Hands-On Intelligent Agents with OpenAI Gym

Wrapping the Gym environment

Finally, we will apply the preceding wrappers that we developed based on the environment configuration we specify using the parameters.JSON file:

def make_env(env_id, env_conf):    env = gym.make(env_id)    if 'NoFrameskip' in env_id:        assert 'NoFrameskip' in env.spec.id        env = NoopResetEnv(env, noop_max=30)        env = MaxAndSkipEnv(env, skip=env_conf['skip_rate'])    if env_conf['episodic_life']:        env = EpisodicLifeEnv(env)    if 'FIRE' in env.unwrapped.get_action_meanings():        env = FireResetEnv(env)    env = AtariRescale(env, env_conf['useful_region'])    env = NormalizedEnv(env)    if env_conf['clip_reward']:        env = ClipRewardEnv(env)    return env

All of the environment wrappers that we discussed previously are implemented and available in ...