Source code for pettingzoo.classic.rlcard_envs.leduc_holdem

# noqa: D212, D415
"""
# Leduc Hold'em

```{figure} classic_leduc_holdem.gif
:width: 140px
:name: leduc_holdem
```

This environment is part of the <a href='..'>classic environments</a>. Please read that page first for general information.

| Import             | `from pettingzoo.classic import leduc_holdem_v4` |
|--------------------|--------------------------------------------------|
| Actions            |                                                  |
| Parallel API       | Yes                                              |
| Manual Control     | No                                               |
| Agents             | `agents= ['player_0', 'player_1']`               |
| Agents             | 2                                                |
| Action Shape       | Discrete(4)                                      |
| Action Values      | Discrete(4)                                      |
| Observation Shape  | (36,)                                            |
| Observation Values | [0, 1]                                           |


Leduc Hold'em is a variation of Limit Texas Hold'em with fixed number of 2 players, 2 rounds and a deck of six cards (Jack, Queen, and King in 2 suits). At the beginning of the game, each player receives one card and, after betting, one public card is revealed.
Another round follows. At the end, the player with the best hand wins and receives a reward (+1) and the loser receives -1. At any time, any player can fold.

Our implementation wraps [RLCard](http://rlcard.org/games.html#leduc-hold-em) and you can refer to its documentation for additional details. Please cite their work if you use this game in research.

### Observation Space

The observation is a dictionary which contains an `'observation'` element which is the usual RL observation described below, and an  `'action_mask'` which holds the legal moves, described in the Legal Actions Mask section.

As described by [RLCard](https://github.com/datamllab/rlcard/blob/master/docs/games#leduc-holdem), the first 3 entries of the main observation space correspond to the player's hand (J, Q, and K) and the next 3 represent the public cards. Indexes 6 to 19 and 20 to 33 encode the number of chips by
the current player and the opponent, respectively.

|  Index  | Description                                                                  |
|:-------:|------------------------------------------------------------------------------|
|  0 - 2  | Current Player's Hand<br>_`0`: J, `1`: Q, `2`: K_                            |
|  3 - 5  | Community Cards<br>_`3`: J, `4`: Q, `5`: K_                                  |
|  6 - 20 | Current Player's Chips<br>_`6`: 0 chips, `7`: 1 chip, ..., `20`: 14 chips_   |
| 21 - 35 | Opponent's Chips<br>_`21`: 0 chips, `22`: 1 chip, ..., `35`: 14 chips_       |


#### Legal Actions Mask

The legal moves available to the current agent are found in the `action_mask` element of the dictionary observation. The `action_mask` is a binary vector where each index of the vector represents whether the action is legal or not. The `action_mask` will be all zeros for any agent except the one
whose turn it is. Taking an illegal move ends the game with a reward of -1 for the illegally moving agent and a reward of 0 for all other agents.

### Action Space

| Action ID | Action |
|:---------:|--------|
|     0     | Call   |
|     1     | Raise  |
|     2     | Fold   |
|     3     | Check  |

### Rewards

|      Winner       |       Loser       |
| :---------------: | :---------------: |
| +raised chips / 2 | -raised chips / 2 |


### Version History

* v4: Upgrade to RLCard 1.0.3 (1.11.0)
* v3: Fixed bug in arbitrary calls to observe() (1.8.0)
* v2: Bumped RLCard version, bug fixes, legal action mask in observation replaced illegal move list in infos (1.5.0)
* v1: Bumped RLCard version, fixed observation space, adopted new agent iteration scheme where all agents are iterated over after they are done (1.4.0)
* v0: Initial versions release (1.0.0)

"""
from __future__ import annotations

import os

import gymnasium
import numpy as np
import pygame
from gymnasium.utils import EzPickle

from pettingzoo.classic.rlcard_envs.rlcard_base import RLCardBase
from pettingzoo.utils import wrappers


def get_image(path):
    from os import path as os_path

    cwd = os_path.dirname(__file__)
    image = pygame.image.load(cwd + "/" + path)
    return image


def get_font(path, size):
    from os import path as os_path

    cwd = os_path.dirname(__file__)
    font = pygame.font.Font((cwd + "/" + path), size)
    return font


