunnableLambda(memory.load_memory_variables)
    ) 
    | prompt
    | model
)

while True:
    user_input = input("你：")
    if user_input.lower() == 'exit':
        break
        
    response = conversation_chain.invoke(
        {"input": user_input},
        config={"callbacks": [memory]}
    )
    
    print(f"AI：{response.content}")

from langchain.agents import tool

@tool
def get_weather(city: str):
    """获取指定城市的实时天气"""
    # 这里可以接入真实天气API
    return f"{city}当前天气：25℃，晴"

# 创建工具包
tools = [get_weather]

# 构建Agent
agent = create_tool_calling_agent(
    model,
    tools,
    prompt
)

advanced_prompt = ChatPromptTemplate.from_messages([
    ("system", """你是一个全能助手，请遵守以下规则：
    1. 当涉及天气查询时，必须使用工具
    2. 技术问题回答需包含示例代码
    3. 保持回答在3句话以内"""),
    MessagesPlaceholder("history"),
    ("human", "{input}"),
    MessagesPlaceholder("agent_scratchpad")
])

import gradio as gr

def chat(message, history):
    response = agent.invoke({
        "input": message,
        "history": history
    })
    return response.content

demo = gr.ChatInterface(
    fn=chat,
    title="DeepSeek AI助手",
    description="体验智能对话"
)

if __name__ == "__main__":
    demo.launch()

from langchain.cache import InMemoryCache
langchain.llm_cache = InMemoryCache()

from langchain_community.chat_models import SafetySettings

safety_config = SafetySettings(
    hate_speech_filter=True,
    self_harm_filter=True
)

model = ChatDeepSeek(..., safety_settings=safety_config)

/my-agent
├── main.py
├── utils/
│   ├── tools.py
│   └── config.py
├── .env
├── requirements.txt
└── README.md