dodox/fancy_gym
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Fix: Make wrappers work with BB and Dict-Space

Browse Source
This commit is contained in:
dodox
2023-06-18 14:25:20 +02:00
parent b032dec5fe
commit 9ade0dcdc4
3 changed files with 150 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/test_black_box.py
+4 -7
View File
@@ -9,7 +9,7 @@ from gymnasium.core import ActType, ObsType
import fancy_gym
from fancy_gym.black_box.raw_interface_wrapper import RawInterfaceWrapper
from fancy_gym.utils.time_aware_observation import TimeAwareObservation
from fancy_gym.utils.wrappers import TimeAwareObservation
SEED = 1
ENV_IDS = ['Reacher5d-v0', 'dmc:ball_in_cup-catch-v0', 'metaworld:reach-v2', 'Reacher-v2']
@@ -17,6 +17,8 @@ WRAPPERS = [fancy_gym.envs.mujoco.reacher.MPWrapper, fancy_gym.dmc.suite.ball_in
fancy_gym.meta.goal_object_change_mp_wrapper.MPWrapper, fancy_gym.open_ai.mujoco.reacher_v2.MPWrapper]
ALL_MP_ENVS = chain(*fancy_gym.ALL_MOVEMENT_PRIMITIVE_ENVIRONMENTS.values())
MAX_STEPS_FALLBACK = 500
class Object(object):
pass
@@ -115,11 +117,6 @@ def test_verbosity(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]]
@pytest.mark.parametrize('mp_type', ['promp', 'dmp', 'prodmp'])
@pytest.mark.parametrize('env_wrap', zip(ENV_IDS, WRAPPERS))
def test_length(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]]):
if not env.spec.max_episode_steps:
# Not all envs expose a max_episode_steps.
# To use those with MPs, they could be put in a time_limit-wrapper.
return True
basis_generator_type = 'prodmp' if mp_type == 'prodmp' else 'rbf'
env_id, wrapper_class = env_wrap
@@ -127,7 +124,7 @@ def test_length(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]]):
{'trajectory_generator_type': mp_type},
{'controller_type': 'motor'},
{'phase_generator_type': 'exp'},
{'basis_generator_type': basis_generator_type})
{'basis_generator_type': basis_generator_type}, fallback_max_steps=MAX_STEPS_FALLBACK)
for i in range(5):
env.reset(seed=SEED)
test/test_replanning_sequencing.py
+19 -18
View File
@@ -11,7 +11,8 @@ from gymnasium import spaces
import fancy_gym
from fancy_gym.black_box.raw_interface_wrapper import RawInterfaceWrapper
from fancy_gym.utils.time_aware_observation import TimeAwareObservation
from fancy_gym.utils.wrappers import TimeAwareObservation
from fancy_gym.utils.make_env_helpers import ensure_finite_time
SEED = 1
ENV_IDS = ['Reacher5d-v0', 'dmc:ball_in_cup-catch-v0', 'metaworld:reach-v2', 'Reacher-v2']
@@ -19,6 +20,8 @@ WRAPPERS = [fancy_gym.envs.mujoco.reacher.MPWrapper, fancy_gym.dmc.suite.ball_in
fancy_gym.meta.goal_object_change_mp_wrapper.MPWrapper, fancy_gym.open_ai.mujoco.reacher_v2.MPWrapper]
ALL_MP_ENVS = chain(*fancy_gym.ALL_MOVEMENT_PRIMITIVE_ENVIRONMENTS.values())
MAX_STEPS_FALLBACK = 100
class ToyEnv(gym.Env):
observation_space = gym.spaces.Box(low=-1, high=1, shape=(1,), dtype=np.float64)
@@ -64,7 +67,7 @@ def setup():
def test_learn_sub_trajectories(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]],
add_time_aware_wrapper_before: bool):
env_id, wrapper_class = env_wrap
env_step = TimeAwareObservation(fancy_gym.make(env_id, SEED))
env_step = TimeAwareObservation(ensure_finite_time(fancy_gym.make(env_id, SEED), MAX_STEPS_FALLBACK))
wrappers = [wrapper_class]
# has time aware wrapper
@@ -75,15 +78,14 @@ def test_learn_sub_trajectories(mp_type: str, env_wrap: Tuple[str, Type[RawInter
{'trajectory_generator_type': mp_type},
{'controller_type': 'motor'},
{'phase_generator_type': 'exp'},
