Merge branch 'dev' into precommit

src/gluonts/core/component.py

@@ -296,12 +296,16 @@ def validator(init):
  init_params = inspect.signature(init).parameters
  init_fields = {
  param.name: (
- param.annotation
- if param.annotation != inspect.Parameter.empty
- else Any,
- param.default
- if param.default != inspect.Parameter.empty
- else ...,
+ (
+ param.annotation
+ if param.annotation != inspect.Parameter.empty
+ else Any
+ ),
+ (
+ param.default
+ if param.default != inspect.Parameter.empty
+ else ...
+ ),
  )
  for param in init_params.values()
  if param.name != "self"

src/gluonts/dataset/arrow/file.py

@@ -51,16 +51,13 @@ def infer(
  return ArrowStreamFile(path)
 
  @abc.abstractmethod
- def metadata(self) -> Dict[str, str]:
- ...
+ def metadata(self) -> Dict[str, str]: ...
 
  @abc.abstractmethod
- def __iter__(self):
- ...
+ def __iter__(self): ...
 
  @abc.abstractmethod
- def __len__(self):
- ...
+ def __len__(self): ...
 
 
 @dataclass

src/gluonts/dataset/artificial/_base.py

@@ -112,7 +112,7 @@ def __init__(
  self,
  num_timeseries: int = 10,
  num_steps: int = 30,
- freq: str = "1H",
+ freq: str = "1h",
  start: str = "2000-01-01 00:00:00",
  # Generates constant dataset of 0s with explicit NaN missing values
  is_nan: bool = False,

src/gluonts/dataset/repository/_lstnet.py

@@ -200,9 +200,9 @@ def generate_lstnet_dataset(
  meta = MetaData(
  **metadata(
  cardinality=ds_info.num_series,
- freq=ds_info.freq
- if ds_info.agg_freq is None
- else ds_info.agg_freq,
+ freq=(
+  ds_info.freq if ds_info.agg_freq is None else ds_info.agg_freq
+ ),
  prediction_length=prediction_length or ds_info.prediction_length,
  )
  )

src/gluonts/dataset/stat.py

@@ -418,12 +418,12 @@ def calculate_dataset_statistics(ts_dataset: Any) -> DatasetStatistics:
  max_target_length=max_target_length,
  min_target=min_target,
  num_missing_values=num_missing_values,
- feat_static_real=observed_feat_static_real
- if observed_feat_static_real
- else [],
- feat_static_cat=observed_feat_static_cat
- if observed_feat_static_cat
- else [],
+ feat_static_real=(
+  observed_feat_static_real if observed_feat_static_real else []
+ ),
+ feat_static_cat=(
+  observed_feat_static_cat if observed_feat_static_cat else []
+ ),
  num_past_feat_dynamic_real=num_past_feat_dynamic_real,
  num_feat_dynamic_real=num_feat_dynamic_real,
  num_feat_dynamic_cat=num_feat_dynamic_cat,

src/gluonts/evaluation/_base.py

@@ -378,9 +378,9 @@ def get_base_metrics(
  return {
  "item_id": forecast.item_id,
  "forecast_start": forecast.start_date,
- "MSE": mse(pred_target, mean_fcst)
- if mean_fcst is not None
- else None,
+ "MSE": (
+  mse(pred_target, mean_fcst) if mean_fcst is not None else None
+ ),
  "abs_error": abs_error(pred_target, median_fcst),
  "abs_target_sum": abs_target_sum(pred_target),
  "abs_target_mean": abs_target_mean(pred_target),

src/gluonts/ext/rotbaum/_model.py

@@ -340,11 +340,11 @@ def _get_and_cache_quantile_computation(
  The quantile of the associated true value bin.
  """
  if feature_vector_in_train not in self.quantile_dicts[quantile]:
- self.quantile_dicts[quantile][
- feature_vector_in_train
- ] = np.percentile(
- self.id_to_bins[self.preds_to_id[feature_vector_in_train]],
- quantile * 100,
+ self.quantile_dicts[quantile][feature_vector_in_train] = (
+ np.percentile(
+  self.id_to_bins[self.preds_to_id[feature_vector_in_train]],
+  quantile * 100,
+ )
  )
  return self.quantile_dicts[quantile][feature_vector_in_train]
 

src/gluonts/ext/rotbaum/_predictor.py

@@ -254,13 +254,17 @@ def train(
  target_data[:, train_QRX_only_using_timestep],
  )
  self.model_list = [
- QRX(
- xgboost_params=self.model_params,
- min_bin_size=self.min_bin_size,
- model=self.model_list[train_QRX_only_using_timestep].model,
+ (
+ QRX(
+ xgboost_params=self.model_params,
+ min_bin_size=self.min_bin_size,
+ model=self.model_list[
+ train_QRX_only_using_timestep
+ ].model,
+ )
+ if i != train_QRX_only_using_timestep
+ else self.model_list[i]
  )
- if i != train_QRX_only_using_timestep
- else self.model_list[i]
  for i in range(n_models)
  ]
  with concurrent.futures.ThreadPoolExecutor(

src/gluonts/itertools.py

@@ -42,11 +42,9 @@
 
 @runtime_checkable
 class SizedIterable(Protocol):
- def __len__(self):
- ...
+ def __len__(self): ...
 
