def test_reset(self):
env = MultiAgentAtariEnv('pong_classic_v0')
state = env.reset()
self.assertEqual(state.observation.shape, (1, 84, 84))
self.assertEqual(state.reward, 0)
self.assertEqual(state.done, False)
self.assertEqual(state.mask, 1.)
def test_init(self):
MultiAgentAtariEnv('pong_classic_v0')
MultiAgentAtariEnv('mario_bros_v1')
MultiAgentAtariEnv('entombed_cooperative_v0')
def test_list_agents(self):
env = MultiAgentAtariEnv('pong_classic_v0')
self.assertEquals(env.agents, ['first_0', 'second_0'])
def test_action_spaces(self):
env = MultiAgentAtariEnv('pong_classic_v0')
action_spaces = env.action_spaces
print(action_spaces)
def test_state_spaces(self):
env = MultiAgentAtariEnv('pong_classic_v0')
state_spaces = env.state_spaces
self.assertEqual(state_spaces['first_0'].shape, (84, 84))
self.assertEqual(state_spaces['second_0'].shape, (84, 84))
def test_name(self):
env = MultiAgentAtariEnv('pong_classic_v0')
self.assertEqual(env.name, 'pong_classic_v0')