Utility functions#

skpro has a number of modules dedicated to utilities:

skpro.datatypes, which contains utilities for data format checks and conversion.
skpro.registry, which contains utilities for estimator and tag search
skpro.utils, which contains generic utility functions.

Data Format Checking and Conversion#

Module exports: data type definitions, checks, validation, fixtures, converters.

`convert_to`(obj, to_type[, as_scitype, ...])	Convert object to a different machine representation, subject to scitype.
`convert`(obj, from_type, to_type[, ...])	Convert objects between different machine representations, subject to scitype.
`check_raise`(obj, mtype[, scitype, var_name])	Check object for compliance with mtype specification, raise errors.
`check_is_mtype`(obj, mtype[, scitype, ...])	Check object for compliance with mtype specification, return metadata.
`check_is_scitype`(obj, scitype[, ...])	Check object for compliance with scitype specification, return metadata.
`mtype`(obj[, as_scitype, exclude_mtypes])	Infer the mtype of an object considered as a specific scitype.
`scitype`(obj[, candidate_scitypes, ...])	Infer the scitype of an object.
`mtype_to_scitype`(mtype[, return_unique, ...])	Infer scitype belonging to mtype.
`scitype_to_mtype`(scitype[, softdeps])	Return list of all mtypes belonging to scitype.

Registry and lookup functionality.

`all_objects`([object_types, filter_tags, ...])	Get a list of all objects from skpro.
`all_tags`([object_types, as_dataframe])	Get a list of all tags from skpro.
`check_tag_is_valid`(tag_name, tag_value)	Check validity of a tag value.

Utility functions for plotting.

`plot_crossplot_interval`(y_true, y_pred[, ...])	Probabilistic cross-plot for regression, truth vs prediction interval.
`plot_crossplot_std`(y_true, y_pred[, ax])	Probabilistic cross-plot for regression, error vs predictive standard deviation.
`plot_crossplot_loss`(y_true, y_pred, metric)	Cross-loss-plot for probabilistic regression.

Estimator checker for extension.

check_estimator(estimator[, ...])

Run all tests on one single estimator.