The Automated Code Correction Framework is an AI-powered debugging tool designed to automate error detection and correction. It streamlines the debugging process by analyzing errors, providing relevant solutions, and applying corrections to faulty code. Implemented in Python, the system leverages LangChain for AI-driven automation.
- Logger: Captures and stores errors from executed programs in a log file.
- AI Agent 1: Extracts errors, explains their causes, and provides relevant examples from sources like Stack Overflow in a GUI.
- AI Agent 2: Extracts faulty code, sends it to a Large Language Model (LLM) for correction, and replaces the erroneous lines with the fixed code.
- Automated Debugging: Reduces manual effort and enhances development efficiency.
- GUI: Shows the error analysis in a GUI.
- Python
- LangChain (for AI agents)
- Large Language Model (LLM) (for code correction)
- Python Tkinter (for displaying error analysis and examples)
- Clone the repository:
git clone https://github.com/Deekshith-poojary/Automated-Code-Correction.git cd Automated-Code-Correction - Install dependencies:
pip install -r requirements.txt
- Set up your LLM API keys in
agentone.pyandagenttwo.py:- The project uses Google Gemini AI API Key by default.
- Users can also configure OpenAI or Claude API keys as alternatives.
- Simply run your error-prone program file by importing the
global_loggermodule:Example usage:import global_logger
import global_logger print(3/0)#it will through division by zero error.
- The framework will automatically capture errors, analyze them, and provide corrected solutions.
- View debugging insights and corrected code in the GUI.
Feel free to contribute to the project by submitting pull requests. Ensure your code follows best practices and includes proper documentation.
This project is licensed under the MIT License.
For queries or contributions, reach out to deekshithpoojary122d@gmail.com.