def test_ppo_model_cc_visual(mock_communicator, mock_launcher):
tf.reset_default_graph()
with tf.Session() as sess:
with tf.variable_scope("FakeGraphScope"):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=2)
env = UnityEnvironment(' ')
model = PPOModel(env.brains["RealFakeBrain"])
init = tf.global_variables_initializer()
sess.run(init)
run_list = [
model.output, model.log_probs, model.value, model.entropy,
model.learning_rate
]
feed_dict = {
model.batch_size: 2,
model.sequence_length: 1,
model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
[3, 4, 5, 3, 4, 5]]),
model.visual_in[0]: np.ones([2, 40, 30, 3]),
model.visual_in[1]: np.ones([2, 40, 30, 3]),
model.epsilon: np.array([[0, 1], [2, 3]])
}
sess.run(run_list, feed_dict=feed_dict)
env.close()
def test_ppo_model_dc_vector_curio(mock_communicator, mock_launcher):
tf.reset_default_graph()
with tf.Session() as sess:
with tf.variable_scope("FakeGraphScope"):
mock_communicator.return_value = MockCommunicator(
discrete_action=True, visual_inputs=0)
env = UnityEnvironment(' ')
model = PPOModel(env.brains["RealFakeBrain"], use_curiosity=True)
init = tf.global_variables_initializer()
sess.run(init)
run_list = [
model.output, model.all_log_probs, model.value, model.entropy,
model.learning_rate, model.intrinsic_reward
]
feed_dict = {
model.batch_size:
2,
model.sequence_length:
1,
model.vector_in:
np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
model.next_vector_in:
np.array([[1, 2, 3, 1, 2, 3], [3, 4, 5, 3, 4, 5]]),
model.action_holder: [[0], [0]],
model.action_masks:
np.ones([2, 2])
}
sess.run(run_list, feed_dict=feed_dict)
env.close()
def test_step(mock_communicator, mock_launcher):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=0)
env = UnityEnvironment(' ')
brain = env.brains['RealFakeBrain']
brain_info = env.reset()
brain_info = env.step([0] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
with pytest.raises(UnityActionException):
env.step([0])
brain_info = env.step([-1] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
with pytest.raises(UnityActionException):
env.step([0] * brain.vector_action_space_size[0] * len(brain_info['RealFakeBrain'].agents))
env.close()
assert env.global_done
assert isinstance(brain_info, dict)
assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
assert isinstance(brain_info['RealFakeBrain'].visual_observations, list)
assert isinstance(brain_info['RealFakeBrain'].vector_observations, np.ndarray)
assert len(brain_info['RealFakeBrain'].visual_observations) == brain.number_visual_observations
assert len(brain_info['RealFakeBrain'].vector_observations) == \
len(brain_info['RealFakeBrain'].agents)
assert len(brain_info['RealFakeBrain'].vector_observations[0]) == \
brain.vector_observation_space_size * brain.num_stacked_vector_observations
print("\n\n\n\n\n\n\n" + str(brain_info['RealFakeBrain'].local_done))
assert not brain_info['RealFakeBrain'].local_done[0]
assert brain_info['RealFakeBrain'].local_done[2]
def test_ppo_model_cc_vector_rnn(mock_communicator, mock_launcher):
tf.reset_default_graph()
with tf.Session() as sess:
with tf.variable_scope("FakeGraphScope"):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=0)
env = UnityEnvironment(' ')
memory_size = 128
model = PPOModel(env.brains["RealFakeBrain"],
use_recurrent=True,
m_size=memory_size)
init = tf.global_variables_initializer()
sess.run(init)
run_list = [
model.output, model.all_log_probs, model.value, model.entropy,
model.learning_rate, model.memory_out
]
feed_dict = {
model.batch_size: 1,
model.sequence_length: 2,
model.memory_in: np.zeros((1, memory_size)),
model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
[3, 4, 5, 3, 4, 5]]),
model.epsilon: np.array([[0, 1]])
}
sess.run(run_list, feed_dict=feed_dict)
env.close()
def test_initialization(mock_communicator, mock_launcher):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=0)
env = UnityEnvironment(' ')
with pytest.raises(UnityActionException):
env.step([0])
assert env.brain_names[0] == 'RealFakeBrain'
env.close()
def test_close(mock_communicator, mock_launcher):
comm = MockCommunicator(
discrete_action=False, visual_inputs=0)
mock_communicator.return_value = comm
env = UnityEnvironment(' ')
assert env._loaded
env.close()
assert not env._loaded
assert comm.has_been_closed
def test_ppo_policy_evaluate(mock_communicator, mock_launcher, dummy_config):
tf.reset_default_graph()
mock_communicator.return_value = MockCommunicator(discrete_action=False,
visual_inputs=0)
env = UnityEnvironment(' ')
brain_infos = env.reset()
brain_info = brain_infos[env.brain_names[0]]
trainer_parameters = dummy_config
model_path = env.brain_names[0]
trainer_parameters['model_path'] = model_path
trainer_parameters['keep_checkpoints'] = 3
policy = PPOPolicy(0, env.brains[env.brain_names[0]], trainer_parameters,
False, False)
run_out = policy.evaluate(brain_info)
assert run_out['action'].shape == (3, 2)
env.close()
def test_reset(mock_communicator, mock_launcher):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=0)
