def main(_):
params = {}
if FLAGS.symbolic:
params = {'seed': FLAGS.seed, 'level_name': FLAGS.level_name}
env_generator = symbolic_alchemy.get_symbolic_alchemy_level
else:
env_settings = dm_alchemy.EnvironmentSettings(
seed=FLAGS.seed, level_name=FLAGS.level_name)
params = {'name': FLAGS.docker_image_name, 'settings': env_settings}
env_generator = dm_alchemy.load_from_docker
with env_generator(**params) as env:
agent = RandomAgent(env.action_spec())
timestep = env.reset()
score = 0
while not timestep.last():
action = agent.act(timestep)
timestep = env.step(action)
if timestep.reward:
score += timestep.reward
print('Total score: {:.2f}, reward: {:.2f}'.format(
score, timestep.reward))
def test_load_level(self, level_name):
self.assertIsNotNone(
dm_alchemy.load_from_docker(
name=FLAGS.docker_image_name,
settings=dm_alchemy.EnvironmentSettings(
seed=123, level_name=level_name)))
def make_object_under_test(self):
return dm_alchemy.load_from_docker(
name=FLAGS.docker_image_name,
settings=dm_alchemy.EnvironmentSettings(seed=123,
level_name=_TEST_LEVEL))
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import sys
sys.path.insert(0, "//")
import nnn
import tensorflow as tf
import tf_agents as tfa
"""from tf_agents.environments import suite_gym
env = suite_gym.load('CartPole-v0')
#env = tfa.environments.tf_py_environment.TFPyEnvironment(env)
print('time step spec', env.time_step_spec())
print('act spec', env.action_spec())"""
import dm_alchemy
LEVEL_NAME = ('alchemy/perceptual_mapping_'
'randomized_with_rotation_and_random_bottleneck')
settings = dm_alchemy.EnvironmentSettings(seed=123, level_name=LEVEL_NAME)
env = dm_alchemy.load_from_docker(settings)
def test_load_level(self, level_name):
self.assertIsNotNone(
dm_alchemy.load_from_disk(FLAGS.path,
settings=dm_alchemy.EnvironmentSettings(
seed=123, level_name=level_name)))
def make_object_under_test(self):
return dm_alchemy.load_from_disk(
FLAGS.path,
settings=dm_alchemy.EnvironmentSettings(seed=123,
level_name=_TEST_LEVEL))
def main(_):
pygame.init()
try:
pygame.mixer.quit()
except NotImplementedError:
pass
pygame.display.set_caption('Alchemy Human Agent')
view_width, view_height = [int(pixels) for pixels in FLAGS.resolution]
env_settings = dm_alchemy.EnvironmentSettings(seed=FLAGS.seed,
level_name=FLAGS.level_name,
width=view_width,
height=view_height)
# Let rendering use all but one CPU cores, but at least 1 core.
num_rendering_cores = max(multiprocessing.cpu_count() - 1, 1)
environment_variables = {'LP_NUM_THREADS': str(num_rendering_cores)}
with dm_alchemy.load_from_docker(
name=FLAGS.docker_image_name,
settings=env_settings,
environment_variables=environment_variables) as env:
screen = pygame.display.set_mode(
(int(FLAGS.screen_size[0]), int(FLAGS.screen_size[1])))
rgb_spec = env.observation_spec()['RGB_INTERLEAVED']
surface = pygame.Surface((rgb_spec.shape[1], rgb_spec.shape[0]))
score = 0
clock = pygame.time.Clock()
while True:
for event in pygame.event.get():
if event.type == pygame.QUIT:
return
elif event.type == pygame.KEYDOWN:
if event.key == pygame.K_q:
return
if event.key == pygame.K_ESCAPE:
_grab_mouse(not pygame.event.get_grab())
elif event.type == pygame.MOUSEBUTTONDOWN:
if event.button == _LEFT_BUTTON:
_grab_mouse()
actions = _NO_ACTION.copy()
keys = pygame.key.get_pressed()
for key, key_actions in _KEYS_TO_ACTION.items():
if not keys[key]:
continue
for name, action in key_actions.items():
actions[name] = action
if pygame.event.get_grab():
left_button_pressed, _, _ = pygame.mouse.get_pressed()
if left_button_pressed:
actions['HAND_GRIP'] = 1
x, y = pygame.mouse.get_rel()
actions['LOOK_LEFT_RIGHT'] = _MOUSE_SENSITIVITY * x
actions['LOOK_DOWN_UP'] = _MOUSE_SENSITIVITY * y
timestep = env.step(actions)
thisframe = np.copy(timestep.observation['RGB_INTERLEAVED'])
frame = np.swapaxes(thisframe, 0, 1)
pygame.surfarray.blit_array(surface, frame)
pygame.transform.smoothscale(surface, screen.get_size(), screen)
info = pygame.display.Info()
rect_x = info.current_w // 2
rect_y = info.current_h // 2
width_line = 2
pygame.draw.line(screen, _CURSOR_COLOR,
(rect_x - _CURSOR_SIZE, rect_y),
(rect_x + _CURSOR_SIZE, rect_y), width_line)
pygame.draw.line(screen, _CURSOR_COLOR,
(rect_x, rect_y - _CURSOR_SIZE),
(rect_x, rect_y + _CURSOR_SIZE), width_line)
pygame.display.update()
if timestep.reward:
score += timestep.reward
logging.info('Total score: %1.1f, reward: %1.1f', score,
timestep.reward)
clock.tick(_FRAMES_PER_SECOND)