Gemini-Code-Fixer

An automated Python script fixer powered by Google's Gemini AI model. This tool automatically detects, analyzes, and fixes errors in Python scripts using AI-powered suggestions.

Features

• Automatic error detection and stack trace analysis
• AI-powered error fixing suggestions
• Handles multiple file dependencies
• Iterative fixing attempts with maximum retry limit
• Detailed execution logging
• Support for common Python libraries (pandas, matplotlib, numpy)

Prerequisites

• Python 3.7+
• Google Cloud Project
• Required Python packages (see requirements.txt)

Installation

1. Clone the repository:

git clone https://github.com/cnemri/gemini-code-fixer.git
cd gemini-code-fixer

2. Install required packages:

pip install -r requirements.txt

3. Set up your Google Cloud Project ID in main.py:

project_id = "your-project-id"  # Replace with your GCP project ID

Usage

1. Place your Python script to be fixed in the same directory as main.py and name it script.py

2. Run the fixer:

python main.py

The tool will:

1. Execute your script
2. If errors occur, analyze the stack trace
3. Generate and apply fixes
4. Repeat until the script runs successfully or reaches maximum iterations
5. Log all actions in full_execution_log.json

Project Structure

• main.py - Main execution script
• script_executor.py - Handles Python script execution
• traceback_parser.py - Parses error stack traces using Gemini AI
• file_handler.py - Manages file operations
• fix_suggester.py - Generates fix suggestions using Gemini AI
• examples/ - Example Python scripts demonstrating common errors
• requirements.txt - Required Python packages

Example Scripts

The repository includes example scripts demonstrating common Python errors:

1. 01-pandas.py - Pandas DataFrame operations with typos
2. 02-matplotlib.py - Matplotlib plotting with incorrect method names
3. 03-numpy.py - NumPy array operations with dimension mismatches
4. 04-combined.py - Combined usage of multiple libraries with syntax errors

How It Works

1. Script Execution: Runs the target Python script and captures any errors
2. Error Analysis: Uses Gemini AI to parse stack traces and identify relevant files
3. Context Collection: Gathers contents of all related Python files
4. Fix Generation: Sends error context to Gemini AI for fix suggestions
5. Fix Application: Applies suggested fixes to the original script
6. Verification: Re-runs the script to verify the fix
7. Logging: Records all actions and responses in a JSON log file

Configuration

• Maximum fix attempts: 10 (configurable in main.py)
• Default Gemini model: "gemini-1.5-flash-002" (configurable)
• Log file: full_execution_log.json

Limitations

• Relies on Gemini knowledge. Google search grounding to be implemented.

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

License

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

Support

For issues and feature requests, please use the GitHub Issues section of the repository.