discontinuum.data_manager

discontinuum.data_manager#

Data preprocessing utilities.

Functions

is_initialized(func)

Decorator checks whether model has been fit.

Classes

Data(target, covariates[, target_unc])

DataManager(target_pipeline, error_pipeline, ...)

class discontinuum.data_manager.Data(target: 'Dataset', covariates: 'Dataset', target_unc: 'Dataset' = None)#
class discontinuum.data_manager.DataManager(target_pipeline: 'Type[Pipeline]' = <class 'discontinuum.pipeline.LogStandardPipeline'>, error_pipeline: 'Type[Pipeline]' = <class 'discontinuum.pipeline.LogErrorPipeline'>, covariate_pipelines: 'Dict[str, Pipeline]' = None)#
property X: ArrayLike#

Convenience function for DataManager.covariates.transform

Xnew(ds: Dataset) → ArrayLike#

Convenience function for DataManager.covariates.transform

error_pipeline#

alias of LogErrorPipeline

fit(target: Dataset, covariates: Dataset, target_unc: Dataset = None)#

Initialize DataManager for a given data distribution.

Parameters:

  • target (Dataset) – Target data.

  • covariates (Dataset) – Covariate data.

  • target_unc (Dataset) – Target uncertainty. Default is None.

get_dim(dim: str) → int#

Get the dimension of a variable.

In other words, its column in the design matrix.

Parameters:

dim (str) – Dimension name.

Returns:

Dimension (column) in design matrix.

Return type:

int

inverse_transform_covariates(X: ArrayLike) → Dataset#

Inverse transform design matrix into covariates

target_pipeline#

alias of LogStandardPipeline

transform_covariates(covariates: Dataset) → ArrayLike#

Transform covariates into design matrix

property y: ArrayLike#

Convenience function for DataManager.target.transform

y_t(y: ArrayLike) → Dataset#

Convenience function for DataManager.target.untransform

property y_unc: ArrayLike#

Convenience function for DataManager.target.transform

discontinuum.data_manager.is_initialized(func)#

Decorator checks whether model has been fit.

Contents