Skip to content

Latest commit

 

History

History
64 lines (44 loc) · 2.58 KB

README.md

File metadata and controls

64 lines (44 loc) · 2.58 KB
Raw

5dchess-tools

Programming set of tools to analyze 5D Chess games.

Note: you are currently seeing an old, unmaintained version of this library. The library is currently being rewritten in its entirety as part of the 0.2 update. You can see and contribute to its progress over on the v2 branch!

Installation

Clone this repository:

git clone https://github.com/adri326/5dchess-tools/
cd 5dchess-tools

Then build or run it using cargo:

cargo run path/to/game.json

The current, included executable will read a JSON file (outputted by this parser) and proceed to run calculations on it.

As a dependency

Add the following to your Cargo.toml:

[dependencies.chess5dtools]
version = "0.1"
git = "https://github.com/adri326/5dchess-tools"

You can then import the different modules in your code, for instance:

use chess5dlib::game::*;

Usage

The library half of this tool is labelled as chess5dlib (the executable and package chess5dtools).

  • The various structures making a game's state can be found in chess5dlib::game (/lib/game.rs).
  • Per-board move-related logic can be found in chess5dlib::moves (/lib/moves.rs).
  • Moveset-related logic can be found in chess5dlib::moveset (/lib/moveset.rs). Note that as I am writing this, these functions are heavily oriented towards a branch factor-limited, tree-based analysis.
  • Board scoring logic can be found in chess5dlib::resolve (/lib/resolve.rs, might be renamed later)
  • αβ-pruned search and other tree-based search algorithms can be found in chess5dlib::tree

Notes

This game can reach very complex states (multi-dimensional series of checks, having to create several timelines in a specific order, etc.). Soundness over checkmate proof is complicated to achieve and has been found to sometimes be extremely computationally expensive.

It could be very expensive to list out all of the possible movesets (sets of moves per turn), thus the algorithms here are based around a lazy method of generating these movesets. This method relies on a two-pass analysis of the moves:

  • moves are listed for every board
  • for every board, moves are scored and illegal moves are pruned (first pass),i n the same way that the game shows you a red indicator when you try out these moves in the original game
  • moves are lazily combined, based on their ordering made in the first pass
  • movesets are scored and illegal movesets are pruned (second pass)