Git-Vector

Git-Vector is a command-line tool that allows developers to interact with their Git repositories using OpenAI models. It provides a conversational interface to help you understand and navigate your codebase by leveraging the full context of your repository.

Table of Contents

• Git-Vector
  • Table of Contents
  • Features
  • Installation
  • Usage
  • Configuration
  • License
  • Author
  • Acknowledgments

Features

• Conversational Interface: Interactively ask questions about your codebase and receive detailed explanations.
• Full Repository Context: The tool indexes your entire Git repository to provide context-aware responses.
• Caching Mechanism: Embeddings are cached to improve performance on subsequent runs.
• Customizable Models: Supports different OpenAI models for both embeddings and chat completions.
• Configurable Parameters: Adjust maximum tokens for prompts and responses to suit your needs.

Installation

pip install git-vector

Usage

1. Set Up Environment Variables:

   Export your OpenAI API key as an environment variable (or use a .env file):

   export OPENAI_API_KEY=your_api_key_here
   

2. Run the Application:

   git-vector --repo-dir /path/to/your/git/repository
   

   This will start the interactive CLI for the user to chat with the codebase.

Configuration

• --repo-dir: (Required) The path to the Git repository.
• --embedding-model: The OpenAI model to use for embeddings (default: text-embedding-3-small).
• --chat-model: The OpenAI model to use for chat completions (default: gpt-4o-mini).
• --max-prompt-tokens: Maximum number of tokens for the prompt (default: 2000).
• --max-response-tokens: Maximum number of tokens for the response (default: 500).

License

This project is licensed under the MIT License. See the LICENSE file for details.

Author

• Markus Blomqvist - [email protected]

Acknowledgments

• OpenAI for providing the API that powers this tool.
• GitPython for interacting with Git repositories.
• NumPy and scikit-learn for numerical and machine learning.