devinyf/langchain_qianwen

灵积-通义千问 Langchain

探索通义千问 Api 在 langchain 中的使用参考借鉴 openai langchain 的实现目前在个人项目工具中使用

NOTE: langchian 已经带有了一个合并的 Tongyi 实现, 当时写这个项目的时候 Tongyi 的功能还不够完善, 不过随着后续的迭代应该已经没问题了建议优先考虑通过以下方式使用

from langchain_community.llms.tongyi import Tongyi
from langchain_community.chat_models.tongyi import ChatTongyi

Install

pip 会同时安装依赖库: Langchain 和 Dashscope-SDK

pip install langchain-qianwen

Clone 项目手动安装

git clone ... && cd langchain_qianwen
pip install -r requirements.txt

 # 建议运行 pytest 单元测试确认功能运行正常，防止依赖库出现 breaking change
pip install pytest
pytest

使用前置条件：

了解 Langchain langchain文档
在阿里云开发参考文档申请并创建API-KEY
设置 api_key 环境变量 export DASHSCOPE_API_KEY="YOUR_DASHSCOPE_API_KEY"

支持 LCEL(LangChain Expression Language) 语法

from langchain.prompts import PromptTemplate
from langchain_qianwen import Qwen_v1

if __name__ == "__main__":
    jock_template = "给我讲个有关 {topic} 的笑话"
    prompt = PromptTemplate.from_template(jock_template)

    llm = Qwen_v1(
        model_name="qwen-turbo",
        temperature=0.18,
        streaming=True,
    )

    chain = prompt | llm

    for s in chain.stream({"topic": "产品经理"}):
        print(s, end="", flush=True)

支持异步调用 async callback handler

p.s. 目前 llm 模型 (Qwen_v1) 可以使用 AsyncIteratorCallbackHandler, chatmodel(ChatQwen_v1 待更新这个我还用不到...)

from langchain.callbacks.streaming_aiter import AsyncIteratorCallbackHandler
from langchain_qianwen import Qwen_v1
from langchain.chains import ConversationChain
from langchain.memory import ConversationBufferMemory

import asyncio

async def use_async_handler(input):
    handler = AsyncIteratorCallbackHandler()
    llm = Qwen_v1(
        model_name="qwen-turbo",
        streaming=True,
        callbacks=[handler], 
    )

    memory = ConversationBufferMemory()
    chain = ConversationChain(
        llm=llm,
        memory=memory,
        verbose=True,
    )

    asyncio.create_task(chain.apredict(input=input))

    return handler.aiter()

async def async_test():
    async_gen = await use_async_handler("hello")
    async for i in async_gen:
        print(i)

if __name__ == "__main__":
    asyncio.run(async_test())

chat_models

from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
from langchain_qianwen import ChatQwen_v1
from langchain.schema import (
    HumanMessage,
)

if __name__ == "__main__":
    chat = ChatQwen_v1(
        model_name="qwen-turbo",
        streaming=True,
        callbacks=[StreamingStdOutCallbackHandler()],
    )

    chat([HumanMessage(content="举例说明一下 PHP 为什么是世界上最好的语言")])

使用 agent 增加网络搜索功能

from langchain.agents import load_tools, AgentType, initialize_agent
from langchain_qianwen import Qwen_v1

if __name__ == "__main__":
    llm = Qwen_v1(
        model_name="qwen-plus",
    )
    ## 需要去 serpapi 官网申请一个 api_key
    tool_names = ["serpapi"]
    tools = load_tools(tool_names)

    agent = initialize_agent(tools=tools,
                             llm=llm,
                             agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
                             verbose=True)
    agent.run("今天北京的天气怎么样?")

使用 embedding 提取文档中的信息

from langchain.embeddings.dashscope import DashScopeEmbeddings
from langchain.vectorstores import Chroma
from langchain.text_splitter import CharacterTextSplitter
from langchain.chains import RetrievalQA
from langchain.document_loaders import DirectoryLoader

from langchain_qianwen import Qwen_v1

if __name__ == "__main__":
    llm = Qwen_v1(
        model_name="qwen-turbo",
    )
    loader = DirectoryLoader("./assets", glob="**/*.txt")
    document = loader.load()

    text_splitter = CharacterTextSplitter(chunk_size=2048, chunk_overlap=0)
    texts = text_splitter.split_documents(document)
    embeddings = DashScopeEmbeddings(
            model="text-embedding-v1",
    )

    print(f"text length: {len(texts)}")

    # 使用 embedding engion 将 text 转换为向量
    db = Chroma.from_documents(texts, embeddings)
    retriever = db.as_retriever()

    # qa = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=retriever, return_source_documents=True)
    qa = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=retriever)
    query = "文章中的工厂模式使用例子有哪些??"
    rsp = qa.run({"query": query})
    print(rsp)

更多使用请查看 langchain 官方文档和 examples 目录

devinyf / langchain_qianwen