Merge branch 'master' into feat/pytorch_lightning

darts/models/forecasting/block_rnn_model.py

@@ -3,10 +3,10 @@
 -------------------------------
 """
 
-import torch.nn as nn
-import torch
+from typing import List, Optional, Tuple, Union
 
-from typing import List, Optional, Union, Tuple
+import torch
+import torch.nn as nn
 
 from darts.logging import raise_if_not, get_logger
 from darts.models.forecasting.pl_forecasting_module import PLPastCovariatesModule
@@ -76,7 +76,7 @@ def __init__(
             Tensor containing the prediction at the last time step of the sequence.
         """
 
-        super(_BlockRNNModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/models/forecasting/forecasting_model.py

@@ -12,31 +12,16 @@
 to obtain forecasts for a desired number of time stamps into the future.
 """
 import copy
-from collections import OrderedDict
-from typing import Optional, Tuple, Union, Any, Callable, Dict, List, Sequence
-from itertools import product
+import inspect
+import time
 from abc import ABC, ABCMeta, abstractmethod
+from collections import OrderedDict
 from itertools import product
 from random import sample
 from typing import Any, Callable, Dict, List, Optional, Sequence, Tuple, Union
 
 import numpy as np
 import pandas as pd
-import time
-
-from darts.timeseries import TimeSeries
-from darts.logging import get_logger, raise_log, raise_if_not, raise_if
-from darts.utils import (
-    _build_tqdm_iterator,
-    _with_sanity_checks,
-    _historical_forecasts_general_checks,
-    _parallel_apply,
-)
-from darts.utils.timeseries_generation import (
-    _generate_new_dates,
-    _build_forecast_series,
-)
-import inspect
 
 from darts import metrics
 from darts.logging import get_logger, raise_if, raise_if_not, raise_log
@@ -47,7 +32,10 @@
     _parallel_apply,
     _with_sanity_checks,
 )
-from darts.utils.timeseries_generation import _generate_index
+from darts.utils.timeseries_generation import (
+    _build_forecast_series,
+    _generate_new_dates,
+)
 
 logger = get_logger(__name__)
 
@@ -101,7 +89,7 @@ def __call__(cls, *args, **kwargs):
         cls._model_call = all_params
 
         # 6) call model
-        return super(ModelMeta, cls).__call__(**all_params)
+        return super().__call__(**all_params)
 
 
 class ForecastingModel(ABC, metaclass=ModelMeta):

darts/models/forecasting/nbeats.py

@@ -3,14 +3,12 @@
 -------
 """
 
-from typing import NewType, Union, List, Tuple
 from enum import Enum
-from typing import List, NewType, Optional, Tuple, Union
+from typing import List, NewType, Tuple, Union
 
 import numpy as np
 import torch
 import torch.nn as nn
-from numpy.random import RandomState
 
 from darts.logging import get_logger, raise_log, raise_if_not
 from darts.models.forecasting.pl_forecasting_module import PLPastCovariatesModule
@@ -358,7 +356,7 @@ def __init__(
             Tensor containing the output of the NBEATS module.
 
         """
-        super(_NBEATSModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/models/forecasting/rnn_model.py

@@ -3,14 +3,15 @@
 -------------------------
 """
 
-import torch.nn as nn
+from typing import Optional, Sequence, Tuple, Union
+
 import torch
-from typing import Sequence, Optional, Union, Tuple
-from darts.timeseries import TimeSeries
+import torch.nn as nn
 
 from darts.logging import raise_if_not, get_logger
 from darts.models.forecasting.pl_forecasting_module import PLDualCovariatesModule
 from darts.models.forecasting.torch_forecasting_model import DualCovariatesTorchModel
+from darts.timeseries import TimeSeries
 from darts.utils.data import DualCovariatesShiftedDataset, TrainingDataset
 
 logger = get_logger(__name__)
@@ -70,7 +71,7 @@ def __init__(
         """
 
         # RNNModule doesn't really need input and output_chunk_length for PLModule
-        super(_RNNModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/models/forecasting/tcn_model.py

