langchain RunableBranch 分类判断选择不同链

最新推荐文章于 2025-05-03 12:49:02 发布

木下瞳

最新推荐文章于 2025-05-03 12:49:02 发布

阅读量651

点赞数 1

分类专栏： langchain 文章标签： langchain

本文链接：https://blog.youkuaiyun.com/zjkpy_5/article/details/137712949

版权

langchain 专栏收录该内容

17 篇文章

订阅专栏

本文介绍了如何使用Python编程构建一个基于LangChain的模型，该模型可以对用户的问题进行分类（关于LangChain或其他），然后针对不同类别提供相应的详细回答或一般性答案。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

import os
from dotenv import load_dotenv
from langchain_community.llms import Tongyi
load_dotenv('key.env')  # 指定加载 env 文件
key = os.getenv('DASHSCOPE_API_KEY')  # 获得指定环境变量
DASHSCOPE_API_KEY = os.environ["DASHSCOPE_API_KEY"]  # 获得指定环境变量
model = Tongyi(temperature=1)

from langchain_core.prompts import PromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_core.runnables import RunnableBranch

# 构建分类判断链：识别用户的问题应该属于哪个（指定的）分类
chain = (
        PromptTemplate.from_template(
            """Given the user question below, classify it as either being about `LangChain` or `Other`.
    
    Do not respond with more than one word.
    
    <question>
    {question}
    </question>
    
    Classification:"""
        )
        | model
        | StrOutputParser()
)

# 构建内容问答链和默认应答链
langchain_chain = (
        PromptTemplate.from_template(
            """You are an expert in LangChain. Respond to the following question in one sentence:
    
    Question: {question}
    Answer:"""
        )
        | model
)
general_chain = (
        PromptTemplate.from_template(
            """Respond to the following question in one sentence:
    
    Question: {question}
    Answer:"""
        )
        | model
)

# 通过 RunnableBranch 构建条件分支并附加到主调用链上
branch = RunnableBranch(
    (lambda x: "langchain" in x["topic"].lower(), langchain_chain),
    general_chain,
)
full_chain = {"topic": chain, "question": lambda x: x["question"]} | branch

print(full_chain.invoke({"question": "什么是 LangChain?"}))
print(full_chain.invoke({"question": "1 + 2 = ?"}))