poker_ai.games.short_deck package¶

Submodules¶

poker_ai.games.short_deck.player module¶

class poker_ai.games.short_deck.player.ShortDeckPokerPlayer(player_i: int, initial_chips: int, pot: poker_ai.poker.pot.Pot)¶

Bases: poker_ai.poker.player.Player

Inherits from Player which will interface easily with the PokerEngine.

This class should manage the state of the players personal pot, and the private cards that are dealt to the player. Also manages whether this player has folded or not.

poker_ai.games.short_deck.state module¶

class poker_ai.games.short_deck.state.ShortDeckPokerState(players: List[poker_ai.games.short_deck.player.ShortDeckPokerPlayer], small_blind: int = 50, big_blind: int = 100, lut_path: str = '.', pickle_dir: bool = False, load_card_lut: bool = True)¶

Bases: object

The state of a Short Deck Poker game at some given point in time.

The class is immutable and new state can be instanciated from once an action is applied via the ShortDeckPokerState.new_state method.

__repr__()¶: Return a helpful description of object in strings and debugger.

_increment_stage()¶: Once betting has finished, increment the stage of the poker game.

_move_to_next_player()¶: Ensure state points to next valid active player.

_reset_betting_round_state()¶: Reset the state related to counting types of actions.

property all_players_have_actioned¶: Return whether all players have made atleast one action.

apply_action(action_str: Optional[str]) → poker_ai.games.short_deck.state.ShortDeckPokerState ¶

Create a new state after applying an action.

Parameters: action_str (str or None) – The description of the action the current player is making. Can be any of {“fold, “call”, “raise”}, the latter two only being possible if the agent hasn’t folded already.
Returns: new_state – A poker state instance that represents the game in the next timestep, after the action has been applied.
Return type: ShortDeckPokerState

property betting_round¶: Betting stagee in integer form.

property betting_stage¶: Return betting stage.

property community_cards¶: Return all shared/public cards.

property current_player¶: Returns a reference to player that makes a move for this state.

property info_set¶: Get the information set for the current player.

property initial_regret¶: Returns the default regret for this state.

property initial_strategy¶: Returns the default strategy for this state.

property is_terminal¶

Returns whether this state is terminal or not.

The state is terminal once all rounds of betting are complete and we are at the show down stage of the game or if all players have folded.

property legal_actions¶: Return the actions that are legal for this game state.

static load_card_lut(lut_path: str = '.', pickle_dir: bool = False) → Dict[str, Dict[Tuple[int, …], str]]¶

Load card information lookup table.

…

Parameters

lut_path (str) – Path to lookupkup table.
pickle_dir (bool) – Whether the lut_path is a path to pickle files or not. Pickle files are deprecated for the lut.

Returns

cad_info_lut – Card information cluster lookup table.

Return type

InfoSetLookupTable

property n_players_started_round¶: Return n_players that started the round.

property payout¶: Return player index to payout number of chips dictionary.

property player_i¶: Get the index of the players turn it is.

property players¶: Returns players in table.

property private_hands¶: Return all private hands.

poker_ai.games.short_deck.state.new_game(n_players: int, card_info_lut: Dict[str, Dict[Tuple[int, …], str]] = {}, **kwargs) → poker_ai.games.short_deck.state.ShortDeckPokerState ¶

Create a new game of short deck poker.

…