{'basis_generator_type': 'rbf'}, seed=SEED)
{'basis_generator_type': 'rbf'}, seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
assert env.learn_sub_trajectories
assert env.spec.max_episode_steps
assert env_step.spec.max_episode_steps
assert env.traj_gen.learn_tau
# This also verifies we are not adding the TimeAwareObservationWrapper twice
if env.observation_space.__class__ in [spaces.Dict]:
assert spaces.flatten_space(env.observation_space) == env_step.observation_space
else:
assert env.observation_space == env_step.observation_space
assert spaces.flatten_space(env_step.observation_space) == spaces.flatten_space(env.observation_space)
done = True
@@ -112,7 +114,7 @@ def test_learn_sub_trajectories(mp_type: str, env_wrap: Tuple[str, Type[RawInter
def test_replanning_time(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWrapper]],
add_time_aware_wrapper_before: bool, replanning_time: int):
env_id, wrapper_class = env_wrap
env_step = TimeAwareObservation(fancy_gym.make(env_id, SEED))
env_step = TimeAwareObservation(ensure_finite_time(fancy_gym.make(env_id, SEED), MAX_STEPS_FALLBACK))
wrappers = [wrapper_class]
# has time aware wrapper
@@ -128,15 +130,14 @@ def test_replanning_time(mp_type: str, env_wrap: Tuple[str, Type[RawInterfaceWra
{'trajectory_generator_type': mp_type},
{'controller_type': 'motor'},
{'phase_generator_type': phase_generator_type},
{'basis_generator_type': basis_generator_type}, seed=SEED)
{'basis_generator_type': basis_generator_type}, seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
assert env.do_replanning
assert env.spec.max_episode_steps
assert env_step.spec.max_episode_steps
assert callable(env.replanning_schedule)
# This also verifies we are not adding the TimeAwareObservationWrapper twice
if env.observation_space.__class__ in [spaces.Dict]:
assert spaces.flatten_space(env.observation_space) == env_step.observation_space
else:
assert env.observation_space == env_step.observation_space
assert spaces.flatten_space(env_step.observation_space) == spaces.flatten_space(env.observation_space)
env.reset(seed=SEED)
@@ -177,7 +178,7 @@ def test_max_planning_times(mp_type: str, max_planning_times: int, sub_segment_s
},
{'basis_generator_type': basis_generator_type,
},
seed=SEED)
seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
_ = env.reset(seed=SEED)
done = False
planning_times = 0
@@ -209,7 +210,7 @@ def test_replanning_with_learn_tau(mp_type: str, max_planning_times: int, sub_se
},
{'basis_generator_type': basis_generator_type,
},
seed=SEED)
seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
_ = env.reset(seed=SEED)
done = False
planning_times = 0
@@ -242,7 +243,7 @@ def test_replanning_with_learn_delay(mp_type: str, max_planning_times: int, sub_
},
{'basis_generator_type': basis_generator_type,
},
seed=SEED)
seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
_ = env.reset(seed=SEED)
done = False
planning_times = 0
@@ -297,7 +298,7 @@ def test_replanning_with_learn_delay_and_tau(mp_type: str, max_planning_times: i
},
{'basis_generator_type': basis_generator_type,
},
seed=SEED)
seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
_ = env.reset(seed=SEED)
done = False
planning_times = 0
@@ -346,7 +347,7 @@ def test_replanning_schedule(mp_type: str, max_planning_times: int, sub_segment_
},
{'basis_generator_type': basis_generator_type,
},
seed=SEED)
seed=SEED, fallback_max_steps=MAX_STEPS_FALLBACK)
_ = env.reset(seed=SEED)
for i in range(max_planning_times):
action = env.action_space.sample()