def test_discrete_space(n):
sp = Discrete(n)
for ii in range(n):
assert sp.contains(ii)
for ii in range(2 * n):
assert sp.contains(sp.sample())
class ActionDictTestEnv(gym.Env):
action_space = Dict({"position": Discrete(1), "velocity": Discrete(1)})
observation_space = Box(low=-1.0, high=2.0, shape=(3, ), dtype=np.float32)
def step(self, action):
observation = np.array([1.0, 1.5, 0.5])
reward = 1
done = True
return observation, reward, done
def reset(self):
return np.array([1.0, 1.5, 0.5])
def test_tuple():
sp1 = Box(0.0, 1.0, shape=(3, 2))
sp2 = Discrete(2)
sp = Tuple([sp1, sp2])
for _ in range(10):
assert sp.contains(sp.sample())
sp.reseed()
def convert_space_from_gym(gym_space):
if isinstance(gym_space, gym.spaces.Discrete):
return Discrete(gym_space.n)
#
#
elif isinstance(gym_space, gym.spaces.Box):
return Box(gym_space.low, gym_space.high, gym_space.shape,
gym_space.dtype)
#
#
elif isinstance(gym_space, gym.spaces.Tuple):
spaces = []
for sp in gym_space.spaces:
spaces.append(convert_space_from_gym(sp))
return Tuple(spaces)
#
#
elif isinstance(gym_space, gym.spaces.MultiDiscrete):
return MultiDiscrete(gym_space.nvec)
#
#
elif isinstance(gym_space, gym.spaces.MultiBinary):
return MultiBinary(gym_space.n)
#
#
elif isinstance(gym_space, gym.spaces.Dict):
spaces = {}
for key in gym_space.spaces:
spaces[key] = convert_space_from_gym(gym_space[key])
return Dict(spaces)
else:
raise ValueError("Unknown space class: {}".format(type(gym_space)))
def test_dict():
nested_observation_space = Dict({
'sensors':
Dict({
'position':
Box(low=-100, high=100, shape=(3, )),
'velocity':
Box(low=-1, high=1, shape=(3, )),
'front_cam':
Tuple((Box(low=0, high=1,
shape=(10, 10, 3)), Box(low=0,
high=1,
shape=(10, 10, 3)))),
'rear_cam':
Box(low=0, high=1, shape=(10, 10, 3)),
}),
'ext_controller':
MultiDiscrete((5, 2, 2)),
'inner_state':
Dict({
'charge':
Discrete(100),
'system_checks':
MultiBinary(10),
'job_status':
Dict({
'task': Discrete(5),
'progress': Box(low=0, high=100, shape=()),
})
})
})
sp = nested_observation_space
for _ in range(10):
assert sp.contains(sp.sample())
sp.reseed()
sp2 = Dict(sp.spaces)
for _ in range(10):
assert sp.contains(sp2.sample())
sp2.reseed()
def test_dict():
nested_observation_space = Dict({
"sensors":
Dict({
"position":
Box(low=-100, high=100, shape=(3, )),
"velocity":
Box(low=-1, high=1, shape=(3, )),
"front_cam":
Tuple((
Box(low=0, high=1, shape=(10, 10, 3)),
Box(low=0, high=1, shape=(10, 10, 3)),
)),
"rear_cam":
Box(low=0, high=1, shape=(10, 10, 3)),
}),
"ext_controller":
MultiDiscrete((5, 2, 2)),
"inner_state":
Dict({
"charge":
Discrete(100),
"system_checks":
MultiBinary(10),
"job_status":
Dict({
"task": Discrete(5),
"progress": Box(low=0, high=100, shape=()),
}),
}),
})
sp = nested_observation_space
for _ in range(10):
assert sp.contains(sp.sample())
sp.reseed()
sp2 = Dict(sp.spaces)
for _ in range(10):
assert sp.contains(sp2.sample())
sp2.reseed()