RoboHub

RoboHub is a neural network based chess engine designed for over the board play.

Usage

Communication with the engine is performed in console.

Interfaces

Limited UCI (Universall Chess Iterface) implementation is available alongside a custom OTB (Over The Board) interaface, which is much simpler and quicker to use via keyboard.

Selection:

• uci
  selects uci interface
  acknowledged with response "uciok"
• ocb
  selects ocb interface
  acknowledged with response "ocbok"

UCI

• ucinewgame
  resets the engine search tree

• isready
  responds 'readyok' when the engine has finished initialization

• position [ fen | startpos ] moves move1 .... movei
  allows for passing a position in format of fen string or starting a new game

• debug [ on | off ]
  toggles debug mode

• go

  • searchmoves move1 .... movei
    restrict search to specified moves
  • ponder
    start evaluating positions resulting from the move suggested for the opponent
  • wtime x
    white has x msec left on the clock
  • btime x
    black has x msec left on the clock
  • winc x
    white increment per move in mseconds if x > 0
  • binc x
    black increment per move in mseconds if x > 0
  • movestogo x
    tells the engine how many moves are left until bonus time
  • depth x
    search x plies only
  • nodes x
    search x nodes only
  • movetime x
    search exactly x mseconds
  • infinite
    search until the "stop" command

• stop
  stop calculation

• ponderhit
  the user has played the expected move. The engine will continue calculation in normal mode

• quit
  quit the program

OTB

• new
  resets the engine search tree and creates a default board

• reset
  resets the engine search tree

• isready
  responds 'readyok' when the engine has finished initialization

• position [fen | startpos ] moves move1 .... movei
  allows for passing a position in format of fen string or starting a new game

• go

  • movetime
    search exactly x seconds.
    If movetime is not given, the engine will adjust search time accordingly to total time it has left

• setoption name [value x]
  used to change the internal parameters

  • autopush
    automatically push the best move after finishing search
  • autoponder
    automatically enter ponder search after pushing the best move autopush has to be enabled

• push move
  push a move on the board
  e.g. push e2e4

Dependencies

PyTorch
PyTorch Lighning
NumPy
python-chess
PyYAML