FeiFei: Automated GitHub Issue Responder Bot

English | 中文

This repository contains a Python script that automates responses to new issues in specified GitHub repositories using a Language Model (LLM). The bot leverages OpenAI's GPT models, Hugging Face models, or other supported models to generate responses based on the repository's content.

Features

• Automated Responses: Monitors specified GitHub repositories for new issues and automatically generates responses using an LLM.
• LLM Integration: Supports OpenAI, Hugging Face, Anthropic, Google, Qianfan, Ollama, or local models for generating responses.
• Vector Stores: Builds or loads a vector store of the repository's content for context-aware responses.
• Environment Configuration: Utilizes environment variables and a YAML configuration file for flexible setup.
• Thread-Safe Operations: Implements locking mechanisms to ensure thread-safe file operations.
• Extensible Design: Modular functions allow for easy customization and extension of capabilities.

Table of Contents

• Prerequisites
• Installation
• Configuration
• Usage
• Customization
• Contributing
• License

Prerequisites

• Python 3.10 or higher
• Access tokens for GitHub, OpenAI, Hugging Face, or other supported backends

Installation

1. Clone the Repository

   git clone https://github.com/GreatV/feifei.git
   cd feifei

2. Create a Virtual Environment

   python3 -m venv venv
   source venv/bin/activate  # On Windows use 'venv\Scripts\activate'

3. Install Dependencies

   pip install -r requirements.txt

Configuration

Environment Variables

Create a .env file in the root directory and add the following variables:

GITHUB_TOKEN=your_github_token
OPENAI_API_KEY=your_openai_api_key  # Optional, if using OpenAI
HUGGINGFACEHUB_API_TOKEN=your_huggingface_token  # Optional, if using Hugging Face

YAML Configuration

Create or modify the config.yaml file in the root directory:

github_base_url: "https://api.github.com"
llm_backend: "openai"  # Options: 'openai', 'huggingface', 'local', 'anthropic', 'google', 'qianfan', 'ollama'
embedding_backend: "openai"  # Options: 'openai', 'huggingface', 'local'
model_name: "gpt-3.5-turbo"  # LLM model name
embedding_model_name: "sentence-transformers/all-MiniLM-L6-v2"
check_interval: 300  # Time in seconds between checks
repositories:
  yourusername/your-repo:
    start_issue_number: 1
    branch: "main"  # Branch to monitor
    recent_period: 30  # Period in days to consider recent issues
enable_compression: true  # Enable or disable response compression

• github_base_url: The base URL for GitHub API requests.
• llm_backend: The backend to use for the language model.
• embedding_backend: The backend to use for embeddings.
• model_name: The name of the LLM model.
• embedding_model_name: The name of the embeddings model.
• check_interval: How often (in seconds) the bot checks for new issues.
• repositories: A list of repositories to monitor and their settings.
  • start_issue_number: The issue number to start monitoring from.
  • branch: The branch to monitor for changes.
  • recent_period: The period in days to consider issues and discussions as recent.
• enable_compression: Enable or disable response compression.

Usage

Run the main script:

python main.py

To enable debug mode, use the --debug command line argument:

python main.py --debug

The bot will start monitoring the specified repositories for new issues and respond automatically.

Customization

Modifying the Prompt Template and Signature

The prompt used to generate responses and the signature appended to each response can now be customized in the config.yaml file instead of being hardcoded in the code.

Update the prompt_template and signature fields in your config.yaml:

prompt_template: |
  You are a smart assistant capable of answering user questions based on provided documents.

  Question: {input}

  Here are some potentially useful documents:
  {context}

  Please provide a detailed answer based on the above information, in the same language as the question. If possible, reference related Issues or Discussions and provide links.

  Answer:

signature: "\n\n---\n*Response generated by 🤖 feifei-bot | {model_name}*"

Changing the LLM Backend

You can switch between OpenAI, Hugging Face, Anthropic, Google, Qianfan, Ollama, or a local model by updating the llm_backend in config.yaml:

llm_backend: "huggingface"  # or 'local', 'anthropic', 'google', 'qianfan', 'ollama'

Ensure that you have the necessary tokens and models installed locally if you choose 'local'.

Adjusting Retrieval Settings

The number of documents retrieved for context can be adjusted in the search_kwargs parameter:

retriever = vectorstore.as_retriever(search_kwargs={"k": 5})

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the Repository

2. Create a New Branch

   git checkout -b feature/your-feature-name

3. Commit Your Changes

   git commit -am 'Add new feature'

4. Push to the Branch

   git push origin feature/your-feature-name

5. Create a Pull Request

License

This project is licensed under the Apache-2.0 license.