def test_forecast_tcmf_distributed(self):
input = dict({'id': self.id, 'y': self.data})
from zoo.orca import init_orca_context, stop_orca_context
init_orca_context(cores=4, spark_log_level="INFO", init_ray_on_spark=True,
object_store_memory="1g")
self.model.fit(input, num_workers=4, **self.fit_params)
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat = self.model.predict(horizon=self.horizon, num_workers=4)
yhat_loaded = loaded_model.predict(horizon=self.horizon, num_workers=4)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_equal(yhat, yhat_loaded)
self.model.fit_incremental({'y': self.data_new})
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value, metric=['mse'])
stop_orca_context()
def test_forecast_tcmf_ndarray(self):
ndarray_input = {'id': self.id, 'y': self.data}
self.model.fit(ndarray_input, **self.fit_params)
assert not self.model.is_xshards_distributed()
# test predict
yhat = self.model.predict(horizon=self.horizon)
# test save load
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat_loaded = loaded_model.predict(horizon=self.horizon)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
# test evaluate
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value, metric=['mse'])
# test fit_incremental
self.model.fit_incremental({'y': self.data_new}) # 1st time
self.model.fit_incremental({'y': self.data_new}) # 2nd time
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
def setUp(self):
self.model = TCMFForecaster()
self.num_samples = 300
self.horizon = np.random.randint(1, 50)
self.seq_len = 480
self.data = np.random.rand(self.num_samples, self.seq_len)
self.id = np.arange(self.num_samples)
self.data_new = np.random.rand(self.num_samples, self.horizon)
self.fit_params = dict(val_len=12,
start_date="2020-1-1",
freq="5min",
y_iters=1,
init_FX_epoch=1,
max_FX_epoch=1,
max_TCN_epoch=1,
alt_iters=2)
def test_forecast_tcmf_without_id(self):
# construct data
input = dict({'y': self.data})
self.model.fit(input, **self.fit_params)
assert not self.model.is_xshards_distributed()
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat = self.model.predict(horizon=self.horizon)
yhat_loaded = loaded_model.predict(horizon=self.horizon)
assert "id" not in yhat_loaded
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
target_value = dict({"y": self.data_new})
self.model.evaluate(target_value=target_value, metric=['mse'])
self.model.fit_incremental({'y': self.data_new}) # 1st time
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
data_new_id = {'id': self.id, 'y': self.data_new}
with self.assertRaises(ValueError) as context:
self.model.fit_incremental(data_new_id)
self.assertTrue('Got valid id in fit_incremental and invalid id in fit.'
in str(context.exception))
def test_tcmf_ndarray_covariates_dti(self):
ndarray_input = {'id': self.id, 'y': self.data}
self.model.fit(ndarray_input,
covariates=np.random.rand(3, self.seq_len),
dti=pd.date_range('20130101', periods=self.seq_len),
**self.fit_params)
future_covariates = np.random.randn(3, self.horizon)
future_dti = pd.date_range('20130101', periods=self.horizon)
# test predict
yhat = self.model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
# test save load
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat_loaded = loaded_model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
# test evaluate
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value,
target_covariates=future_covariates,
target_dti=future_dti,
metric=['mse'])
# test fit_incremental
self.model.fit_incremental({'y': self.data_new},
covariates_incr=future_covariates,
dti_incr=future_dti,)
yhat_incr = self.model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
class TestChronosModelTCMFForecaster(TestCase):
def setUp(self):
self.model = TCMFForecaster()
self.num_samples = 300
self.horizon = np.random.randint(1, 50)
self.seq_len = 480
self.data = np.random.rand(self.num_samples, self.seq_len)
self.id = np.arange(self.num_samples)
self.data_new = np.random.rand(self.num_samples, self.horizon)
self.fit_params = dict(val_len=12,
start_date="2020-1-1",
freq="5min",
y_iters=1,
init_FX_epoch=1,
max_FX_epoch=1,
max_TCN_epoch=1,
alt_iters=2)
def test_forecast_tcmf_ndarray(self):
