1. 初始化模型
首先,我们需要初始化要使用的模型。
# First we initialize the model we want to use.
from langchain_openai import ChatOpenAImodel = ChatOpenAI(temperature=0,model="glm-4-plus",openai_api_key="your_api_key",openai_api_base="https://open.bigmodel.cn/api/paas/v4/"
)
2. 定义自定义工具
我们将使用一个自定义工具来返回预定义的天气信息。
from typing import Literal
from langchain_core.tools import tool@tool
def get_weather(location: str):"""Use this to get weather information from a given location."""if location.lower() in ["nyc", "new york"]:return "It might be cloudy in nyc"elif location.lower() in ["sf", "san francisco"]:return "It's always sunny in sf"else:raise AssertionError("Unknown Location, please try again use different expression")tools = [get_weather]
3. 设置检查点
为了支持人机交互模式,我们需要一个检查点。
from langgraph.checkpoint.memory import MemorySavermemory = MemorySaver()
4. 定义图形
使用LangGraph的预构建功能来创建一个REACT代理。
from langgraph.prebuilt import create_react_agentgraph = create_react_agent(model, tools=tools, interrupt_before=["tools"], checkpointer=memory
)
5. 打印工具流
定义一个工具来美化打印流。
def print_stream(stream):"""A utility to pretty print the stream."""for s in stream:message = s["messages"][-1]if isinstance(message, tuple):print(message)else:message.pretty_print()
6. 发送查询并打印结果
发送一个查询并打印结果。
from langchain_core.messages import HumanMessageconfig = {"configurable": {"thread_id": "42"}}
inputs = {"messages": [("user", "what is the weather in SF, CA?")]}print_stream(graph.stream(inputs, config, stream_mode="values"))
输出示例:
================================[1m Human Message [0m=================================
what is the weather in SF, CA?
================================[1m Ai Message [0m==================================
Tool Calls:get_weather (call_9208192248525421766)Call ID: call_9208192248525421766Args:location: SF, CA
7. 获取当前状态
获取当前状态以了解下一步操作。
snapshot = graph.get_state(config)
print("Next step: ", snapshot.next)
输出示例:
Next step: ('tools',)
8. 继续处理流
继续处理流以获取更多结果。
print_stream(graph.stream(None, config, stream_mode="values"))
输出示例:
================================[1m Ai Message [0m==================================
Tool Calls:get_weather (call_9208192248525421766)Call ID: call_9208192248525421766Args:location: SF, CA
================================[1m Tool Message [0m=================================
Name: get_weatherError: AssertionError('Unknown Location, please try again use different expression')Please fix your mistakes.
================================[1m Ai Message [0m==================================
Tool Calls:get_weather (call_9208192214165585200)Call ID: call_9208192214165585200Args:location: San Francisco, CA
9. 更新状态
根据工具的错误信息更新状态。
state = graph.get_state(config)last_message = state.values["messages"][-1]
last_message.tool_calls[0]["args"] = {"location": "San Francisco"}graph.update_state(config, {"messages": [last_message]})
输出示例:
{'configurable': {'thread_id': '42','checkpoint_ns': '','checkpoint_id': '1efa3728-5fb2-6e2f-8004-c28044c43c8b'}}
10. 再次打印流
再次打印流以获取最终结果。
print_stream(graph.stream(None, config, stream_mode="values"))
输出示例:
================================[1m Ai Message [0m==================================
Tool Calls:get_weather (call_9208192214165585200)Call ID: call_9208192214165585200Args:location: San Francisco
================================[1m Tool Message [0m=================================
Name: get_weatherIt's always sunny in sf
================================[1m Ai Message [0m==================================The weather in San Francisco is always sunny!
参考链接:https://langchain-ai.github.io/langgraph/how-tos/create-react-agent-hitl/