@@ -4,18 +4,17 @@
 """
 
 import math
-from typing import Optional, Sequence, Tuple, Union
+from typing import Optional, Sequence, Tuple
 
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-from typing import Optional, Sequence, Tuple
-from darts.timeseries import TimeSeries
-from darts.utils.data import PastCovariatesShiftedDataset
 
-from darts.logging import raise_if_not, get_logger
+from darts.logging import get_logger, raise_if_not
 from darts.models.forecasting.pl_forecasting_module import PLPastCovariatesModule
 from darts.models.forecasting.torch_forecasting_model import PastCovariatesTorchModel
+from darts.timeseries import TimeSeries
+from darts.utils.data import PastCovariatesShiftedDataset
 
 logger = get_logger(__name__)
 
@@ -182,7 +181,7 @@ def __init__(
             leading up to the first prediction, all in chronological order.
         """
 
-        super(_TCNModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/models/forecasting/tft_model.py

@@ -6,42 +6,27 @@
 from typing import Dict, List, Optional, Sequence, Tuple, Union
 
 import numpy as np
-
 import torch
-from numpy.random import RandomState
 from torch import nn
 from torch.nn import LSTM as _LSTM
 
 from darts import TimeSeries
-
-from darts.logging import get_logger, raise_if_not, raise_if
-from darts.utils.likelihood_models import QuantileRegression, Likelihood
-from darts.utils.data import (
-    TrainingDataset,
-    MixedCovariatesSequentialDataset,
-    MixedCovariatesTrainingDataset,
-    MixedCovariatesInferenceDataset,
-)
+from darts.logging import get_logger, raise_if, raise_if_not
 from darts.models.forecasting.pl_forecasting_module import PLMixedCovariatesModule
-from darts.models.forecasting.torch_forecasting_model import MixedCovariatesTorchModel
 from darts.models.forecasting.tft_submodels import (
     _GateAddNorm,
     _GatedResidualNetwork,
     _InterpretableMultiHeadAttention,
     _VariableSelectionNetwork,
 )
-from darts.models.forecasting.torch_forecasting_model import (
-    MixedCovariatesTorchModel,
-    TorchParametricProbabilisticForecastingModel,
-)
+from darts.models.forecasting.torch_forecasting_model import MixedCovariatesTorchModel
 from darts.utils.data import (
     MixedCovariatesInferenceDataset,
     MixedCovariatesSequentialDataset,
     MixedCovariatesTrainingDataset,
     TrainingDataset,
 )
 from darts.utils.likelihood_models import Likelihood, QuantileRegression
-from darts.utils.torch import random_method
 
 logger = get_logger(__name__)
 
@@ -101,7 +86,7 @@ def __init__(
             all parameters required for :class:`darts.model.forecasting_models.PLForecastingModule` base class.
         """
 
-        super(_TFTModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/models/forecasting/torch_forecasting_model.py

@@ -17,46 +17,33 @@
       forecasting models.
 """
 
+import datetime
+import inspect
 import os
 import shutil
-import inspect
-import datetime
-import numpy as np
-from glob import glob
+
 from abc import ABC, abstractmethod
+from glob import glob
 from typing import Optional, Dict, Tuple, Union, Sequence, List
 
+import numpy as np
+import pytorch_lightning as pl
 import torch
+from pytorch_lightning import loggers as pl_loggers
 from torch import Tensor
 from torch.utils.data import DataLoader
 
