Human and computer Go players each have their weaknesses and blind spots, and until now, they have been pitted against each other. The GoFamiliar is software designed to instead work alongside human Go players to find better moves together. The GoFamiliar is built on the contention that the combination of human abilities and computer algorithms, through cunning information delivery, is better than either is alone.
The main use of it will be in a style of play called Hydra, Advanced or Free Style, where a "player" can be any combination of humans and computers.
No use cases completed.
See Project Documents/Objectives and Principles.md
Incomplete, and I will be seeking advice or mentoring on how to create it as an installable application.
No API at present
Tests written using pytest for most testing, with some doctests used to document simple use cases for functions and method.
Just me at the moment. I am very open to contributors, but it is a side of development I am inexperienced with. Suggestions, advice and/or mentoring would be welcomed.
I studied the code of the michi project extensively prior to beginning this project. It helped greatly in understanding MCTS and how to implement Go in Python.
https://github.com/pasky/michi
My fork of this project shows my attempt to understand MCTS and pasky's implementation of it by refactoring, and adding tests to michi.py as well as trying to find speed ups.
https://github.com/salvor7/michi
None at present