- def __iter__(self):
- ...
+ def __iter__(self): ...
 
 
 T = TypeVar("T")

src/gluonts/model/forecast_generator.py

@@ -132,9 +132,11 @@ def __call__(
  yield QuantileForecast(
  output.T,
  start_date=batch[FieldName.FORECAST_START][i],
- item_id=batch[FieldName.ITEM_ID][i]
- if FieldName.ITEM_ID in batch
- else None,
+ item_id=(
+ batch[FieldName.ITEM_ID][i]
+ if FieldName.ITEM_ID in batch
+ else None
+ ),
  info=batch["info"][i] if "info" in batch else None,
  forecast_keys=self.quantiles,
  )
@@ -181,9 +183,11 @@ def __call__(
  yield SampleForecast(
  output,
  start_date=batch[FieldName.FORECAST_START][i],
- item_id=batch[FieldName.ITEM_ID][i]
- if FieldName.ITEM_ID in batch
- else None,
+ item_id=(
+ batch[FieldName.ITEM_ID][i]
+ if FieldName.ITEM_ID in batch
+ else None
+ ),
  info=batch["info"][i] if "info" in batch else None,
  )
  assert i + 1 == len(batch[FieldName.FORECAST_START])
@@ -221,9 +225,11 @@ def __call__(
  yield make_distribution_forecast(
  distr,
  start_date=batch[FieldName.FORECAST_START][i],
- item_id=batch[FieldName.ITEM_ID][i]
- if FieldName.ITEM_ID in batch
- else None,
+ item_id=(
+ batch[FieldName.ITEM_ID][i]
+ if FieldName.ITEM_ID in batch
+ else None
+ ),
  info=batch["info"][i] if "info" in batch else None,
  )
  assert i + 1 == len(batch[FieldName.FORECAST_START])

src/gluonts/mx/batchify.py

@@ -117,13 +117,16 @@ def as_in_context(batch: dict, ctx: mx.Context = None) -> DataBatch:
  Move data into new context, should only be in main process.
  """
  batch = {
- k: v.as_in_context(ctx) if isinstance(v, mx.nd.NDArray)
- # Workaround due to MXNet not being able to handle NDArrays with 0 in
- # shape properly:
- else (
- stack(v, ctx=ctx, dtype=v.dtype, variable_length=False)
- if isinstance(v[0], np.ndarray) and 0 in v[0].shape
- else v
+ k: (
+ v.as_in_context(ctx)
+ if isinstance(v, mx.nd.NDArray)
+ # Workaround due to MXNet not being able to handle NDArrays with 0 in
+ # shape properly:
+ else (
+ stack(v, ctx=ctx, dtype=v.dtype, variable_length=False)
+ if isinstance(v[0], np.ndarray) and 0 in v[0].shape
+ else v
+ )
  )
  for k, v in batch.items()
  }

src/gluonts/mx/block/dropout.py

@@ -231,9 +231,11 @@ def mask(p, like):
  # mask as output, instead of simply copy output to the first element
  # in case that the base cell is ResidualCell
  new_states = [
- F.where(output_mask, next_states[0], states[0])
- if p_outputs != 0.0
- else next_states[0]
+ (
+ F.where(output_mask, next_states[0], states[0])
+ if p_outputs != 0.0
+ else next_states[0]
+ )
  ]
  new_states.extend(
  [

src/gluonts/mx/distribution/box_cox_transform.py

@@ -116,6 +116,7 @@ class BoxCoxTransform(Bijection):
  `tol_lambda_1`
  F
  """
+
  arg_names = ["box_cox.lambda_1", "box_cox.lambda_2"]
 
  @validated()

src/gluonts/mx/distribution/inflated_beta.py

@@ -113,6 +113,7 @@ class ZeroInflatedBeta(ZeroAndOneInflatedBeta):
  `(*batch_shape, *event_shape)`.
  F
  """
+
  is_reparameterizable = False
 
  @validated()
@@ -145,6 +146,7 @@ class OneInflatedBeta(ZeroAndOneInflatedBeta):
  `(*batch_shape, *event_shape)`.
  F
  """
+
  is_reparameterizable = False
 
