langchain wechat

langchain-wechat

A WeChat chatbot based on fastapi + langchain + itchat, inspired by the chatgpt-on-wechat project

Your star is the biggest encouragement to me?

---

Note

This project is only for personal learning and communication. Users must abide by OpenAI's terms of use and **Chinese laws and regulations** and may not use it for illegal purposes.

According to the requirements of the "Interim Measures for the Management of Generative Artificial Intelligence Services", please do not provide any unregistered generative artificial intelligence services to the public in China.

Function

Assistants

When executed in the default assistant mode, various tools will be configured, and LLM automatically selects appropriate tools to call, such as calling the DuCKDuckGo search engine, summarizing the search results and returning them to the user.

Supported message types

• Receive individual text message replies
• Receive group message@Current user text message reply
• ...

Supported models

• openai
• ...

tools

• duckduckgo search
• ...

RAGBot (under development)

You need to provide a set of documents and then, based on the user's question, return the most relevant document snippets.

Environment setup

1. The OpenAI account preparation project uses the OpenAI interface by default. You need to go to the OpenAI registration page to create an account. After creating the account, go to the API management page to create an API Key and save it. The interface requires overseas network access and binding of credit card payment.

2. Add API Key to environment variables

    export OPENAI_API_KEY= " sk-xxxxxxxxxxxxxxxxxxxxxx "
   

3. Install Python The code is developed and tested through Python3.10. You need to install python 3.10. After downloading, click the installation fool-proof next step.

Install redis

redis is used to store WeChat user chat records

docker run -d -p 6379:6379 -p 8001:8001 redis/redis-stack:latest

Install dependencies

poetry install

or

pip install -r requirements.txt

copy .env

cp config/.env.example config/.env

run

uvicorn src.main:app

Configuration items

Configure in config/.env file

 # openai 模型
OPENAI_MODEL= " gpt-3.5-turbo-1106 "
# openai api key
OPENAI_API_KEY= " your-openai-api-key "

# 模型 temperature
AI_TEMPERATURE=0.7
# 模型系统角色提示词
AI_SYSTEM_ROLE_PROMPT= "系统： "

# 聊天记录保存最大长度
CHAT_MAX_MESSAGE_HISTORY_LENGTH=10
# 触发聊天记录总结的阈值
CHAT_MESSAGE_HISTORY_SUMMARY_THRESHOLD=5


# 微信是否开启热重载
WECHAT_HOT_RELOAD=False
# 微信用户数据保存路径
WECHAT_USER_DATA_STORAGE_PATH= " wechat.pkl "