def test_mujoco():
env = neorl.make("HalfCheetah-v3")
train_data, val_data = env.get_dataset(train_num=100,
data_type="L",
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.1) == len(val_data["index"])
env = neorl.make("Walker2d-v3")
train_data, val_data = env.get_dataset(train_num=10,
data_type="m",
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.1) == len(val_data["index"])
env = neorl.make("Hopper-v3")
train_data, val_data = env.get_dataset(train_num=0,
data_type="e",
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.1) == 0 and len(
val_data["index"]) == 0
def get_env(task):
try:
if task.startswith("HalfCheetah-v3"):
env = neorl.make("HalfCheetah-v3")
elif task.startswith("Hopper-v3"):
env = neorl.make("Hopper-v3")
elif task.startswith("Walker2d-v3"):
env = neorl.make("Walker2d-v3")
elif task.startswith('d4rl'):
import d4rl
env = gym.make(task[5:])
else:
task_name = task.strip().split("-")[0]
env = neorl.make(task_name)
except:
raise NotImplementedError
return env
def load_data_from_neorl(task, task_data_type="low", task_train_num=99):
import neorl
env = neorl.make(task)
train_data, val_data = env.get_dataset(data_type=task_data_type,
train_num=task_train_num)
train_buffer, val_buffer = load_neorl_buffer(
train_data), load_neorl_buffer(val_data)
return train_buffer, val_buffer
def test_finance():
env = neorl.make("finance")
train_data, val_data = env.get_dataset(train_num=100,
data_type="H",
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.1) == len(val_data["index"])
def test_citylearn():
env = neorl.make("citylearn")
TEST_NUM = 9
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
path=TEST_DATA_PATH)
assert len(train_data["obs"]) == TEST_NUM * 1000
assert len(train_data["next_obs"]) == TEST_NUM * 1000
assert len(train_data["action"]) == TEST_NUM * 1000
assert len(train_data["reward"]) == TEST_NUM * 1000
assert len(train_data["done"]) == TEST_NUM * 1000
assert len(train_data["index"]) == TEST_NUM
TEST_NUM = 10
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
path=TEST_DATA_PATH)
assert len(train_data["obs"]) == TEST_NUM * 1000
assert len(train_data["next_obs"]) == TEST_NUM * 1000
assert len(train_data["action"]) == TEST_NUM * 1000
assert len(train_data["reward"]) == TEST_NUM * 1000
assert len(train_data["done"]) == TEST_NUM * 1000
assert len(train_data["index"]) == TEST_NUM
TEST_NUM = 99
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
path=TEST_DATA_PATH)
assert len(train_data["obs"]) == TEST_NUM * 1000
assert len(train_data["next_obs"]) == TEST_NUM * 1000
assert len(train_data["action"]) == TEST_NUM * 1000
assert len(train_data["reward"]) == TEST_NUM * 1000
assert len(train_data["done"]) == TEST_NUM * 1000
assert len(train_data["index"]) == TEST_NUM
TEST_NUM = 100
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
path=TEST_DATA_PATH)
assert len(train_data["obs"]) == TEST_NUM * 1000
assert len(train_data["next_obs"]) == TEST_NUM * 1000
assert len(train_data["action"]) == TEST_NUM * 1000
assert len(train_data["reward"]) == TEST_NUM * 1000
assert len(train_data["done"]) == TEST_NUM * 1000
assert len(train_data["index"]) == TEST_NUM
TEST_NUM = 999
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
path=TEST_DATA_PATH)
assert len(train_data["obs"]) == TEST_NUM * 1000
assert len(train_data["next_obs"]) == TEST_NUM * 1000
assert len(train_data["action"]) == TEST_NUM * 1000
assert len(train_data["reward"]) == TEST_NUM * 1000
assert len(train_data["done"]) == TEST_NUM * 1000
assert len(train_data["index"]) == TEST_NUM
TEST_NUM = 99
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
data_type="medium",
path=TEST_DATA_PATH)
assert len(train_data["index"]) == TEST_NUM
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
data_type="low",
path=TEST_DATA_PATH)
assert len(train_data["index"]) == TEST_NUM
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
need_val=False,
path=TEST_DATA_PATH)
assert val_data is None
train_data, val_data = env.get_dataset(train_num=TEST_NUM,
val_ratio=0.3,
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.3) == len(val_data["index"])
def customized_reward_func(data):
obs = data["obs"]
return np.ones((len(obs), 1))
env = neorl.make("ib", reward_func=customized_reward_func)
train_data, val_data = env.get_dataset(data_type="high",
train_num=50,
need_val=False,
use_data_reward=False)
assert len(train_data["index"]) == 50
assert np.all(train_data["reward"] == np.ones_like(train_data["reward"]))
assert val_data is None
import neorl
env = neorl.make("citylearn")
train_data, val_data = env.get_dataset(data_type="medium",
train_num=99,
need_val=True)
print("citylearn:", train_data, val_data)
def test_ib():
env = neorl.make("ib")
train_data, val_data = env.get_dataset(train_num=99,
data_type="M",
path=TEST_DATA_PATH)
assert int(len(train_data["index"]) * 0.1) == len(val_data["index"])
import neorl
if __name__ == '__main__':
for domain in [
'HalfCheetah-v3', 'Hopper-v3', 'Walker2d-v3', 'ib', 'finance',
'citylearn'
]:
for level in ['low', 'medium', 'high']:
for amount in [100, 1000, 10000
] if not domain == 'finance' else [100, 1000]:
env = neorl.make(domain)
env.get_dataset(data_type=level, train_num=amount)