columns

Column-oriented transformers for pandas providers.

type pysatl_cpd.data.providers.transformers.columns.ColumnSelector = str | Sequence[str]
class pysatl_cpd.data.providers.transformers.columns.ColumnsSelectorTransformer(columns, *, rename_provider=False)[source]

Bases: IDataTransformer[PandasProviderT, PandasProviderT]

Transformer that selects specific columns from labeled data.

Parameters:

  • columns (str | Sequence[str]) – Column name or sequence of column names to select.

  • rename_provider (bool) – Whether to update the provider annotation name.

__init__(columns, *, rename_provider=False)[source]
Parameters:
Return type:

None

property annotation: str

Transformer annotation string.

Returns:

Annotation describing selected columns.

Return type:

annotation

transform(provider)[source]

Apply column selection to provider.

Parameters:

provider (TypeVar(PandasProviderT, PandasDataProvider[Any], PandasLabeledData[Any])) – Input provider to transform.

Returns:

Provider with selected columns.

Return type:

TypeVar(PandasProviderT, PandasDataProvider[Any], PandasLabeledData[Any])

Raises:

TypeError – If provider is not a PandasLabeledData.

class pysatl_cpd.data.providers.transformers.columns.ColumnFeatureCreator(name, mapping, *, rename_provider=False)[source]

Bases: IDataTransformer[PandasProviderT, PandasProviderT]

Transformer that appends a derived feature column.

Parameters:

  • name (str) – Name of the derived feature column.

  • mapping (Any) – Callable applied to each pandas row.

  • rename_provider (bool) – Whether to update the provider annotation name.

__init__(name, mapping, *, rename_provider=False)[source]
Parameters:
Return type:

None

property annotation: str

Transformer annotation string.

transform(provider)[source]

Append a derived feature column to a pandas provider.

Parameters:

provider (TypeVar(PandasProviderT, PandasDataProvider[Any], PandasLabeledData[Any])) – Input pandas provider to transform.

Returns:

Provider with the appended feature column.

Return type:

TypeVar(PandasProviderT, PandasDataProvider[Any], PandasLabeledData[Any])

Raises:

  • TypeError – If provider is not a PandasDataProvider or PandasLabeledData, or if mapping is not callable.

  • ValueError – If the feature column name is empty.