ndarray_input = {'id': self.id, 'y': self.data}
self.model.fit(ndarray_input, **self.fit_params)
assert not self.model.is_xshards_distributed()
# test predict
yhat = self.model.predict(horizon=self.horizon)
# test save load
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat_loaded = loaded_model.predict(horizon=self.horizon)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
# test evaluate
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value, metric=['mse'])
# test fit_incremental
self.model.fit_incremental({'y': self.data_new}) # 1st time
self.model.fit_incremental({'y': self.data_new}) # 2nd time
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
def test_tcmf_ndarray_covariates_dti(self):
ndarray_input = {'id': self.id, 'y': self.data}
self.model.fit(ndarray_input,
covariates=np.random.rand(3, self.seq_len),
dti=pd.date_range('20130101', periods=self.seq_len),
**self.fit_params)
future_covariates = np.random.randn(3, self.horizon)
future_dti = pd.date_range('20130101', periods=self.horizon)
# test predict
yhat = self.model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
# test save load
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat_loaded = loaded_model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
# test evaluate
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value,
target_covariates=future_covariates,
target_dti=future_dti,
metric=['mse'])
# test fit_incremental
self.model.fit_incremental({'y': self.data_new},
covariates_incr=future_covariates,
dti_incr=future_dti,)
yhat_incr = self.model.predict(horizon=self.horizon,
future_covariates=future_covariates,
future_dti=future_dti,
)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
def test_forecast_ndarray_error(self):
# is_xshards_distributed
with self.assertRaises(Exception) as context:
self.model.is_xshards_distributed()
self.assertTrue('You should run fit before calling is_xshards_distributed()'
in str(context.exception))
# fit
input = dict({'data': self.data})
with self.assertRaises(Exception) as context:
self.model.fit(input)
self.assertTrue("key `y` doesn't exist in x" in str(context.exception))
input = dict({'y': "abc"})
with self.assertRaises(Exception) as context:
self.model.fit(input)
self.assertTrue("the value of y should be an ndarray" in str(context.exception))
id_diff = np.arange(200)
input = dict({'id': id_diff, 'y': self.data})
with self.assertRaises(Exception) as context:
self.model.fit(input)
self.assertTrue("the length of the id array should be equal to the number of"
in str(context.exception))
input_right = dict({'id': self.id, 'y': self.data})
self.model.fit(input_right, **self.fit_params)
with self.assertRaises(Exception) as context:
self.model.fit(input_right)
self.assertTrue('This model has already been fully trained' in str(context.exception))
# fit_incremental
data_id_diff = {'id': self.id - 1, 'y': self.data_new}
with self.assertRaises(ValueError) as context:
self.model.fit_incremental(data_id_diff)
self.assertTrue('The input ids in fit_incremental differs from input ids in fit'
in str(context.exception))
# evaluate
target_value_fake = dict({"data": self.data_new})
with self.assertRaises(Exception) as context:
self.model.evaluate(target_value=target_value_fake, metric=['mse'])
self.assertTrue("key `y` doesn't exist in x" in str(context.exception))
def test_forecast_tcmf_without_id(self):
# construct data
input = dict({'y': self.data})
self.model.fit(input, **self.fit_params)
assert not self.model.is_xshards_distributed()
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat = self.model.predict(horizon=self.horizon)
yhat_loaded = loaded_model.predict(horizon=self.horizon)
assert "id" not in yhat_loaded
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_array_almost_equal(yhat, yhat_loaded, decimal=4)
target_value = dict({"y": self.data_new})
self.model.evaluate(target_value=target_value, metric=['mse'])
self.model.fit_incremental({'y': self.data_new}) # 1st time
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
data_new_id = {'id': self.id, 'y': self.data_new}
with self.assertRaises(ValueError) as context:
self.model.fit_incremental(data_new_id)
self.assertTrue('Got valid id in fit_incremental and invalid id in fit.'
