AI-Powered Codebase Refactoring Agent (Gemini)

An intelligent codebase analysis and refactoring system that helps improve code quality, maintainability, and generates insights about your projects using Google's Gemini AI.

Features

• 🔍 Codebase Analysis: Analyzes code structure, complexity, and quality metrics
• 🤖 AI-Powered Insights: Uses Google Gemini AI for intelligent code analysis and suggestions
• ✨ Code Refactoring: Automated code improvements while preserving functionality
• 🌐 Multi-language Support: Python, JavaScript, TypeScript, Java, C/C++, Go, Rust, and more
• 🗺️ Repository Mapping: Complete file tree analysis with metadata
• 🎙️ Interview Questions: Generates technical interview questions based on code analysis
• 📊 Comprehensive Reports: Detailed recommendations for codebase improvements
• ⚡ Rate Limit Management: Built-in delays and retry logic for API stability

Supported Languages

• Python (.py)
• JavaScript/TypeScript (.js, .jsx, .ts, .tsx)
• Java (.java)
• C/C++ (.c, .cpp, .h, .hpp)
• C# (.cs)
• Go (.go)
• Rust (.rs)
• PHP (.php)
• Ruby (.rb)
• Swift (.swift)
• Kotlin (.kt)
• Scala (.scala)
• R (.r)
• HTML/CSS (.html, .css)
• Shell scripts (.sh)
• SQL (.sql)
• And many more...

Installation

Quick Setup

1. Clone the repository:

   git clone <repository-url>
   cd code-refactoring-agent

2. Run the setup script:

   python setup.py

3. Get your Gemini API key:

   • Visit Google AI Studio
   • Create a new API key
   • Add it to your .env file:

     GEMINI_API_KEY=your_api_key_here
     

Manual Installation

1. Install dependencies:

   pip install -r requirements.txt

2. Set up environment variables:

   cp .env.example .env
   # Edit .env and add your GEMINI_API_KEY

Usage

Basic Usage

Analyze a local codebase:

python main.py /path/to/your/project

Analyze a GitHub repository:

python main.py https://github.com/username/repository.git

Specify custom output directory:

python main.py /path/to/project --output-dir my_refactored_code

Model Selection

Choose between Gemini models based on your needs:

Test your API key and find working models:

python test_gemini_models.py

Fast processing (recommended for most cases):

python main.py /path/to/project --model gemini-1.5-flash-latest

Higher quality analysis and refactoring:

python main.py /path/to/project --model gemini-1.5-pro-latest

Rate Limit Management

Adjust delay between API calls:

python main.py /path/to/project --delay 5  # 5 seconds between calls

Process only specific phases:

python main.py /path/to/project --skip-refactoring  # Analysis only
python main.py /path/to/project --skip-analysis     # Refactoring only

Advanced Options

Get help:

python main.py --help

Complete example with all options:

python main.py https://github.com/user/repo.git \
  --output-dir custom_output \
  --model gemini-1.5-pro-latest \
  --delay 3

Output Structure

The tool generates several outputs in your specified directory:

refactored_codebase/
├── 📁 src/                          # Refactored source code
│   ├── 🐍 main.py                   # Improved Python files
│   └── 🟨 utils.js                  # Refactored JavaScript
├── 📝 CODEBASE_RECOMMENDATIONS.md   # Overall improvement suggestions
└── 📝 INTERVIEW_QUESTIONS.md        # Generated interview questions

How It Works

1. Analysis Phase:

   • Scans your codebase for supported file types
   • Analyzes each file for complexity, maintainability, and code smells
   • Generates detailed quality reports using Gemini AI

2. Refactoring Phase:

   • Applies AI-powered improvements to your code
   • Maintains original functionality while improving readability
   • Adds comments and documentation where appropriate

3. Insights Generation:

   • Creates overall codebase recommendations
   • Generates technical interview questions
   • Provides actionable improvement suggestions

Configuration

Environment Variables

You can customize the behavior using environment variables in your .env file:

# Required
GEMINI_API_KEY=your_api_key_here

# Optional model configuration
GEMINI_MODEL=gemini-1.5-flash-latest     # or gemini-1.5-pro-latest
GEMINI_TEMPERATURE=0.2                 # Creativity level (0.0-1.0)
GEMINI_MAX_RETRIES=3                   # API retry attempts
GEMINI_REQUEST_TIMEOUT=30              # Request timeout in seconds

# File processing limits
MAX_FILE_SIZE_FOR_REFACTORING=15000    # Max characters for refactoring
MAX_FILE_SIZE_FOR_ANALYSIS=12000       # Max characters for analysis

Gemini Models Comparison

Model	Speed	Quality	Use Case
gemini-1.5-flash-latest	⚡ Fast	Good	General analysis, large codebases
gemini-1.5-pro-latest	🐌 Slower	Excellent	Complex refactoring, detailed analysis
gemini-pro	⚡ Fast	Good	Legacy model, stable

Ignored Patterns

The tool automatically ignores common directories and files:

• Version control: .git, .svn
• Dependencies: node_modules, venv, __pycache__
• Build outputs: dist, build, target
• IDE files: .idea, .vscode

Troubleshooting

Common Issues

"GEMINI_API_KEY not found":

• Ensure your .env file exists and contains your API key
• Get your API key from Google AI Studio

"Rate limit exceeded" or "Quota exceeded":

• Increase the --delay parameter (try --delay 5 or higher)
• Use --skip-analysis or --skip-refactoring to reduce API calls
• Check your Gemini API quota in Google AI Studio

Model not found (404 error):

• Run python test_gemini_models.py to check available models
• Use gemini-1.5-flash-latest or gemini-pro instead
• Your API key might not have access to newer models "Git command not found":
• Install Git: https://git-scm.com/downloads
• Ensure Git is in your system PATH

Large files being skipped:

• Files over 15KB are copied without refactoring
• Files over 12KB are truncated for analysis
• This prevents API token limit issues

Performance Tips

1. For large codebases:

   • Start with --skip-refactoring to get analysis results quickly
   • Use gemini-1.5-flash-latest for faster processing
   • Run python test_gemini_models.py first to find working models
   • Increase --delay to avoid rate limits

2. For detailed analysis:

   • Use gemini-1.5-pro-latest for better quality
   • Process smaller codebases or specific directories
   • Use --skip-analysis if you only need refactored code

3. Rate limit management:

   • The tool automatically adds delays between requests
   • Implements exponential backoff for retries
   • Monitors for rate limit responses

API Limits and Pricing

• Gemini 1.5 Flash: Fast and efficient, generous free tier
• Gemini 1.5 Pro: More capable, limited free requests per minute
• Check current limits at: Google AI Studio

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

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

Acknowledgments

• Built with Google Gemini AI for powerful code analysis
• Uses the latest Gemini 1.5 models for optimal performance
• Designed with rate limiting and error handling for production use

Roadmap

• Support for more programming languages
• Custom refactoring rules and templates
• Integration with popular IDEs
• Batch processing for multiple repositories
• Web-based interface
• Custom analysis rules
• Team collaboration features

Support

For issues, feature requests, or questions:

1. Check the Issues page
2. Create a new issue with detailed information
3. Include error messages and your configuration

Common support scenarios:

• Rate limiting issues → Adjust --delay parameter
• Large codebases → Use --skip-analysis or --skip-refactoring
• API errors → Check your Gemini API key and quota
• Installation issues → Run python setup.py again

Version History

• v2.0.0: Migrated to Google Gemini AI with improved rate limiting
• v1.0.0: Initial release with Groq API support