basic structure of an openai gym env

custom_open_ai_gym_env.py

import gym
from gym import spaces

class MyEnv(gym.Env):
    """Custom Environment"""
    metadata = {'render.modes': ['human']}
    
    def __init__(self):
        super(Snake, self).__init__()
        # necessary internal env id
        self.id = "MyEnv"
        # defines the kind of actions you can take
        # below it is an example where the possible actions are only the values 0 or 1
        self.action_space = spaces.Discrete(2)
        # defines the kind of observations you can make
        # below it is an example where the possible observations are only the values 0 or 1
        self.observation_space = spaces.Discrete(2)
       
    def reset(self):
      ''' Resets the game and returns the initial state
      '''
      pass
    
    def render(self, mode='human'):
      ''' renders the game
      '''
      pass
    
    def step(self, action):
      ''' performs the action and observes the state, get the reward, check if the game
      is over and get more info
      '''
      obs = None
      reward = None
      done = None
      info = {}
      return obs, reward, done, info