Skip to main content

A tool to generate commit messages using AI

Project description

lazym

lazym is an AI-powered tool that automatically generates meaningful commit messages for your Git repositories.

lazym is inspired by jen-Ya/commitgpt

Features

  • Generates commit messages based on staged changes
  • Without the risk of leaking code snippets to external LLM services
  • Integrates seamlessly with your Git workflow
  • Uses advanced language models for intelligent message generation

Requirements

Command to pull Llama 3.1:8b model for Ollama:

$ ollama run llama3.1:8b

Installation

To install lazym, follow these steps:

  1. Ensure you have Python 3.7+ and pip installed on your system.

  2. Install lazym using pip:

    pip install lazym
    
  3. After installation, navigate to any Git repository and run the following command to set up the Git hook:

    lazym install
    

    This will install the necessary Git hook to enable automatic commit message generation.

Usage

Once installed and set up, lazym works in two ways:

  1. Automatically when you make a commit: Simply stage your changes as usual and run git commit. lazym will generate a commit message based on your staged changes.

  2. Manually by running lazym ci "<hints for LLM>": This command allows you to generate a commit message with additional context provided to the LLM.

After generating the commit message, you'll be presented with three options:

  1. Accept and commit: Use the generated message as-is and commit.
  2. Edit message: Modify the generated message before committing.
  3. Cancel commit: Abort the commit process.

This allows you to benefit from the AI-generated suggestions while maintaining full control over your commit messages. When editing, you can start with the AI-generated message and make any necessary adjustments.

Commands

  • lazym install: Install the prepare-commit-msg hook in the current Git repository.
  • lazym uninstall: Uninstall the prepare-commit-msg hook from the current Git repository.
  • lazym ci "<optional hints>": Generate a commit message with optional additional context.

Configuration

lazym can be configured using a config.ini file located at ~/.config/lazym/config.ini. The following options are available:

[DEFAULT]

  • model: The AI model to use for generating commit messages.
    • Default: "llama3.1:8b"
    • Options:
      • Ollama models: Use the model name directly (e.g., "llama3.1:8b", "codellama:7b")
      • Groq models: Prefix with "groq:" (e.g., "groq:mixtral-8x7b-32768", "groq:llama2-70b-4096")

[!NOTE] To use Groq models, you need to:

  1. Sign up at groq.com to get an API key
  2. Set the GROQ_API_KEY environment variable with your API key
  • temperature: Controls the randomness of the AI's responses.

    • Default: 0.8
    • Range: 0.0 to 1.0
    • Lower values make responses more focused and deterministic
    • Higher values make responses more creative and varied
  • message_format: The format to apply to the generated commit message.

    • Default: "lowercase"
    • Options:
      • "lowercase": Converts the first character of the message to lowercase.
      • "sentence case": Capitalizes the first character of each sentence.
      • Any other value will keep the original format as generated by the AI.
  • prompt: The prompt template used to generate commit messages.

    • Must include the {diff} placeholder where the code changes should be inserted
    • Default: empty (uses built-in prompt)
    • You can customize this to match your team's commit message style

Customizing Prompts

You can customize how lazym generates commit messages by providing your own prompt template. There are two ways to do this:

  1. Using config.ini: Suitable for simple, one-line prompts. Specify the prompt option in your config.ini file. This method is ideal for straightforward prompts that don't require complex formatting or multiple lines.

    Example:

    [DEFAULT]
    prompt = "Generate a concise commit message for the following changes: {diff}"
    
  2. Using prompt.txt: Suitable for more complex, multi-line prompts. Create a prompt.txt file in the ~/.config/lazym/ directory. If this file exists, its contents will override the prompt setting in config.ini. This method is ideal for detailed prompts that require specific formatting or guidelines.

    Example prompt.txt:

    Generate a conventional commit message for these changes.
    Format: <type>(<scope>): <description>
    Types: feat, fix, docs, style, refactor, test, chore
    Keep the message under 72 chars.
    
    Changes:
    {diff}
    

If no prompt is specified in either config.ini or prompt.txt, lazym will use its built-in prompt template that follows general Git best practices.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

lazym-0.12.0.tar.gz (10.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

lazym-0.12.0-py3-none-any.whl (9.5 kB view details)

Uploaded Python 3

File details

Details for the file lazym-0.12.0.tar.gz.

File metadata

  • Download URL: lazym-0.12.0.tar.gz
  • Upload date:
  • Size: 10.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.6

File hashes

Hashes for lazym-0.12.0.tar.gz
Algorithm Hash digest
SHA256 f59c3862e76f59cf40ae0cd52b6fb497b6fd6a1def79be3c2a4c465d2e9d98dc
MD5 741038ad6fcacfaad9cbb4aad189f0b6
BLAKE2b-256 3d9b055d3c1c2b1f1154e550d0aa385e366c41cabcc411cb786f945f1909287e

See more details on using hashes here.

File details

Details for the file lazym-0.12.0-py3-none-any.whl.

File metadata

  • Download URL: lazym-0.12.0-py3-none-any.whl
  • Upload date:
  • Size: 9.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.6

File hashes

Hashes for lazym-0.12.0-py3-none-any.whl
Algorithm Hash digest
SHA256 194b48626e6cedff74553e7059695f2d75cb406bbbf880d3e60142c97c57c8b0
MD5 4b2c37d44b57ea20e716165bc762e0b3
BLAKE2b-256 2df9a8933527be9e4463996cbbfb653d5a9b7803fd85cee0652783bfce20ca41

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page