LangChain 入门教程

LangChain是什么

LangChain是一个agent工程的平台。官网原话为：LangChain is the platform for agent engineering
LangChain 目前有两个开源框架：

• LangChain: 可以帮助你快速的构建agent, 可以选择任何模型的提供商。
• LangGraph：允许你通过底层编排、记忆存储，人工交互的方式控制自定义agnet的每一步（原文：allows you to control every step of your custom agent with low-level orchestration, memory, and human-in-the-loop support）

LangChain更新迭代

目前大模型技术更新迭代非常快，导致LangChain的开源框架也在一直的迭代更新。目前LangChain 跟 LangGraphm目前都已经迭代到v1.0版本了。

LangChain教程

LangChain github的仓库：https://github.com/langchain-ai/langchain

• 安装依赖包

pip install -U langchain

• 安装适配的大语言模型库

pip install -U langchain
pip install langchain-deepseek
pip install openai

创建第一个agent

from langchain.agents import create_agent
import os

# 去deepseek 官网购买tokern,并将DEEPSEEK_API_KEY设置到环境变量中
os.environ['DEEPSEEK_API_KEY'] = 'xxxxxxx'

# function call 回掉查询天气的工具
def get_weather(city: str) -> str:
    """get weather data for a given city"""
    return f"It's always sunny in {city}"

agent = create_agent(
    model='deepseek-chat',
    tools=[get_weather],
    system_prompt='You are a helpful assistant',
)

message = agent.invoke(
    {
        "message": [{"role": "user", "content": "what is the weather in sf"}]
    }
)
print(message['messages'][0].content)

建立真实世界Agent

创建一个实用的天气预报agent需要以下核心的几个步骤：

1.Detailed system prompts for better agent behavior
2.Create tools that integrate with external data
3.Model configuration for consistent responses
4.Structured output for predictable results
5.Conversational memory for chat-like interactions
6.Create and run the agent create a fully functional agent

快速实操：

• 详细的系统提示词让agent的角色跟行为更好

SYSTEM_PROMPT = """You are an expert weather forecaster, who speaks in puns.

You have access to two tools:

- get_weather_for_location: use this to get the weather for a specific location
- get_user_location: use this to get the user's location

If a user asks you for the weather, make sure you know the location. If you can tell from the question that they mean wherever they are, use the get_user_location tool to find their location."""

• 创建工具

from dataclasses import dataclass
from typing import Any
from langchain.tools import tool, ToolRuntime

@tool
def get_weather_for_loction(city: str) -> str:
    """Get weather for a given city."""
    return f"It's always sunny in {city}"

@dataclass
class Context:
    """Custom runtime context schema."""
    user_id: str
@tool
def get_user_location(runtime: ToolRuntime[Context, Any]) -> str:
    """Retrieve user information based on user ID."""
    user_id = runtime.context.user_id
    return f"Florida" if user_id == "1" else "SF"

• 配置模型

model = init_chat_model(
    "deepsee_chat",
    temperature=0.5,
    timeout=10,
    max_tokens=1000
)

• 定义返回格式

@dataclass
class ResponseFormat:
    """Response schema for the agent."""
    # A punny response (always required)
    punny_response: str
    # Any interesting information about the weather if available
    weather_conditions: str | None = None

• 会话记忆

from langgraph.checkpoint.memory import InMemorySaver
checkpointer = InMemorySaver()

• 创建并运行agent

from langchain.agents import create_agent
agent = create_agent(
    model=model,
    system_prompt=SYSTEM_PROMPT,
    tools=[get_weather_for_loction, get_user_location],
    context_schema=Context,
    response_format=ResponseFormat,
    checkpointer=checkpointer
)

config = {"configurable": {"thread_id": "1"}}

response = agent.invoke(
    {"messages": [{"role": "user", "content": "what is the weather outside?"}]},
    config=config,
    context=Context(user_id="1")
)
print(response['structured_response'])

response = agent.invoke(
    {"messages": [{"role": "user", "content": "thank you"}]},
    config=config,
    context=Context(user_id="1")
)
print(response['structured_response'])

完整demo事例

import os
from dataclasses import dataclass
from typing import Any

from langchain.agents import create_agent
from langchain.tools import tool, ToolRuntime
from langchain.chat_models import  init_chat_model
from langgraph.checkpoint.memory import InMemorySaver

os.environ['DEEPSEEK_API_KEY'] = 'xxxxxxxxxxxxx'

SYSTEM_PROMPT = """You are an expert weather forecaster, who speaks in puns.

You have access to two tools:

- get_weather_for_location: use this to get the weather for a specific location
- get_user_location: use this to get the user's location

If a user asks you for the weather, make sure you know the location. If you can tell from the question that they mean wherever they are, use the get_user_location tool to find their location."""

@tool
def get_weather_for_loction(city: str) -> str:
    """Get weather for a given city."""
    return f"It's always sunny in {city}"

@dataclass
class Context:
    """Custom runtime context schema."""
    user_id: str
@tool
def get_user_location(runtime: ToolRuntime[Context, Any]) -> str:
    """Retrieve user information based on user ID."""
    user_id = runtime.context.user_id
    return f"Florida" if user_id == "1" else "SF"

@dataclass
class ResponseFormat:
    """Response schema for the agent."""
    # A punny response (always required)
    punny_response: str
    # Any interesting information about the weather if available
    weather_conditions: str | None = None

model = init_chat_model(
    "deepseek-chat",
    temperature=0.5,
    timeout=10,
    max_tokens=1000
)

checkpointer = InMemorySaver()


agent = create_agent(
    model=model,
    system_prompt=SYSTEM_PROMPT,
    tools=[get_weather_for_loction, get_user_location],
    context_schema=Context,
    response_format=ResponseFormat,
    checkpointer=checkpointer
)

config = {"configurable": {"thread_id": "1"}}

response = agent.invoke(
    {"messages": [{"role": "user", "content": "what is the weather outside?"}]},
    config=config,
    context=Context(user_id="1")
)
print(response['structured_response'])

response = agent.invoke(
    {"messages": [{"role": "user", "content": "thank you"}]},
    config=config,
    context=Context(user_id="1")
)
print(response['structured_response'])