hyphi_gym.common.board

Module Contents

Classes

Board

Base for grid-based games managing

Attributes

CELLS
CHARS
ACTIONS

hyphi_gym.common.board.CELLS

hyphi_gym.common.board.CHARS

hyphi_gym.common.board.ACTIONS

class hyphi_gym.common.board.Board(size: tuple[int, int], layout: list[str] | None, bound=None, **kwargs)

Bases: hyphi_gym.common.base.Base

Base for grid-based games managing • A layout of a variable size • Containing CELLS ∈ [WALL, FIELD, AGENT, TARGET, HOLE] • Navigable with ACTIONS ∈ [UP, RIGHT, DOWN,LEFT]

board: numpy.ndarray

size: tuple[int, int]

layout: numpy.ndarray | None

ascii(grid: numpy.ndarray | None = None) → list[str]

Transform 2D-INT Array to list of strings

_grid(ascii: list[str]) → numpy.ndarray

Transform 1D string to 2D-INT Array

getpos(board: numpy.ndarray | None = None, cell: str = AGENT) → numpy.ndarray

Position helper for finding the vector position of cell on the board or internal board

newpos(position: numpy.ndarray | tuple[int, int], action: int, n=1) → tuple[int, int]

Action helper mutating a position tuple by appying action n-times

iterate_actions(p: tuple[int, int], n=1, condition=lambda act, pos, n: ...) → dict[int, tuple[int, int]]

Return possible n actions in a bounded box given a position p and their mutated positions

action_possible(act: int, pos: tuple[int, int], n=1) → bool

Return possible n actions act in a bounded box given a position pos

_generate()

Random generator for flat grids, creating a grid of size

_validate(board, error=True, setup=False)

Overwrite this function to validate layout return steps to solution

_update(key: str, oldpos, newpos)

_randomize(board: numpy.ndarray, key: str)

Mutation function to randomize the position of cell on board

reset(**kwargs) → tuple[gymnasium.spaces.Space, dict]

Gymnasium compliant function to reset the environment