in str(context.exception))
def test_forecast_tcmf_xshards(self):
from zoo.orca import OrcaContext
import zoo.orca.data.pandas
import pandas as pd
OrcaContext.pandas_read_backend = "pandas"
def preprocessing(df, id_name, y_name):
id = df.index
data = df.to_numpy()
result = dict({id_name: id, y_name: data})
return result
def postprocessing(pred_results, output_dt_col_name):
id_arr = pred_results["id"]
pred_results = pred_results["prediction"]
pred_results = np.concatenate((np.expand_dims(id_arr, axis=1), pred_results), axis=1)
final_df = pd.DataFrame(pred_results, columns=["id"] + output_dt_col_name)
final_df.id = final_df.id.astype("int")
final_df = final_df.set_index("id")
final_df.columns.name = "datetime"
final_df = final_df.unstack().reset_index().rename({0: "prediction"}, axis=1)
return final_df
def get_pred(d):
return d["prediction"]
with tempfile.NamedTemporaryFile() as temp:
data = np.random.rand(300, 480)
df = pd.DataFrame(data)
df.to_csv(temp.name)
shard = zoo.orca.data.pandas.read_csv(temp.name)
shard.cache()
shard_train = shard.transform_shard(preprocessing, 'id', 'data')
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue("key `y` doesn't exist in x" in str(context.exception))
shard_train = shard.transform_shard(preprocessing, 'cid', 'y')
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue("key `id` doesn't exist in x" in str(context.exception))
with self.assertRaises(Exception) as context:
self.model.is_xshards_distributed()
self.assertTrue('You should run fit before calling is_xshards_distributed()'
in str(context.exception))
shard_train = shard.transform_shard(preprocessing, 'id', 'y')
self.model.fit(shard_train, **self.fit_params)
assert self.model.is_xshards_distributed()
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue('This model has already been fully trained' in str(context.exception))
with self.assertRaises(Exception) as context:
self.model.fit_incremental(shard_train)
self.assertTrue('NotImplementedError' in context.exception.__class__.__name__)
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname + "/model")
loaded_model = TCMFForecaster.load(tempdirname + "/model", is_xshards_distributed=True)
horizon = np.random.randint(1, 50)
yhat_shard_origin = self.model.predict(horizon=horizon)
yhat_list_origin = yhat_shard_origin.collect()
yhat_list_origin = list(map(get_pred, yhat_list_origin))
yhat_shard = loaded_model.predict(horizon=horizon)
yhat_list = yhat_shard.collect()
yhat_list = list(map(get_pred, yhat_list))
yhat_origin = np.concatenate(yhat_list_origin)
yhat = np.concatenate(yhat_list)
assert yhat.shape == (300, horizon)
np.testing.assert_equal(yhat, yhat_origin)
output_dt_col_name = pd.date_range(start='2020-05-01', periods=horizon, freq='H').to_list()
yhat_df_shards = yhat_shard.transform_shard(postprocessing, output_dt_col_name)
final_df_list = yhat_df_shards.collect()
final_df = pd.concat(final_df_list)
final_df.sort_values("datetime", inplace=True)
assert final_df.shape == (300 * horizon, 3)
OrcaContext.pandas_read_backend = "spark"
def test_forecast_tcmf_distributed(self):
input = dict({'id': self.id, 'y': self.data})
from zoo.orca import init_orca_context, stop_orca_context
init_orca_context(cores=4, spark_log_level="INFO", init_ray_on_spark=True,
object_store_memory="1g")
self.model.fit(input, num_workers=4, **self.fit_params)
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname)
loaded_model = TCMFForecaster.load(tempdirname, is_xshards_distributed=False)
yhat = self.model.predict(horizon=self.horizon, num_workers=4)
yhat_loaded = loaded_model.predict(horizon=self.horizon, num_workers=4)
yhat_id = yhat_loaded["id"]
np.testing.assert_equal(yhat_id, self.id)
yhat = yhat["prediction"]
yhat_loaded = yhat_loaded["prediction"]
assert yhat.shape == (self.num_samples, self.horizon)
np.testing.assert_equal(yhat, yhat_loaded)
self.model.fit_incremental({'y': self.data_new})
yhat_incr = self.model.predict(horizon=self.horizon)
yhat_incr = yhat_incr["prediction"]
assert yhat_incr.shape == (self.num_samples, self.horizon)
np.testing.assert_raises(AssertionError, np.testing.assert_array_equal, yhat, yhat_incr)
