ChessAI Bot

A multi‐mode chess AI project featuring:

• Easy mode: a Deep Q-Network (DQN) agent trained via OpenAI Gym & TensorFlow
• Medium/Hard modes: classical minimax search with alpha-beta pruning and a phase-aware static evaluation
• Interactive Pygame interface for human vs. AI play

---

Demo

demo.mp4

---

Features

• Custom Gym environment (ChessEnv.py) for RL training
• Fast static evaluation tuned for opening development, middlegame tactics and endgame conversion
• Adaptive search depth based on material count (quicker in opening, deeper in endgame)
• Pygame GUI with piece sprites, menus & animations
• Clear project structure and comprehensive documentation
• Bots in three levels, with the hardest being at 1400 elo

---

Installation

1. Clone the repo

   git clone https://github.com/AlirezaPrz/ChessAI.git

2. Create & activate a virtual environment

   python3 -m venv venv
   source venv/bin/activate       # macOS/Linux
   venv\Scripts\activate          # Windows

3. Install dependencies

   pip install -r requirements.txt

---

Usage

1. Train the (RL) Bots

cd Chessbot
python src/train_model.py

This will spin up the Gym environment, train the DQN agent, and save model weights under weights/.

2. Play against the AI

cd Chessbot
python src/chess_main.py

• On launch, choose Easy, Medium, or Hard mode.
• Use mouse clicks to move pieces.
• Z to undo, R to restart, Q to quit to menu.

---

Project Layout

Chess/                      
├── images/                 # Piece sprites & menu graphics  
│   ├── set1/  
│   ├── set2/  
│   └── start/  
├── src/                    # Source code  
│   ├── weights/            # Trained DQN model files
│   ├── button.py           # UI button class  
│   ├── chess_engine.py      # Core engine & move generation  
│   ├── chess_env.py         # Gym environment wrapper for RL  
│   ├── minimax_bot.py       # Minimax + evaluation + search  
│   ├── train_model.py       # RL training script for DQN agent  
│   ├── chess_main.py        # Pygame entry-point & menus  
│   └── help.py             # Misc utility functions  
├── requirements.txt        # `pip install` dependency list  
├── CONTRIBUTING.md         # Contribution guidelines  
├── LICENSE                 # MIT License  
└── README.md               # This file  

---

Contributing

Please see CONTRIBUTING.md for:

1. Forking & cloning
2. Branch naming conventions
3. Coding style checks & pull request process

---

License

This project is licensed under the MIT License. See LICENSE for details.

---

Happy coding & enjoy your games!