Trains the detector and detects anomalies on the same series. Parameters ---------- series Time series to be used

Read a pickled ``TimeSeries``. Parameters ---------- path : string path pointing to a pickle file that will be lo

Returns causal mask to apply for self-attention layer that acts on future input only. The model will attend to all `False` values.

Extract and return the components of target series, static covariate, past and future covariates series. Parameters ---------- model

Returns the feature importances for the encoder, decoder and static covariates as pd.DataFrames. If multiple series were used in :fun

Helper function called by `get_feature_times` that extracts all of the times within `future_covariates` that can be used to create fe

Helper function called by `get_feature_times` that extracts all times within `future_covariates` that can be used to create features.

Helper function called by `get_feature_times` that extracts all times within `past_covariates` that can be used to create features. M

Helper function called by `get_feature_times` that extracts all times within a `target_series` that can be used to create a feature a

Helper function called by `get_feature_times` that extracts all of the times within `target_series` *or* `past_covariates` that can b

Helper function that returns all the times within `target_series` that can be used to create features for prediction. Since

Helper function that returns all the times within `target_series` that can be used to create features and labels for training.

Returns the G matrix given a specified reconciliation method.

Returns scaled time index relative to prediction point.

Helper function to test that the model raises an error when calling `predict_from_dataset()` after `fit_from_dataset()` with missing inputs.

prediction should be almost equal to y_true. Absolute tolarance is set to 0.2 to give some flexibility

Interval Coverage (IC). IC gives a binary outcome with `1` if the observation is within the interval, and `0` otherwise. For the true series

Recursive function filling S for a given base component and all its ancestors

Inverse transforms a (sequence of) series by calling the user-implemented `ts_inverse_transform` method. In case a sequence or list of lists

Loads a model from a given path or file handle. Parameters ---------- path Path or file handle from whic

Load the model from automatically saved checkpoints under '{work_dir}/darts_logs/{model_name}/checkpoints/'. This method is used for

Load the model weights from a safetensors file. Parameters ---------- module The PyTorch module to load the weigh

Load validation inputs for TimesFM2p5Model fidelity tests. The data imports here are adapted from the `20-SKLearnModel-examples` notebook.

Load validation inputs for TiRexModel fidelity tests.

Load validation inputs for PatchTSTFMModel fidelity tests. The data imports here are adapted from the `20-SKLearnModel-examples` notebook.

Load validation inputs for Chronos2Model fidelity tests. The data imports here are adapted from the `20-SKLearnModel-examples` notebook.

Load only the weights from automatically saved checkpoints under '{work_dir}/darts_logs/{model_name}/ checkpoints/'. This method is u

Mean Interval Coverage (MIC). MIC gives the time-aggregated Interval Coverage :func:`~darts.metrics.metrics.ic` - the ratio of observations b

Context manager to temporarily set options.

Plots the attention heads of the `TFTModel`. Parameters ---------- expl_result A `TFTExplainabilityResult` object

Plots the Cross Correlation Function (CCF) between `ts` and `ts_other`, highlighting it at lag `m`, with corresponding significance interval.

This function plots the histogram of values in a TimeSeries instance or an array-like. All types of TimeSeries are supported (uni-, multivariate,

Plots the Partial Autocorrelation Function (PACF) of `ts`, highlighting it at lag `m`, with corresponding significance interval. Uses :func:`

Plots the variable selection / feature importances of the `TFTModel` based on the input. The figure includes three subplots: - encode

Returns the distribution parameters as a array, extracted from the raw model outputs.

Returns the output of the base Facebook Prophet model in form of a pandas DataFrame. Note however, that the output of this method is not suppo

Filters the given sequence of target time series with the filtering model. Parameters ---------- series The seque

Abstract method to implement the generation of predictions for the input `series`.

Computes the historical forecasts that would have been obtained by the underlying forecasting model on `series`. `retrain` is set to

performs the prediction step batch output of Darts' :class:`TorchInferenceDataset` - tuple of ``(past target, past cov,

Helper function to slice and concatenate the input features. In order to remove the dependency on the `model` argument, it can be decomposed into

Processes the input horizons and target component names. horizons Optionally, an integer or sequence of integers representing the future

Coefficient of Determination :math:`R^2` (see [1]_ for more details). For the true series :math:`y` and predicted series :math:`\\hat{y}` of leng

Creates a random walk univariate TimeSeries, where each step is obtained by sampling a gaussian distribution with mean `mean` and standard de

Resets the model object and removes all stored data - model, checkpoints, loggers and training history.

Reset option(s) to default value(s).

Trims all series in the provided list, if necessary, so that the returned time series have a common span (corresponding to largest time sub-i

Samples a prediction from the likelihood distribution and the predicted parameters.

Samples a prediction from the likelihood distribution and the predicted parameters.

Saves the model under a given path. Creates two files under ``path`` (model object) and ``path``.ckpt (checkpoint). Note: P

Performs a grid search over all hyper parameters to select the best model, using the fitted values on the training series `ts`.

Set the option value with validation.

Set the number of processors to be used by the transformer while processing multiple ``TimeSeries``. Parameters ---------- va

Set the verbosity status. `True` for enabling the detailed report about scaler's operation progress, `False` for no additional inform

Similar to :func:`shap_explanations()`, but computes SHAP explanations for only one forecasted timestamp, which corresponds to the la

Updates the PyTorch Lightning Trainer parameters to move the model to CPU the next time :func:`fit()` or :func:`predict()` is called.

Returns ------- Dense n x m numpy array, where empty cells are set to np.inf

Save the series in pickle format. Parameters ---------- path : string path to a file where current object will be

This method applies transformation to the non-transformed encoded covariates output of `SequentialEncoder._encode_sequence()` after being merg

Returns a tuple of (past transformer, future transformer).

The function that will be applied to each series when :func:`fit` is called. If the `global_fit` attribute is set to `False`, then `ts_fit` s

The function that will be applied to each series when :func:`inverse_transform` is called. The function must take as first argument a ``TimeS

The function that will be applied to each series when :func:`transform()` is called. This method is not implemented in the base class and mus

Returns a new (untrained) model instance created with the same parameters.