Blocku Doku Bot is a sophisticated agent designed to play the BlockuDoku game by analyzing screenshots and determining the best set of moves using heuristics developed through genetic algorithm training. This bot utilizes a combination of board recognition algorithms and a Python server-client setup to process game states and compute optimal strategies.
- Board Recognition Algorithm: Detects occupied cells on the game board.
- Block Detection: Identifies blocks based on their indices from the game screenshot.
- Optimal Move Calculation: Uses a backtracking algorithm to compute the best possible set of moves.
- Heuristic Training: Currently leveraging features and trained constants from another project, with plans to develop proprietary heuristics.
To use the Blocku Doku Bot, follow these steps:
- Ensure you have Java installed on your machine.
- Download the
BlockuDokuBot.jarfile from the repository. - Place your screenshot (1080x2400 PNG) in the
Server/Server_Imgdirectory.
Run the bot with the following command:
java .\Blocku-Doku-Bot.jar .\Server\Server_Img\{test image}.pngReplace {img file path} with the path to your game screenshot.
You can run this command with preset images Here is an example of such a use:
java -jar .\Blocku-Doku-Bot.jar .\Server\Server_Img\20240424235322.pngFor the Python server-client:
- Flask
- os
- subprocess
- datetime
You can install the necessary Python libraries using:
pip install FlaskContributions are welcome! Please feel free to fork the repository, make changes, and submit a pull request. If you plan to propose significant changes, please open an issue first to discuss your ideas.
- GaBlockSudoku: This project was instrumental in providing the initial set of features and training models for my genetic algorithm.
This project is released under the MIT License.