Codebuddy

Codebuddy is an advanced command-line tool designed for executing user queries using OpenAI's GPT models within a Git-integrated shell environment. It provides intelligent automation of project file and configuration management, enabling efficient and dynamic interaction with projects.

Table of Contents

• Features
• Installation
• Usage
• Configuration
• Architecture
• Dependencies
• Development
• Contributing
• License

Features

• Interactive Shell: Engage in an interactive shell environment with Git-based rollback capabilities to ensure safe modifications.
• AI-Driven Query Processing: Leverage OpenAI's GPT models to intelligently handle user queries and automate project management tasks.
• Seamless File Management: Facilitates project updates, file modifications, and configuration adjustments in response to user interactions.
• Robust Command Execution: Securely execute system commands as part of an integrated workflow.

Installation

Codebuddy requires Python 3.8 or later. Install using Poetry with:

pip install pycodebuddy

Usage

Enter the interactive shell ...

$ codebuddy
$ Entering interactive shell. Type 'exit' to exit or use Ctrl+C to abort.
$ User:

... or execute direct queries via:

codebuddy "<your request here>"

Ensure your OpenAI API key is configured in the config.yaml.

Configuration

Codebuddy uses a configuration file for setting various parameters. This can be found at .user_config.yaml in the user's home directory, but it can be overridden by a config.yaml located in the project's root directory. Below are the configuration parameters:

• api_key: The API key for authentication (e.g., OpenAI). Default is 'api-key'.
• instructions: Additional instructions or notes used within to prepare the chat bot for his task.
• allow_command_execution: Boolean flag for allowing/disallowing command execution. Default is False.
• client_type: The type of client, like 'openai'. Default is 'openai'.
• base_url: Optional base URL for the openai backend to use. useful to run local models.
• history_size: Size of the history to maintain. Default is 10.
• model: The specified model to use, such as 'gpt-4o'.

Example config.yaml file:

api_key: 'your-api-key-here'
instructions: 'additional setup instructions'
allow_command_execution: true
client_type: 'openai'
base_url: 'http://localhost:11434/v1' # to use ollama
history_size: 10
model: 'gpt-4o'

Architecture

Codebuddy's architecture is modular and comprises the following components:

• main.py: The entry point for the interactive shell and query execution.
• openai_client.py: Manages API integration with OpenAI for query processing.
• file_handler.py: Oversees file operations, including file management.
• command_executor.py: Responsible for executing system commands securely.
• config_loader.py: Loads configuration parameters from the YAML files.
• transaction_handler.py: Utilizes Git for safe rollbacks and transaction management.

Dependencies

Managed via Poetry and specified within pyproject.toml. Key dependencies include:

• openai
• PyYAML

Development

Install additional development tools using:

poetry install --with dev

Use black for formatting, flake8 for style checks, mypy for type checks, and isort for organizing imports.

Contributing

Contributions are encouraged. Fork, implement improvements, and submit a pull request with test coverage for new features.

License

Licensed under MIT License. See LICENSE for more information.