env = UnityEnvironment(' ')
brain = env.brains['RealFakeBrain']
brain_info = env.reset()
env.close()
assert not env.global_done
assert isinstance(brain_info, dict)
assert isinstance(brain_info['RealFakeBrain'], BrainInfo)
assert isinstance(brain_info['RealFakeBrain'].visual_observations, list)
assert isinstance(brain_info['RealFakeBrain'].vector_observations, np.ndarray)
assert len(brain_info['RealFakeBrain'].visual_observations) == brain.number_visual_observations
assert len(brain_info['RealFakeBrain'].vector_observations) == \
len(brain_info['RealFakeBrain'].agents)
assert len(brain_info['RealFakeBrain'].vector_observations[0]) == \
brain.vector_observation_space_size * brain.num_stacked_vector_observations
def test_cc_bc_model(mock_communicator, mock_launcher):
tf.reset_default_graph()
with tf.Session() as sess:
with tf.variable_scope("FakeGraphScope"):
mock_communicator.return_value = MockCommunicator(
discrete_action=False, visual_inputs=0)
env = UnityEnvironment(' ')
model = BehavioralCloningModel(env.brains["RealFakeBrain"])
init = tf.global_variables_initializer()
sess.run(init)
run_list = [model.sample_action, model.policy]
feed_dict = {
model.batch_size: 2,
model.sequence_length: 1,
model.vector_in: np.array([[1, 2, 3, 1, 2, 3],
[3, 4, 5, 3, 4, 5]])
}
sess.run(run_list, feed_dict=feed_dict)
env.close()
class UnityEnv(gym.Env):
"""
Provides Gym wrapper for Unity Learning Environments.
Multi-agent environments use lists for object types, as done here:
https://github.com/openai/multiagent-particle-envs
"""
def __init__(self,
environment_filename: str,
worker_id=0,
use_visual=False,
uint8_visual=False,
multiagent=False,
flatten_branched=False):
"""
Environment initialization
:param environment_filename: The UnityEnvironment path or file to be wrapped in the gym.
:param worker_id: Worker number for environment.
:param use_visual: Whether to use visual observation or vector observation.
:param uint8_visual: Return visual observations as uint8 (0-255) matrices instead of float (0.0-1.0).
:param multiagent: Whether to run in multi-agent mode (lists of obs, reward, done).
:param flatten_branched: If True, turn branched discrete action spaces into a Discrete space rather than MultiDiscrete.
"""
self._env = UnityEnvironment(environment_filename, worker_id)
self.name = self._env.academy_name
self.visual_obs = None
self._current_state = None
self._n_agents = None
self._multiagent = multiagent
self._flattener = None
self.game_over = False # Hidden flag used by Atari environments to determine if the game is over
# Check brain configuration
if len(self._env.brains) != 1:
raise UnityGymException(
"There can only be one brain in a UnityEnvironment "
"if it is wrapped in a gym.")
self.brain_name = self._env.external_brain_names[0]
brain = self._env.brains[self.brain_name]
if use_visual and brain.number_visual_observations == 0:
raise UnityGymException(
"`use_visual` was set to True, however there are no"
" visual observations as part of this environment.")
self.use_visual = brain.number_visual_observations >= 1 and use_visual
if not use_visual and uint8_visual:
logger.warning(
"`uint8_visual was set to true, but visual observations are not in use. "
"This setting will not have any effect.")
else:
self.uint8_visual = uint8_visual
if brain.number_visual_observations > 1:
logger.warning(
"The environment contains more than one visual observation. "
"Please note that only the first will be provided in the observation."
)
if brain.num_stacked_vector_observations != 1:
raise UnityGymException(
"There can only be one stacked vector observation in a UnityEnvironment "
"if it is wrapped in a gym.")
# Check for number of agents in scene.
initial_info = self._env.reset()[self.brain_name]
self._check_agents(len(initial_info.agents))
# Set observation and action spaces
if brain.vector_action_space_type == "discrete":
if len(brain.vector_action_space_size) == 1:
self._action_space = spaces.Discrete(
brain.vector_action_space_size[0])
else:
if flatten_branched:
self._flattener = ActionFlattener(
brain.vector_action_space_size)
self._action_space = self._flattener.action_space
else:
self._action_space = spaces.MultiDiscrete(
brain.vector_action_space_size)
else:
if flatten_branched:
logger.warning(
"The environment has a non-discrete action space. It will "
"not be flattened.")
high = np.array([1] * brain.vector_action_space_size[0])
self._action_space = spaces.Box(-high, high, dtype=np.float32)
high = np.array([np.inf] * brain.vector_observation_space_size)
self.action_meanings = brain.vector_action_descriptions
if self.use_visual:
if brain.camera_resolutions[0]["blackAndWhite"]:
depth = 1
else:
depth = 3
self._observation_space = spaces.Box(
0,
1,
dtype=np.float32,
shape=(brain.camera_resolutions[0]["height"],
brain.camera_resolutions[0]["width"], depth))
else:
self._observation_space = spaces.Box(-high, high, dtype=np.float32)
def reset(self):
"""Resets the state of the environment and returns an initial observation.