[docs] def env(**kwargs): env = raw_env(**kwargs) env = wrappers.TerminateIllegalWrapper(env, illegal_reward=-1) env = wrappers.AssertOutOfBoundsWrapper(env) env = wrappers.OrderEnforcingWrapper(env) return env
[docs] class raw_env(RLCardBase, EzPickle): metadata = { "render_modes": ["human", "rgb_array"], "name": "leduc_holdem_v4", "is_parallelizable": False, "render_fps": 1, } def __init__( self, render_mode: str | None = None, screen_height: int | None = 1000, ): EzPickle.__init__(self, render_mode, screen_height) super().__init__("leduc-holdem", 2, (36,)) self.render_mode = render_mode self.screen_height = screen_height if self.render_mode == "human": self.clock = pygame.time.Clock()
[docs] def step(self, action): super().step(action) if self.render_mode == "human": self.render()
[docs] def render(self): if self.render_mode is None: gymnasium.logger.warn( "You are calling render method without specifying any render mode." ) return def calculate_width(self, screen_width, i): return int( ( screen_width / (np.ceil(len(self.possible_agents) / 2) + 1) * np.ceil((i + 1) / 2) ) + (tile_size * 31 / 616) ) def calculate_offset(tile_size): return int(tile_size * 23 / 28) # - ((j) * (tile_size * 23 / 28)) def calculate_height(screen_height, divisor, multiplier, tile_size, offset): return int(multiplier * screen_height / divisor + tile_size * offset) screen_height = self.screen_height screen_width = int( screen_height * (1 / 20) + np.ceil(len(self.possible_agents) / 2) * (screen_height * 1 / 2) ) if self.screen is None: pygame.init() if self.render_mode == "human": self.screen = pygame.display.set_mode((screen_width, screen_height)) pygame.display.set_caption("Leduc Hold'em") else: self.screen = pygame.Surface((screen_width, screen_height)) # Setup dimensions for card size and setup for colors tile_size = screen_height * 2 / 10 bg_color = (7, 99, 36) white = (255, 255, 255) self.screen.fill(bg_color) chips = { 0: {"value": 10000, "img": "ChipOrange.png", "number": 0}, 1: {"value": 5000, "img": "ChipPink.png", "number": 0}, 2: {"value": 1000, "img": "ChipYellow.png", "number": 0}, 3: {"value": 100, "img": "ChipBlack.png", "number": 0}, 4: {"value": 50, "img": "ChipBlue.png", "number": 0}, 5: {"value": 25, "img": "ChipGreen.png", "number": 0}, 6: {"value": 10, "img": "ChipLightBlue.png", "number": 0}, 7: {"value": 5, "img": "ChipRed.png", "number": 0}, 8: {"value": 1, "img": "ChipWhite.png", "number": 0}, } # Load and blit all images for each card in each player's hand for i, player in enumerate(self.possible_agents): state = self.env.game.get_state(self._name_to_int(player)) # Load specified card # Each player holds only one card. Unlike Texas Hold'em, state['hand'] = str, and not a list card = state["hand"] card_img = get_image(os.path.join("img", card + ".png")) card_img = pygame.transform.scale( card_img, (int(tile_size * (142 / 197)), int(tile_size)) ) # Players with even id go above public cards if i % 2 == 0: self.screen.blit( card_img, ( ( calculate_width(self, screen_width, i) - calculate_offset(tile_size) ), calculate_height(screen_height, 4, 1, tile_size, -1), ), ) # Players with odd id go below public cards else: self.screen.blit( card_img, ( ( calculate_width(self, screen_width, i) - calculate_offset(tile_size) ), calculate_height(screen_height, 4, 3, tile_size, 0), ), ) # Load and blit text for player name font = get_font(os.path.join("font", "Minecraft.ttf"), 36) text = font.render("Player " + str(i + 1), True, white) textRect = text.get_rect() if i % 2 == 0: textRect.center = ( ( screen_width / (np.ceil(len(self.possible_agents) / 2) + 1) * np.ceil((i + 1) / 2) - tile_size * (4 / 10) ), calculate_height(screen_height, 4, 1, tile_size, -(22 / 20)), ) else: textRect.center = ( ( screen_width / (np.ceil(len(self.possible_agents) / 2) + 1) * np.ceil((i + 1) / 2) - tile_size * (4 / 10) ), calculate_height(screen_height, 4, 3, tile_size, (23 / 20)), ) self.screen.blit(text, textRect) # Load and blit number of poker chips for each player font = get_font(os.path.join("font", "Minecraft.ttf"), 24) text = font.render(str(state["my_chips"]), True, white) textRect = text.get_rect() # Calculate number of each chip total = state["my_chips"] height = 0 for key in chips: num = total / chips[key]["value"] chips[key]["number"] = int(num) total %= chips[key]["value"] chip_img = get_image(os.path.join("img", chips[key]["img"])) chip_img = pygame.transform.scale( chip_img, (int(tile_size / 2), int(tile_size * 16 / 45)) ) # Blit poker chip img for j in range(0, int(chips[key]["number"])): if i % 2 == 0: self.screen.blit( chip_img, ( ( calculate_width(self, screen_width, i) + tile_size * (2 / 10) ), calculate_height(screen_height, 4, 1, tile_size, -1 / 2) - ((j + height) * tile_size / 15), ), ) else: self.screen.blit( chip_img, ( ( calculate_width(self, screen_width, i) + tile_size * (2 / 10) ), calculate_height(screen_height, 4, 3, tile_size, 1 / 2) - ((j + height) * tile_size / 15), ), ) height += chips[key]["number"] # Blit text number if i % 2 == 0: textRect.center = ( (calculate_width(self, screen_width, i) + tile_size * (9 / 20)), calculate_height(screen_height, 4, 1, tile_size, -1 / 2) - ((height + 1) * tile_size / 15), ) else: textRect.center = ( (calculate_width(self, screen_width, i) + tile_size * (9 / 20)), calculate_height(screen_height, 4, 3, tile_size, 1 / 2) - ((height + 1) * tile_size / 15), ) self.screen.blit(text, textRect) # Load and blit public cards if state["public_card"] is not None: card = state["public_card"] card_img = get_image(os.path.join("img", card + ".png")) card_img = pygame.transform.scale( card_img, (int(tile_size * (142 / 197)), int(tile_size)) ) self.screen.blit( card_img, ( ( ( ((screen_width / 2) + (tile_size * 31 / 616)) - calculate_offset(tile_size) + (tile_size / 2) ), calculate_height(screen_height, 2, 1, tile_size, -(1 / 2)), ) ), ) if self.render_mode == "human": pygame.display.update() self.clock.tick(self.metadata["render_fps"]) observation = np.array(pygame.surfarray.pixels3d(self.screen)) return ( np.transpose(observation, axes=(1, 0, 2)) if self.render_mode == "rgb_array" else None )