RefactoAI is a full-stack web platform and AI assistant designed to help developers master software design patterns, clean code practices, and refactoring skills through various exercises.

The system combines a FastAPI backend for data management and code execution with a React + Vite frontend for an interactive learning experience.
It leverages static analysis, AI feedback, and live code testing to create an environment where developers can continuously improve code quality and reasoning.

• 🧠 AI Code Review Assistant
  Automatically analyzes submitted code using Pylint and an AI model, providing structured, educational feedback and improvement hints.

• 🧩 Interactive Refactoring Tasks
  Each challenge presents a piece of “messy code” and asks users to refactor it based on design patterns (e.g., Factory, Adapter, Strategy).

• ⚙️ Real-Time Code Execution
  Run your Python code in a controlled sandbox environment and receive instant feedback.

• 📚 Topic-Based Roadmap
  Tasks are grouped by programming patterns or principles, allowing learners to progress from beginner to advanced topics.

• 💬 AI Chat Support
  Users can discuss code with an AI assistant trained to explain best practices and clean code concepts.

• FastAPI – High-performance Python web framework
• SQLAlchemy – ORM for database modeling
• SQLite – Lightweight local database
• Pylint – Static analysis tool for Python code

• React + Vite – Fast modern frontend framework
• TailwindCSS – Utility-first CSS styling
• TypeScript – Type-safe codebase
• Monaco Editor – Coding environment

• OpenAI Integration
  The system uses an LLM to interpret Pylint feedback, identify design issues, and provide conceptual explanations.

1. Task Retrieval – The user selects a programming task from a categorized roadmap.
2. Code Submission – The user writes or refactors Python code.
3. Code Check – The system runs:
   • Static analysis (Pylint) to detect code quality issues
   • AI Evaluation using LLM for conceptual and structural guidance
4. Feedback Delivery – The assistant returns structured JSON feedback containing:
   • A summary of weaknesses
   • Hints with explanations
   • A numerical score for code clarity and structure

cd server
python3 -m venv .venv
source .venv/bin/activate
pip install fastapi uvicorn sqlalchemy python-dotenv pylint
python scripts/fill_db.py
uvicorn main:app --reload

cd client
npm install
npm run dev