target_value = dict({"y": self.data_new})
assert self.model.evaluate(target_value=target_value, metric=['mse'])
stop_orca_context()
def test_forecast_tcmf_xshards(self):
from zoo.orca import OrcaContext
import zoo.orca.data.pandas
import pandas as pd
OrcaContext.pandas_read_backend = "pandas"
def preprocessing(df, id_name, y_name):
id = df.index
data = df.to_numpy()
result = dict({id_name: id, y_name: data})
return result
def postprocessing(pred_results, output_dt_col_name):
id_arr = pred_results["id"]
pred_results = pred_results["prediction"]
pred_results = np.concatenate((np.expand_dims(id_arr, axis=1), pred_results), axis=1)
final_df = pd.DataFrame(pred_results, columns=["id"] + output_dt_col_name)
final_df.id = final_df.id.astype("int")
final_df = final_df.set_index("id")
final_df.columns.name = "datetime"
final_df = final_df.unstack().reset_index().rename({0: "prediction"}, axis=1)
return final_df
def get_pred(d):
return d["prediction"]
with tempfile.NamedTemporaryFile() as temp:
data = np.random.rand(300, 480)
df = pd.DataFrame(data)
df.to_csv(temp.name)
shard = zoo.orca.data.pandas.read_csv(temp.name)
shard.cache()
shard_train = shard.transform_shard(preprocessing, 'id', 'data')
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue("key `y` doesn't exist in x" in str(context.exception))
shard_train = shard.transform_shard(preprocessing, 'cid', 'y')
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue("key `id` doesn't exist in x" in str(context.exception))
with self.assertRaises(Exception) as context:
self.model.is_xshards_distributed()
self.assertTrue('You should run fit before calling is_xshards_distributed()'
in str(context.exception))
shard_train = shard.transform_shard(preprocessing, 'id', 'y')
self.model.fit(shard_train, **self.fit_params)
assert self.model.is_xshards_distributed()
with self.assertRaises(Exception) as context:
self.model.fit(shard_train)
self.assertTrue('This model has already been fully trained' in str(context.exception))
with self.assertRaises(Exception) as context:
self.model.fit_incremental(shard_train)
self.assertTrue('NotImplementedError' in context.exception.__class__.__name__)
with tempfile.TemporaryDirectory() as tempdirname:
self.model.save(tempdirname + "/model")
loaded_model = TCMFForecaster.load(tempdirname + "/model", is_xshards_distributed=True)
horizon = np.random.randint(1, 50)
yhat_shard_origin = self.model.predict(horizon=horizon)
yhat_list_origin = yhat_shard_origin.collect()
yhat_list_origin = list(map(get_pred, yhat_list_origin))
yhat_shard = loaded_model.predict(horizon=horizon)
yhat_list = yhat_shard.collect()
yhat_list = list(map(get_pred, yhat_list))
yhat_origin = np.concatenate(yhat_list_origin)
yhat = np.concatenate(yhat_list)
assert yhat.shape == (300, horizon)
np.testing.assert_equal(yhat, yhat_origin)
output_dt_col_name = pd.date_range(start='2020-05-01', periods=horizon, freq='H').to_list()
yhat_df_shards = yhat_shard.transform_shard(postprocessing, output_dt_col_name)
final_df_list = yhat_df_shards.collect()
final_df = pd.concat(final_df_list)
final_df.sort_values("datetime", inplace=True)
assert final_df.shape == (300 * horizon, 3)
OrcaContext.pandas_read_backend = "spark"
args = parser.parse_args()
num_nodes = 1 if args.cluster_mode == "local" else args.num_workers
init_orca_context(cluster_mode=args.cluster_mode, cores=args.cores, num_nodes=num_nodes,
memory=args.memory, init_ray_on_spark=True)
if not args.use_dummy_data:
assert args.data_dir is not None, "--data_dir must be provided if not using dummy data"
logger.info('Initalizing TCMFForecaster.')
model = TCMFForecaster(
vbsize=128,
hbsize=256,
num_channels_X=[32, 32, 32, 32, 32, 1],
num_channels_Y=[32, 32, 32, 32, 32, 1],
kernel_size=7,
dropout=0.2,
rank=64,
kernel_size_Y=7,
learning_rate=0.0005,
normalize=False,
use_time=True,
svd=True,
)
ymat = np.load(args.data_dir) if not args.use_dummy_data else get_dummy_data()
horizon = 24
train_data = ymat[:, : -2 * horizon]
target_data = ymat[:, -2 * horizon: -horizon]
incr_target_data = ymat[:, -horizon:]
logger.info('Start fitting.')
model.fit({'y': train_data},
val_len=24,