GPTComet

GPTComet

Made with VHS

Description

GPTComet is an AI-powered command-line tool that generates and creates conventional Git commit messages based on staged changes. It leverages various Large Language Models (LLMs) to analyze the diff and produce meaningful commit messages, streamlining the development workflow.

Features

Multiple LLM Providers: Supports a wide range of LLM providers, including OpenAI, Claude, Gemini, Mistral, Ollama, and more.
Interactive Provider Configuration: Easily configure new providers and their settings through an interactive command-line interface.
Customizable Commit Message Style: Generate both brief and detailed commit messages using configurable prompt templates.
Commit Message Translation: Translate generated commit messages into various languages.
Filtered Diff: Ignores specified files (e.g., lock files, specific file types) in the diff to generate more relevant commit messages.
Flexible Configuration: Manage settings through a YAML configuration file, supporting options like API keys, models, output language, and more.
Dry Run Mode: Preview the generated commit message without actually creating a commit.
Rich Commit Messages: Generate detailed commit messages with a title, summary, and bullet points describing the changes when using the --rich flag.
SVN Support: Support for working with SVN repositories, allowing for commit messages that include changes from both Git and SVN repositories.

Installation

Release

You can install by install script:

curl -sSL https://raw.githubusercontent.com/belingud/go-gptcomet/refs/heads/master/install.sh | bash

Windows:

irm https://raw.githubusercontent.com/belingud/go-gptcomet/refs/heads/master/install.ps1 | iex

Or download from Releases.

Build

Prerequisites

Go (version 1.18 or later)

Build from Source

Clone the repository:

git clone https://github.com/belingud/go-gptcomet.git
cd go-gptcomet

Build the project:
```
go build -o gptcomet .
```
(Optional) Move the gptcomet binary to a directory in your PATH for easy access.

Setup

Configuration Methods

Direct Configuration
- Configure directly in ~/.config/gptcomet/gptcomet.yaml.
Interactive Setup
- Use the gmsg newprovider command for guided setup.

Setup Guide

./gptcomet newprovider

    Select Provider           
                              
  > 1. azure                  
    2. chatglm                
    3. claude                 
    4. cohere                 
    5. deepseek               
    6. gemini                 
    7. kimi                   
    8. mistral                
    9. ollama                 
    10. openai                
    11. sambanova             
    12. silicon               
    13. tongyi                
    14. vertex                
    15. xai                   
    16. Input Manually        
                              
    ↑/k up • ↓/j down • ? more

Select the provider you want to configure. Then enter the API base, API key, max tokens and model name. Some provider need you input more information.

gmsg newprovider
Selected provider: silicon
Configure provider:

Previous inputs:
  Enter Silicon API base: https://api.siliconflow.cn/v1
  Enter API key: sk-awz*********************************************
  Enter max tokens: 1024

Enter Enter model name (default: Qwen/Qwen2.5-7B-Instruct):
> Qwen/Qwen2.5-7B-Instruct                 

Provider silicon already has a configuration. Do you want to overwrite it? (y/N): y

Provider silicon configured successfully!

Usage

Basic Usage

Stage your changes in a Git repository:
```
git add .
```
Run GPTComet to generate and create a commit message:
```
./gptcomet commit
```
The tool will analyze the staged changes, generate a commit message, and prompt you to confirm or edit it before creating the commit.

Using a specific language

You can change the output language by setting the output.lang configuration option:

./gptcomet config set output.lang zh-cn

Dry Run

To preview the generated commit message without committing, use the --dry-run flag:

./gptcomet commit --dry-run

SVN

To use SVN instead of Git, set the --svn flag:

./gptcomet commit --svn

Configuring a New Provider

To configure a new LLM provider:

./gptcomet newprovider

This will guide you through selecting a provider and entering the required configuration values (e.g., API key, model name).

Managing Configuration

The gptcomet config command provides subcommands for managing the configuration file:

get <key>: Get the value of a configuration key.
list: List the entire configuration content.
reset: Reset the configuration to default values (optionally reset only the prompt section with --prompt).
set <key> <value>: Set a configuration value.
path: Get the configuration file path.
remove <key> [value]: Remove a configuration key or a value from a list.
append <key> <value>: Append a value to a list configuration.
keys: List all supported configuration keys.

Example:

./gptcomet config set openai.api_key "your_openai_api_key"
./gptcomet config get output.lang
./gptcomet config list

Generating Rich Commit Messages

To generate a more detailed commit message, use the --rich flag:

./gptcomet commit --rich

This will use the rich_commit_message prompt template, resulting in a commit message with a title, summary, and potentially bullet points outlining the changes.

Configuration

GPTComet stores its configuration in a YAML file located at ~/.config/gptcomet/gptcomet.yaml.

Supported Configuration Keys

Here's a summary of the main configuration keys:

Key	Description	Default Value
`provider`	The name of the LLM provider to use.	`openai`
`file_ignore`	A list of file patterns to ignore in the diff.	(See `config.go`)
`output.lang`	The language for commit message generation.	`en`
`output.rich_template`	The template to use for rich commit messages.	`<title>:<summary>\n\n<detail>`
`console.verbose`	Enable verbose output.	`true`
`<provider>.api_base`	The API base URL for the provider.	(Provider-specific)
`<provider>.api_key`	The API key for the provider.
`<provider>.model`	The model name to use.	(Provider-specific)
`<provider>.retries`	The number of retry attempts for API requests.	`2`
`<provider>.proxy`	The proxy URL to use (if needed).
`<provider>.max_tokens`	The maximum number of tokens to generate.	`2048`
`<provider>.top_p`	The top-p value for nucleus sampling.	`0.7`
`<provider>.temperature`	The temperature value for controlling randomness.	`0.7`
`<provider>.frequency_penalty`	The frequency penalty value.	`0`
`<provider>.extra_headers`	Extra headers to include in API requests (JSON string).	`{}`
`<provider>.completion_path`	The API path for completion requests.	(Provider-specific)
`<provider>.answer_path`	The JSON path to extract the answer from the API response.	(Provider-specific)
`prompt.brief_commit_message`	The prompt template for generating brief commit messages.	(See `defaults/defaults.go`)
`prompt.rich_commit_message`	The prompt template for generating rich commit messages.	(See `defaults/defaults.go`)
`prompt.translation`	The prompt template for translating commit messages.	(See `defaults/defaults.go`)

Note: <provider> should be replaced with the actual provider name (e.g., openai, gemini, claude).

Contribution

Contributions to GPTComet are welcome! Please refer to the CONTRIBUTING.rst file for guidelines on how to contribute.

Testing

License

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

Acknowledgements

This project is inspired by various AI-powered Git tools and utilizes several open-source libraries.
Thanks to the developers and maintainers of the LLMs used in this project.
The gptcomet directory contains a Python implementation of a similar tool, which served as a reference for certain features and concepts.

# README