Random players playing bird cards

In [ ]:

import wingspan_gym
from wingspan_gym._internal import StepResult
import random

import wingspan_gym from wingspan_gym._internal import StepResult import random

In [ ]:

import wingspan_gym.game

import wingspan_gym.game

In [ ]:

VERBOSE = True

VERBOSE = True

In [ ]:

class DummyHeuristic:
    def __init__(self, env: wingspan_gym.game.WingspanEnv):
        self.env = env
        self.tried_to_play_a_card = False
        self.played_a_card = False

    def get_next_action(self):
        action_type = self.env.next_action()
        if not self.played_a_card and str(action_type) == "ChooseAction":
            if not self.tried_to_play_a_card:
                self.tried_to_play_a_card = True
                return 0
            else:
                # Tried to play a card and failed
                # Get food then
                return 1

        # No longer choose action, so this works I think
        if self.tried_to_play_a_card:
            self.played_a_card = True

        self.tried_to_play_a_card = False

        return random.randint(0, self.env.action_space_size())

class DummyHeuristic: def __init__(self, env: wingspan_gym.game.WingspanEnv): self.env = env self.tried_to_play_a_card = False self.played_a_card = False def get_next_action(self): action_type = self.env.next_action() if not self.played_a_card and str(action_type) == "ChooseAction": if not self.tried_to_play_a_card: self.tried_to_play_a_card = True return 0 else: # Tried to play a card and failed # Get food then return 1 # No longer choose action, so this works I think if self.tried_to_play_a_card: self.played_a_card = True self.tried_to_play_a_card = False return random.randint(0, self.env.action_space_size())

In [ ]:

env = wingspan_gym.game.WingspanEnv()

env = wingspan_gym.game.WingspanEnv()

In [ ]:

env.reset()
player = [DummyHeuristic(env) for _ in range(2)]
# observation, info = env.reset()

env.reset() player = [DummyHeuristic(env) for _ in range(2)] # observation, info = env.reset()

In [ ]:

episode_over = False
max_steps = 10000
step_idx = 0
while not episode_over and step_idx < max_steps:
    step_idx += 1
    # action = env.action_space.sample()  # agent policy that uses the observation and info

    action = player[env.cur_player_idx()].get_next_action()
    if VERBOSE:
        print("=========")
        print(env._debug_print_state())
        print(f"Chose action: {action}")
        print(f"Step: {step_idx}/{max_steps}")

    if env.step(action) == StepResult.Terminated:
        print(f"Score: {env.points()}")
        break

episode_over = False max_steps = 10000 step_idx = 0 while not episode_over and step_idx < max_steps: step_idx += 1 # action = env.action_space.sample() # agent policy that uses the observation and info action = player[env.cur_player_idx()].get_next_action() if VERBOSE: print("=========") print(env._debug_print_state()) print(f"Chose action: {action}") print(f"Step: {step_idx}/{max_steps}") if env.step(action) == StepResult.Terminated: print(f"Score: {env.points()}") break

In [ ]:

env.close()

env.close()