mirror of
https://github.com/Farama-Foundation/Gymnasium.git
synced 2025-08-01 22:11:25 +00:00
c364506710
* Ditch most of the seeding.py and replace np_random with the numpy default_rng. Let's see if tests pass * Updated a bunch of RNG calls from the RandomState API to Generator API * black; didn't expect that, did ya? * Undo a typo * blaaack * More typo fixes * Fixed setting/getting state in multidiscrete spaces * Fix typo, fix a test to work with the new sampling * Correctly (?) pass the randomly generated seed if np_random is called with None as seed * Convert the Discrete sample to a python int (as opposed to np.int64) * Remove some redundant imports * First version of the compatibility layer for old-style RNG. Mainly to trigger tests. * Removed redundant f-strings * Style fixes, removing unused imports * Try to make tests pass by removing atari from the dockerfile * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * First attempt at deprecating `env.seed` and supporting `env.reset(seed=seed)` instead. Tests should hopefully pass but throw up a million warnings. * black; didn't expect that, didya? * Rename the reset parameter in VecEnvs back to `seed` * Updated tests to use the new seeding method * Removed a bunch of old `seed` calls. Fixed a bug in AsyncVectorEnv * Stop Discrete envs from doing part of the setup (and using the randomness) in init (as opposed to reset) * Add explicit seed to wrappers reset * Remove an accidental return * Re-add some legacy functions with a warning. * Use deprecation instead of regular warnings for the newly deprecated methods/functions
90 lines
2.9 KiB
Python
90 lines
2.9 KiB
Python
import numpy as np
|
|
import gym
|
|
from gym.wrappers import AtariPreprocessing
|
|
import pytest
|
|
|
|
pytest.importorskip("gym.envs.atari")
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def env_fn():
|
|
return lambda: gym.make("PongNoFrameskip-v4")
|
|
|
|
|
|
def test_atari_preprocessing_grayscale(env_fn):
|
|
import cv2
|
|
|
|
env1 = env_fn()
|
|
env2 = AtariPreprocessing(
|
|
env_fn(), screen_size=84, grayscale_obs=True, frame_skip=1, noop_max=0
|
|
)
|
|
env3 = AtariPreprocessing(
|
|
env_fn(), screen_size=84, grayscale_obs=False, frame_skip=1, noop_max=0
|
|
)
|
|
env4 = AtariPreprocessing(
|
|
env_fn(),
|
|
screen_size=84,
|
|
grayscale_obs=True,
|
|
frame_skip=1,
|
|
noop_max=0,
|
|
grayscale_newaxis=True,
|
|
)
|
|
obs1 = env1.reset(seed=0)
|
|
obs2 = env2.reset(seed=0)
|
|
obs3 = env3.reset(seed=0)
|
|
obs4 = env4.reset(seed=0)
|
|
assert env1.observation_space.shape == (210, 160, 3)
|
|
assert env2.observation_space.shape == (84, 84)
|
|
assert env3.observation_space.shape == (84, 84, 3)
|
|
assert env4.observation_space.shape == (84, 84, 1)
|
|
assert obs1.shape == (210, 160, 3)
|
|
assert obs2.shape == (84, 84)
|
|
assert obs3.shape == (84, 84, 3)
|
|
assert obs4.shape == (84, 84, 1)
|
|
assert np.allclose(obs3, cv2.resize(obs1, (84, 84), interpolation=cv2.INTER_AREA))
|
|
obs3_gray = cv2.cvtColor(obs3, cv2.COLOR_RGB2GRAY)
|
|
# the edges of the numbers do not render quite the same in the grayscale, so we ignore them
|
|
assert np.allclose(obs2[10:38], obs3_gray[10:38])
|
|
# the paddle also do not render quite the same
|
|
assert np.allclose(obs2[44:], obs3_gray[44:])
|
|
# now add a channel axis and re-test
|
|
obs3_gray = obs3_gray.reshape(84, 84, 1)
|
|
assert np.allclose(obs4[10:38], obs3_gray[10:38])
|
|
assert np.allclose(obs4[44:], obs3_gray[44:])
|
|
|
|
env1.close()
|
|
env2.close()
|
|
env3.close()
|
|
env4.close()
|
|
|
|
|
|
def test_atari_preprocessing_scale(env_fn):
|
|
# arbitrarily chosen number for stepping into env. and ensuring all observations are in the required range
|
|
max_test_steps = 10
|
|
|
|
for grayscale in [True, False]:
|
|
for scaled in [True, False]:
|
|
env = AtariPreprocessing(
|
|
env_fn(),
|
|
screen_size=84,
|
|
grayscale_obs=grayscale,
|
|
scale_obs=scaled,
|
|
frame_skip=1,
|
|
noop_max=0,
|
|
)
|
|
obs = env.reset().flatten()
|
|
done, step_i = False, 0
|
|
max_obs = 1 if scaled else 255
|
|
assert (0 <= obs).all() and (
|
|
obs <= max_obs
|
|
).all(), "Obs. must be in range [0,{}]".format(max_obs)
|
|
while not done or step_i <= max_test_steps:
|
|
obs, _, done, _ = env.step(env.action_space.sample())
|
|
obs = obs.flatten()
|
|
assert (0 <= obs).all() and (
|
|
obs <= max_obs
|
|
).all(), "Obs. must be in range [0,{}]".format(max_obs)
|
|
step_i += 1
|
|
|
|
env.close()
|