-import pytorch_lightning as pl
-from pytorch_lightning import loggers as pl_loggers
-
-from darts.logging import get_logger, raise_if, raise_if_not, raise_log
-from darts.models.forecasting.forecasting_model import GlobalForecastingModel
-from darts.timeseries import TimeSeries
-from darts.utils.data.encoders import SequentialEncoder
-from darts.utils.likelihood_models import Likelihood
-from darts.utils.torch import random_method
-from darts.models.forecasting.forecasting_model import GlobalForecastingModel
-from darts.models.forecasting.pl_forecasting_module import PLForecastingModule
 from darts.logging import (
-    raise_if_not,
     get_logger,
-    raise_log,
-    raise_if,
     raise_deprecation_warning,
+    raise_if,
+    raise_if_not,
+    raise_log,
     suppress_lightning_warnings,
 )
-from darts.utils.data.training_dataset import (
-    TrainingDataset,
-    PastCovariatesTrainingDataset,
-    FutureCovariatesTrainingDataset,
-    DualCovariatesTrainingDataset,
-    MixedCovariatesTrainingDataset,
-    SplitCovariatesTrainingDataset,
-)
+from darts.models.forecasting.forecasting_model import GlobalForecastingModel
+from darts.models.forecasting.pl_forecasting_module import PLForecastingModule
+from darts.timeseries import TimeSeries
 from darts.utils.data.inference_dataset import (
     DualCovariatesInferenceDataset,
     FutureCovariatesInferenceDataset,
@@ -72,6 +59,17 @@
     PastCovariatesSequentialDataset,
     SplitCovariatesSequentialDataset,
 )
+from darts.utils.data.training_dataset import (
+    DualCovariatesTrainingDataset,
+    FutureCovariatesTrainingDataset,
+    MixedCovariatesTrainingDataset,
+    PastCovariatesTrainingDataset,
+    SplitCovariatesTrainingDataset,
+    TrainingDataset,
+)
+from darts.utils.data.encoders import SequentialEncoder
+from darts.utils.likelihood_models import Likelihood
+from darts.utils.torch import random_method
 
 DEFAULT_DARTS_FOLDER = "darts_logs"
 CHECKPOINTS_FOLDER = "checkpoints"
@@ -401,7 +399,8 @@ def _extract_torch_model_params(**kwargs):
         get_params.remove("self")
         return {kwarg: kwargs.get(kwarg) for kwarg in get_params if kwarg in kwargs}
 
-    def _extract_pl_module_params(self, **kwargs):
+    @staticmethod
+    def _extract_pl_module_params(**kwargs):
         """Extract params from model creation to set up PLForecastingModule (the actual torch.nn.Module)"""
         get_params = list(
             inspect.signature(PLForecastingModule.__init__).parameters.keys()

darts/models/forecasting/transformer_model.py

@@ -4,11 +4,10 @@
 """
 
 import math
-from typing import Optional, Tuple, Union
+from typing import Optional, Tuple
 
 import torch
 import torch.nn as nn
-from typing import Optional, Tuple
 
 from darts.logging import get_logger
 from darts.models.forecasting.pl_forecasting_module import PLPastCovariatesModule
@@ -123,7 +122,7 @@ def __init__(
             Tensor containing the prediction at the last time step of the sequence.
         """
 
-        super(_TransformerModule, self).__init__(**kwargs)
+        super().__init__(**kwargs)
 
         # required for all modules -> saves hparams for checkpoints
         self.save_hyperparameters()

darts/tests/models/forecasting/test_global_forecasting_models.py

@@ -1,6 +1,7 @@
-import numpy as np
 from copy import deepcopy
-from unittest.mock import patch, ANY
+from unittest.mock import ANY, patch
+
+import numpy as np
 
 from darts.dataprocessing.transformers import Scaler
 from darts.datasets import AirPassengersDataset

darts/tests/models/forecasting/test_ptl_trainer.py

@@ -1,10 +1,11 @@
-import numpy as np
-import tempfile
 import shutil
+import tempfile
+
+import numpy as np
 import pytorch_lightning as pl
 
-from darts.tests.base_test_class import DartsBaseTestClass
 from darts.logging import get_logger
+from darts.tests.base_test_class import DartsBaseTestClass
 from darts.utils.timeseries_generation import linear_timeseries
 
 logger = get_logger(__name__)

requirements/torch.txt

@@ -1,2 +1,2 @@
-torch>=1.8.0
 pytorch-lightning>=1.5.0
+torch>=1.8.0