Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/Sweet-Papa-Technologies/egit

A.I. tools and workflows for Git
https://github.com/Sweet-Papa-Technologies/egit

Last synced: 8 days ago
JSON representation

A.I. tools and workflows for Git

Awesome Lists containing this project

README 

        
# eGit - Extended Git CLI with LLM Capabilities



eGit is a CLI tool that extends Git with LLM capabilities, making it easier to work with commit messages, generate release notes, and understand code changes.



## Table of Contents



- [Features](#features)

- [Requirements](#requirements)

- [Installation](#installation)

- [Usage](#usage)

- [Environment Variables](#environment-variables)

- [Contributing](#contributing)

- [License](#license)

- [Full Documentation](#full-documentation)



## Features

- 🤖 AI-powered commit message generation from staged changes

- 📝 Intelligent release notes generation with automatic tagging

- 🔍 Smart change summarization for commits and branches

- ⚙️ Flexible configuration for different LLM providers



## Requirements

- macOS or Windows (Not tested on Linux, but should work and will be fully supported in the future)

- Python 3.10 or higher preinstalled

- Git preinstalled

- One of the following LLM providers:

  - Ollama (default)

  - LM Studio (Recommended for local models)

  - OpenAI

  - Anthropic

  - Google Gemini



## Installation



### Windows

```powershell

# Download eGit

git clone https://github.com/Sweet-Papa-Technologies/egit.git

cd egit



# Run the installer

.\install.ps1

```



### macOS/Linux

```bash

# Download eGit

git clone https://github.com/Sweet-Papa-Technologies/egit.git

cd egit



# Make the installer executable

chmod +x install.sh



# Run the installer

./install.sh

```



The installer will:

1. Check for Python 3.10+ and Git

2. Create a virtual environment

3. Install required dependencies

4. Install eGit in development mode

5. Add eGit to your PATH (may require admin privileges on Windows)



### Manual PATH Configuration

If you don't have admin privileges on Windows, you can manually add eGit to your PATH:

1. Run the `setx` command shown during installation, or

2. Add the installation directory to your PATH environment variable



Alternatively, you can manually call eGit from the install directory:



```shell

# macOS or Linux

~/.egit/.venv/bin/egit



# Windows

C:\ProgramData\egit\.venv\Scripts\egit.exe

```



## Usage

NOTE: API KEY value must be set, otherwise LLM calls will fail. If your endpoint does not take an API key, please set a fake value such as `sk-123`. The value will be ignored if your endpoint does not need it.



#### How this would typically be used:

```bash

git add .

egit summarize --commit

git push

```



### Summarize and Commit Changes

```bash

# View summary of staged changes

egit summarize --staged



# View summary of current branch changes

egit summarize --branch



# Generate summary and automatically commit staged changes

egit summarize --commit

```



### Generate Release Notes and Tags

```bash

# Generate release notes for version (draft mode)

egit release-notes 1.0.0 --draft



# Create and push an annotated tag with release notes

egit release-notes 1.0.0 --tag

```



### Configuration Management

```bash

# View current configuration

egit config --show



# Set configuration values

egit config --set llm_provider --value ollama

egit config --set llm_model --value ollama/llama3.2:3b

egit config --set llm_api_key --value your_api_key

```



### Version Information

```bash

egit --version

# or

egit -v

```



## Environment Variables

Settings can be set via Env Vars or via CLI flags. These are the available environment variables:



- `GIT_EXECUTABLE`: Path to Git executable (default: system git)

- `LLM_PROVIDER`: LLM provider to use (default: ollama)

- `LLM_MODEL`: Model name (default: ollama/llama3.2:3b)

- `LLM_API_KEY`: API key for the LLM service (default: sk-123 for Ollama)

- `LLM_API_BASE`: Base URL for the LLM API (default: http://localhost:11434)

- `LLM_MAX_TOKENS`: Maximum tokens for LLM response (default: 4096)

- `LLM_TEMPERATURE`: Temperature for LLM sampling (default: 0.7)



## Example LLM Provider Setup and Commands

Settings can be set via Env Vars or via CLI flags. These are the CLI Flags



Max Tokens and Temperature are optional, but to set them, use the following commands:



`egit config --set llm_max_tokens --value 4096` 

`egit config --set llm_temperature --value 0.7`



### Ollama

```bash

egit config --set llm_provider --value ollama

egit config --set llm_model --value ollama/llama3.2:3b

egit config --set llm_api_key --value sk-123

egit config --set llm_api_base --value http://localhost:11434

```



### LM Studio

```bash

egit config --set llm_provider --value lmstudio

egit config --set llm_model --value lm_studio/hermes-3-llama-3.1-8b

egit config --set llm_api_key --value sk-123

egit config --set llm_api_base --value http://localhost:1221/v1

```



### OpenAI

```bash

egit config --set llm_provider --value openai

egit config --set llm_model --value openai/gpt-4o

egit config --set llm_api_key --value myopenaiapikey

egit config --set llm_api_base --value https://api.openai.com/v1

```



### Anthropic

```bash

egit config --set llm_provider --value anthropic

egit config --set llm_model --value anthropic/claude-3-5-sonnet-20241022

egit config --set llm_api_key --value myanthropicapikey

egit config --set llm_api_base --value https://api.anthropic.com

egit config --set llm_max_tokens --value 8192 # NOTE: 8192 is the max for Anthropic

```



### Gemini

```bash

egit config --set llm_provider --value gemini

egit config --set llm_model --value gemini/gemini-1.5-pro

egit config --set llm_api_key --value mygeminiapikey

```



### Google Vertex AI

Not Supported yet (Coming Soon!)



## Tested Models

These models have been tested with eGit so far, and should work as expected. Most models that are 8b or highershould work (Even 1b and 3b seem to work OK via LM Studio), so feel free to try out whatever you like.



- ollama/llama3.1:8b # Seems to work OK, except for larger contexts (may be an Ollama config issue)

- lm_studio/hermes-3-llama-3.1-8b # Works Great

- openai/gpt-4o # Works Excellent

- anthropic/claude-3-5-sonnet-20241022 # Works Excellent

- gemini/gemini-1.5-pro # Works Excellent



NOTE: We use LiteLLM to manage the LLM API, so we can use any LLM provider that LiteLLM supports generally. Please see [LiteLLM Providers](https://docs.litellm.ai/docs/providers) for more information on supported LLM providers.



## Contributing



Contributions are welcome! Please feel free to submit a Pull Request.



## License



This project is released under the [Apache License](LICENSE).



## Full Documentation



For more information, visit [Documentation](docs/USER_DOCS/README.md).