LangChain-10(2) 加餐 编写Agent获取本地Docker运行情况 无技术含量只是思路

可以先查看 上一节内容，会对本节有更好的理解。

安装依赖

pip install langchainhub

编写代码

核心代码

@tooldefget_docker_info(docker_name:str)->str:"""Get information about a docker pod container info."""
    result = subprocess.run(['docker','inspect',str(docker_name)], capture_output=True, text=True)return result.stdout

这里是通过执行 Shell的方式来获取状态的。
通过执行

Docker

指令之后，可以获取到一大段的文本内容，此时把这些内容交给大模型去处理，大模型对内容进行提取和推理，最终回答我们。

• 注意@tool注解，没有这个注解的话，无法使用
• 注意要写"""xxx""" 要写明该工具的介绍，大模型将根据介绍来选择是否调用
• 如果3.5的效果不好，可以尝试使用4

from langchain import hub
from langchain.agents import AgentExecutor, tool
from langchain.agents.output_parsers import XMLAgentOutputParser
from langchain_openai import ChatOpenAI
import subprocess

model = ChatOpenAI(
    model="gpt-3.5-turbo",)@tooldefsearch(query:str)->str:"""Search things about current events."""return"32 degrees"@tooldefget_docker_info(docker_name:str)->str:"""Get information about a docker pod container info."""
    result = subprocess.run(['docker','inspect',str(docker_name)], capture_output=True, text=True)return result.stdout

tool_list =[search, get_docker_info]# Get the prompt to use - you can modify this!
prompt = hub.pull("hwchase17/xml-agent-convo")# Logic for going from intermediate steps to a string to pass into model# This is pretty tied to the promptdefconvert_intermediate_steps(intermediate_steps):
    log =""for action, observation in intermediate_steps:
        log +=(f"<tool>{action.tool}</tool><tool_input>{action.tool_input}"f"</tool_input><observation>{observation}</observation>")return log

# Logic for converting tools to string to go in promptdefconvert_tools(tools):return"\n".join([f"{tool.name}: {tool.description}"for tool in tools])

agent =({"input":lambda x: x["input"],"agent_scratchpad":lambda x: convert_intermediate_steps(
            x["intermediate_steps"]),}| prompt.partial(tools=convert_tools(tool_list))| model.bind(stop=["</tool_input>","</final_answer>"])| XMLAgentOutputParser())

agent_executor = AgentExecutor(agent=agent, tools=tool_list)
message1 = agent_executor.invoke({"input":"whats the weather in New york?"})print(f"message1: {message1}")

message2 = agent_executor.invoke({"input":"what is docker pod which name 'lobe-chat-wzk' info? I want to know it 'Image' url"})print(f"message2: {message2}")

执行代码

➜ python3 test10.py
message1: {'input':'whats the weather in New york?', 'output':'The weather in New York is 32 degrees'}
message2: {'input':"what is docker pod which name 'lobe-chat-wzk' info? I want to know it 'Image' url", 'output':'The Image URL for the docker pod named \'lobe-chat-wzk\' is "lobehub/lobe-chat"'}