HCheckers is a relatively simple implementation of checkers board game (also known as "draughts"). The core is written in Haskell, and the GUI is written in Python + Qt5. Some day, probably, there will be a web-based JS client to play from browser.
It is possible to play:
- Human vs computer (either user or computer plays white);
- Human vs human.
HCheckers is not about to compete with well-known and highly optimized commercial checkers software. It will hardly do any good in playing versus human checkers grossmaster. HCheckers does not contain any pre-populated openings or endgames database, and it actually does not know how to play checkers - it only knows the rules.
But this is not either a "look-what-i-did-in-one-evening" project. HCheckers has full support of several rules, has decent UI, supports UI localization, and plays not so bad for amateur.
- Fun of development. For not-so-seasoned Haskell programmers, or people who are not-so-expirienced in writing games, this can show some examples. For that, I'm not going to optimize every possible bit: code readability is in priority. For that, HCheckers is not going to be faster than software written in C++ with economy of every bit of memory.
- Fun of game. HCheckers can play well enough for not-so-seasoned draughtsmen.
The code is general enough to implement a wide range of checkers variants. The following are implemented at the moment:
- Russian
- Simple russian (russian draughts without kings)
- Diagonal russian (russian draughts with different initial setup)
- Spancirety (russian draughts on 8x10 board)
- English (checkers)
- Czech draughts (8x8)
- International draughts (10x10)
- Killer draughts (10x10)
- Frisian draughts (10x10)
- Brazilian (rules of international draughts on 8x8 board)
- Canadian draughts (12x12)
- Turkish draughts (orthogonal)
- Armenian draughts (Tama)
It is possible to implement different AI algorithms; currently there is only one, based on standard alpha-beta pruning. The algorithm has some number of parameters, which can be tuned to choose between better play and performance.
At the moment, HCheckers has most of core functionality implemented, but there are some outstanding issues (please refer to github's issue tracker). Most wanted planned things to do are:
- User documentation (#22)
- Code documentation (#1)
- Spectators support (#9)
For the server part, there are two options available.
Last package that I released is available in github releases. Please refer to https://github.com/portnov/hcheckers/releases .
To build a newer package,
$ git clone https://github.com/portnov/hcheckers.git
$ cd hcheckers/docker/
$ ./build-ubuntu-package.sh
The package will be available under docker/target subdirectory.
Use sudo dpkg -i hcheckersd_0.1.0.0-1_amd64.deb to install it.
For non-debian based systems, the only "easily distributed" form for now is the docker container.
$ git clone https://github.com/portnov/hcheckers.git
$ cd hcheckers/docker/
$ ./build-plain-builder.sh
$ ./build-plain.sh
$ ./run-plain.sh
Python client can be installed in two ways:
- Via
pip:
$ cd hcheckers/python/
$ sudo pip3 install .
- Using debian package (on debian-based systems).
Last package that I released is available in github releases. Please refer to https://github.com/portnov/hcheckers/releases .
To build a newer package, execute
$ sudo apt-get install python3-stdeb
$ cd hcheckers/python/
$ ./build_deb.sh
To install a package, do
$ sudo apt install python3-pyqt5 python3-pyqt5.qtsvg python3-pyqt5.qtmultimedia
$ sudo dpkg -i deb_dist/python3-hcheckers_0.1.0.0-1_all.deb
After client is installed (either via pip or deb package), you can run it with
$ hcheckersc.py
You can run HCheckers without actually installing it; it is mostly useful while developing it.
$ sudo apt-get install stack
$ cd hcheckers/
$ stack build
Run server:
$ stack exec hcheckersd
Run client:
$ cd python/
$ python hcheckersc.py