state_dataset

Datasets of fixed-state labeled time series.

class pysatl_cpd.data.dataset.state_dataset.StateDataset(timeseries, *, state=None)[source]

Bases: IDataset[DataT, NoChangeSeriesAnnotation], Generic

Dataset of fixed-state series without change points.

All series in this dataset share a common state and have no change points within them. Typically created from a larger Dataset by slicing fixed-state segments.

Parameters:

timeseries (Sequence[LabeledData[TypeVar(DataT), NoChangeSeriesAnnotation]]) – Sequence of labeled data instances sharing the same state.
state (StateDescriptor | None) – Optional explicit state descriptor; derived from the first timeseries if not provided.

__init__(timeseries, *, state=None)[source]

Parameters:

timeseries (Sequence[LabeledData[DataT, NoChangeSeriesAnnotation]])
state (StateDescriptor | None)

Return type:

None

property state: StateDescriptor

Return the fixed state descriptor for this dataset.

Returns:: State descriptor for all timeseries in the dataset.
Return type:: state

classmethod from_dataset(dataset, slice_length, *, state, keep_remainder=False)[source]

Create a StateDataset from a Dataset by filtering and slicing.

Parameters:

dataset (Dataset[TypeVar(DataT), Any]) – Source dataset to extract state segments from.
slice_length (int) – Length of each slice to create.
state (StateDescriptor) – State to filter segments by.
keep_remainder (bool) – Whether to keep remainder as final slice.

Returns:

New StateDataset with sliced timeseries.

Return type:

StateDataset[TypeVar(DataT)]

Raises:

ValueError – If slice_length is not positive, or no segments are found for the given state.