Rubik's Cube Solver (CST401 KTU)

This project is a full‑stack AI Rubik's Cube solver designed for the CST401 (Artificial Intelligence) course at KTU. It demonstrates state‑space representation, goal testing, search strategies (BFS, A*, IDA*), and performance metrics.

Tech Stack

• Backend: FastAPI (Python) for solver algorithms and REST API
• Frontend: React + Three.js (react‑three‑fiber) for 3D cube visualization

Run Backend

cd backend
py -m venv .venv
.venv\Scripts\pip install -r requirements.txt
python -m uvicorn main:app --reload --port 8000

Endpoints:

• GET /health – health check
• GET /scramble – sample scrambled state (54 facelets)
• POST /solve – body: { state: string[54], algorithm: "bfs"|"astar"|"ida*" }

Frontend

If Node.js 20.19+ is available:

cd frontend
npm create vite@latest rubiks-ui -- --template react-ts
cd rubiks-ui
npm install three @react-three/fiber @react-three/drei axios zustand tailwindcss postcss autoprefixer
npx tailwindcss init -p
npm run dev

If your Node version is below Vite's requirement, either upgrade Node or use an alternative like CRA or a prebuilt static HTML + CDN Three.js setup.

AI Concepts Mapping

• State space: 54‑length facelet vector in backend/main.py (state: List[str])
• Initial state: user‑provided or /scramble
• Goal test: is_goal(state) checks uniform colors per face
• Actions: quarter turns (U, R, F and inverses); apply_move
• Transition model: apply_move (currently a placeholder for index permutations)
• Search strategies: bfs_solve, astar_solve, ida_star_solve
• Heuristic: heuristic_misplaced_faces (facelet mismatch count)
• Metrics: explored states, depth, elapsed time (ms)

Next Steps

• Implement full cube move permutations for all faces (U, D, L, R, F, B and 2/prime variants)
• Wire the interactive 3D cube to produce/consume the 54‑facelet state
• Visualize the search process and step‑by‑step solution playback

Academic Notes

Code comments highlight AI components for evaluation. The heuristic here is simple and intended for demonstration; you can replace it with domain heuristics (e.g., pattern databases, corner/edge orientation counts) and compare A* vs IDA* empirically using the provided metrics.