idataset

Base dataset sequence abstractions for labeled time series.

class pysatl_cpd.data.dataset.idataset.IDataset(timeseries)

Bases: Sequence[LabeledData], Generic

Common sequence interface for labeled time series collections.

Parameters:

timeseries (Sequence[LabeledData[TypeVar(DataT), TypeVar(AnnotationT, bound= TimeseriesAnnotation)]]) – Sequence of labeled data instances.

__init__(timeseries)

Parameters:

timeseries (Sequence[LabeledData[DataT, AnnotationT]])

Return type:

None

__getitem__(index)

Return one labeled time series by index.

Parameters:

index (int) – Position of the item to return.

Returns:

Labeled time series stored at the given position.

Return type:

LabeledData[TypeVar(DataT), TypeVar(AnnotationT, bound= TimeseriesAnnotation)]

__iter__()

Iterate over stored labeled time series.

Returns:

Iterator over stored labeled providers.

Return type:

Iterator[LabeledData[TypeVar(DataT), TypeVar(AnnotationT, bound= TimeseriesAnnotation)]]

__len__()

Return the dataset size.

Returns:

Number of labeled time series in the dataset.

Return type:

int

property timeseries: MutableSequence[LabeledData[DataT, AnnotationT]]

Return a copy of the stored labeled providers.

Returns:

Copy of the internal list of labeled data instances.

Return type:

timeseries

property states: set[StateDescriptor]

Return the union of all states from the dataset.

Returns:

Set of all distinct state descriptors across stored series.

Return type:

states

property transitions: set[TransitionDescriptor]

Return the union of all transitions from the dataset.

Returns:

Set of all distinct transition descriptors across stored series.

Return type:

transitions

train_test_split(test_size, random_state=None)

Split the dataset into train and test subsets.

Parameters:

• test_size (float) – Fraction of items to place into the test split.

• random_state (int | None) – Optional random seed for reproducible shuffling.

Returns:

Train and test datasets of the same concrete type.

Return type:

tuple[Self, Self]

Raises:

ValueError – If test_size is not between 0 and 1.

merge()

Merge all stored providers into a single labeled provider.

Returns:

Labeled provider containing all time series in sequence.

Return type:

LabeledData[TypeVar(DataT), TimeseriesAnnotation]

Raises:

ValueError – If the dataset is empty.