testing becca gym

becca_gym.py

import numpy as np
import becca.brain as becca_brain
from becca.base_world import World as BaseWorld
import gym

class World(BaseWorld):

    def __init__(self, env_name, seed=None):
        BaseWorld.__init__(self)
        self.name = 'gym_cartPole'
        self.visualize_interval = 1e4


        self.env = gym.make(env_name)
        if seed is not None:
            self.env.seed(seed)

        self.n_sensors, self.n_actions = self.env.observation_space.shape[0], self.env.action_space.n
        print('states are', self.n_sensors, self.n_actions)
        self.sensors = self.env.reset()
        self.done = False

    def step(self, actions):

        if actions[0] == .5:
            raise Exception("input actions are float -", actions[0])


        self.sensors, reward, self.done, _ = self.env.step(np.argmax(actions))

        if self.timestep % self.visualize_interval == 0:
            report = "got reward of " + str(reward)
            # Give an update
            print(report)


        if self.done is True:
            print("done")
            self.sensors = self.env.reset()
        return self.sensors, reward

    # def is_alive(self):
    #     return not self.done

if __name__ == "__main__":
    becca_brain.run(World("CartPole-v0"))