  @validated()

src/gluonts/mx/distribution/lds.py

@@ -409,9 +409,11 @@ def sample(
  # (num_samples, batch_size, latent_dim, latent_dim)
  # innovation_coeff_t: (num_samples, batch_size, 1, latent_dim)
  emission_coeff_t, transition_coeff_t, innovation_coeff_t = (
- _broadcast_param(coeff, axes=[0], sizes=[num_samples])
- if num_samples is not None
- else coeff
+ (
+ _broadcast_param(coeff, axes=[0], sizes=[num_samples])
+ if num_samples is not None
+ else coeff
+ )
  for coeff in [
  self.emission_coeff[t],
  self.transition_coeff[t],
@@ -458,9 +460,11 @@ def sample(
  if scale is None
  else F.broadcast_mul(
  samples,
- scale.expand_dims(axis=1).expand_dims(axis=0)
- if num_samples is not None
- else scale.expand_dims(axis=1),
+ (
+ scale.expand_dims(axis=1).expand_dims(axis=0)
+ if num_samples is not None
+ else scale.expand_dims(axis=1)
+ ),
  )
  )
 

src/gluonts/mx/distribution/lowrank_gp.py

@@ -101,9 +101,7 @@ def hybrid_forward(self, F, x: Tensor) -> Tuple[Tensor, Tensor, Tensor]:
  D_vector = self.proj[1](x_plus_w).squeeze(axis=-1)
 
  d_bias = (
- 0.0
- if self.sigma_init == 0.0
- else inv_softplus(self.sigma_init**2)
+ 0.0 if self.sigma_init == 0.0 else inv_softplus(self.sigma_init**2)
  )
 
  D_positive = (

src/gluonts/mx/distribution/lowrank_multivariate_gaussian.py

@@ -408,9 +408,7 @@ def domain_map(self, F, mu_vector, D_vector, W_vector=None):
  """
 
  d_bias = (
- inv_softplus(self.sigma_init**2)
- if self.sigma_init > 0.0
- else 0.0
+ inv_softplus(self.sigma_init**2) if self.sigma_init > 0.0 else 0.0
  )
 
  # sigma_minimum helps avoiding cholesky problems, we could also jitter

src/gluonts/mx/model/deepar/_network.py

@@ -572,9 +572,11 @@ def unroll_encoder_imputation(
  static_feat = F.concat(
  embedded_cat,
  feat_static_real,
- F.log(scale)
- if len(self.target_shape) == 0
- else F.log(scale.squeeze(axis=1)),
+ (
+ F.log(scale)
+ if len(self.target_shape) == 0
+ else F.log(scale.squeeze(axis=1))
+ ),
  dim=1,
  )
 
@@ -603,9 +605,9 @@ def unroll_encoder_imputation(
  begin_state = self.rnn.begin_state(
  func=F.zeros,
  dtype=self.dtype,
- batch_size=inputs.shape[0]
- if isinstance(inputs, mx.nd.NDArray)
- else 0,
+ batch_size=(
+  inputs.shape[0] if isinstance(inputs, mx.nd.NDArray) else 0
+ ),
  )
 
  unroll_results = self.imputation_rnn_unroll(
@@ -726,9 +728,11 @@ def unroll_encoder_default(
  static_feat = F.concat(
  embedded_cat,
  feat_static_real,
- F.log(scale)
- if len(self.target_shape) == 0
- else F.log(scale.squeeze(axis=1)),
+ (
+ F.log(scale)
+ if len(self.target_shape) == 0
+ else F.log(scale.squeeze(axis=1))
+ ),
  dim=1,
  )
 
@@ -757,9 +761,9 @@ def unroll_encoder_default(
  begin_state = self.rnn.begin_state(
  func=F.zeros,
  dtype=self.dtype,
- batch_size=inputs.shape[0]
- if isinstance(inputs, mx.nd.NDArray)
- else 0,
+ batch_size=(
+  inputs.shape[0] if isinstance(inputs, mx.nd.NDArray) else 0
+ ),
  )
  state = begin_state
  # This is a dummy computation to avoid deferred initialization error

src/gluonts/mx/model/deepstate/_estimator.py

@@ -69,11 +69,14 @@
 # series in the dataset.
 FREQ_LONGEST_PERIOD_DICT = {
  "M": 12, # yearly seasonality
+ "ME": 12, # yearly seasonality
  "W": 52, # yearly seasonality
  "D": 31, # monthly seasonality
  "B": 22, # monthly seasonality
  "H": 168, # weekly seasonality
+ "h": 168, # weekly seasonality
  "T": 1440, # daily seasonality
+ "min": 1440, # daily seasonality
 }