In the case of multi-agent environments, this is a list.
Returns: observation (object/list): the initial observation of the
space.
"""
info = self._env.reset()[self.brain_name]
n_agents = len(info.agents)
self._check_agents(n_agents)
self.game_over = False
if not self._multiagent:
obs, reward, done, info = self._single_step(info)
else:
obs, reward, done, info = self._multi_step(info)
return obs
def step(self, action):
"""Run one timestep of the environment's dynamics. When end of
episode is reached, you are responsible for calling `reset()`
to reset this environment's state.
Accepts an action and returns a tuple (observation, reward, done, info).
In the case of multi-agent environments, these are lists.
Args:
action (object/list): an action provided by the environment
Returns:
observation (object/list): agent's observation of the current environment
reward (float/list) : amount of reward returned after previous action
done (boolean/list): whether the episode has ended.
info (dict): contains auxiliary diagnostic information, including BrainInfo.
"""
# Use random actions for all other agents in environment.
if self._multiagent:
if not isinstance(action, list):
raise UnityGymException(
"The environment was expecting `action` to be a list.")
if len(action) != self._n_agents:
raise UnityGymException(
"The environment was expecting a list of {} actions.".
format(self._n_agents))
else:
if self._flattener is not None:
# Action space is discrete and flattened - we expect a list of scalars
action = [
self._flattener.lookup_action(_act) for _act in action
]
action = np.array(action)
else:
if self._flattener is not None:
# Translate action into list
action = self._flattener.lookup_action(action)
info = self._env.step(action)[self.brain_name]
n_agents = len(info.agents)
self._check_agents(n_agents)
self._current_state = info
if not self._multiagent:
obs, reward, done, info = self._single_step(info)
self.game_over = done
else:
obs, reward, done, info = self._multi_step(info)
self.game_over = all(done)
return obs, reward, done, info
def _single_step(self, info):
if self.use_visual:
self.visual_obs = self._preprocess_single(
info.visual_observations[0][0, :, :, :])
default_observation = self.visual_obs
else:
default_observation = info.vector_observations[0, :]
return default_observation, info.rewards[0], info.local_done[0], {
"text_observation": info.text_observations[0],
"brain_info": info
}
def _preprocess_single(self, single_visual_obs):
if self.uint8_visual:
return (255.0 * single_visual_obs).astype(np.uint8)
else:
return single_visual_obs
def _multi_step(self, info):
if self.use_visual:
self.visual_obs = self._preprocess_multi(info.visual_observations)
default_observation = self.visual_obs
else:
default_observation = info.vector_observations
return list(default_observation), info.rewards, info.local_done, {
"text_observation": info.text_observations,
"brain_info": info
}
def _preprocess_multi(self, multiple_visual_obs):
if self.uint8_visual:
return [(255.0 * _visual_obs).astype(np.uint8)
for _visual_obs in multiple_visual_obs]
else:
return multiple_visual_obs
def render(self, mode='rgb_array'):
return self.visual_obs
def close(self):
"""Override _close in your subclass to perform any necessary cleanup.
Environments will automatically close() themselves when
garbage collected or when the program exits.
"""
self._env.close()
def get_action_meanings(self):
return self.action_meanings
def seed(self, seed=None):
"""Sets the seed for this env's random number generator(s).
Currently not implemented.
"""
logger.warn("Could not seed environment %s", self.name)
return
def _check_agents(self, n_agents):
if not self._multiagent and n_agents > 1:
raise UnityGymException(
"The environment was launched as a single-agent environment, however"
"there is more than one agent in the scene.")
elif self._multiagent and n_agents <= 1:
raise UnityGymException(
"The environment was launched as a mutli-agent environment, however"
"there is only one agent in the scene.")
if self._n_agents is None:
self._n_agents = n_agents
logger.info("{} agents within environment.".format(n_agents))
elif self._n_agents != n_agents:
raise UnityGymException(
"The number of agents in the environment has changed since "
"initialization. This is not supported.")
@property
def metadata(self):
return {'render.modes': ['rgb_array']}
@property
def reward_range(self):
return -float('inf'), float('inf')
@property
def spec(self):
return None
@property
def action_space(self):
return self._action_space
@property
def observation_space(self):
return self._observation_space
@property
def number_agents(self):
return self._n_agents