A Rust-based tool that analyzes Git commit histories and rates each committer's coding style based on performance and maintainability metrics using the Deepseek AI API.
Just a fun project to re-learn Rust and use Deepseek-r1, but afaik for something like this it doesn't really matter which model you use.
- Analyzes both local and public Git repositories
- Generates performance and maintainability scores for each committer
- Processes entire commit history
- Provides detailed per-committer statistics
- Rust (latest stable version)
- Git
- Deepseek API key
- Clone this repository:
git clone https://github.com/t3nsed/feb.git
cd feb- Build the project:
cargo build --releaseThe compiled binary will be available at target/release/feb
Set your Deepseek API key using one of these methods:
- Environment variable:
export DEEPSEEK_API_KEY=your_api_key_here- Command-line argument:
feb /path/to/repo --api-key your_api_key_hereNote: Command-line argument takes precedence over environment variable.
Basic usage:
feb /path/to/repositoryThe program will:
- Analyze each commit in the repository's history
- Extract code changes and commit messages
- Send the data to Deepseek API for analysis
- Calculate and display average scores for each committer
The program outputs statistics for each committer in the following format:
Committer: John Doe <[email protected]>
Performance Score: 8.45
Maintainability Score: 7.92
Committer: Jane Smith <[email protected]>
Performance Score: 9.12
Maintainability Score: 8.78
Scores are on a scale of 0-10, where:
- Performance Score: Indicates code efficiency and optimization
- Maintainability Score: Indicates code readability and maintainability
MIT