High-level primitives base classes¶
High-level primitives base classes provides tools to the developers to easily create new primitives by abstracting some unnecessary and repetitive work.
Primitives base classes¶
FileReaderPrimitiveBase: A primitive base class for reading files referenced in columns.
DatasetSplitPrimitiveBase: A base class for primitives which fit on a
Dataset object to produce splits of that
Dataset when producing.
TabularSplitPrimitiveBase: A primitive base class for splitting tabular datasets.
Examples of primitives using these base classes can be found in this repository:
A primitive which reads columns referencing image files.
FixedSplitDatasetSplitPrimitive A primitive which splits a tabular Dataset in a way that uses for the test (score) split a fixed list of primary index values or row indices of the main resource to be used. All other rows are added used for the train split.
KFoldDatasetSplitPrimitive A primitive which splits a tabular Dataset for k-fold cross-validation.
KFoldTimeSeriesSplitPrimitive A primitive which splits a tabular time-series Dataset for k-fold cross-validation.
NoSplitDatasetSplitPrimitive A primitive which splits a tabular Dataset in a way that for all splits it produces the same (full) Dataset.
TrainScoreDatasetSplitPrimitive A primitive which splits a tabular Dataset into random train and score subsets.