nathanaelbosch / GO_DILab

A machine learning how to play GO - DataInnovation-Lab - WS20172018

A machine learning how to play GO

🎲 TUM Data Innovation Lab WS2017/18, Team Hikaru

Setup

• install Python 3.6 if not already present
• create a virtual environment in the repository: python -m venv .venv (or python3, check python --version)
• activate it: source activate .venv on macOS, source .venv/bin/activate on Linux or .\.venv\Scripts\activate on Windows
• run pip install -r requirements.txt or let your IDE help you with installing the required modules

Usage

To start up a controller with two engines run src/play/run.py either via a run configuration in your IDE or via command line. In run.py you can configure the type of players.

To start just one engine, run src/play/controller/GTPengine.py and use GTP commands to communicate with it.

To build an executable engine (to attach it to a Go GUI or run it by double-clicking), install pyinstaller (pip install pyinstaller) and run pyinstaller --onefile src/play/controller/GTPengine.py. On Windows this will create an .exe in a new dist/ folder.