v0.1.1 (June 3, 2022)#

This is a minor release with some bug fixes and adaptations to changes in watex dependencies. There are also several new features. This release is the beginning of writing the project to fit the common open-source library design in the literature. The release v0.1.1 can be considered as the starting point of the watex project design.

New features#

Implement a single reducers Reducers class focused on the unsupervised methods dimensionality reduction ( PCA, kPCA, LLE and iPCA).
Improve the transformer module (watex.transformers) by adding the watex.transformers.StratifiedUsingBaseCategory for auto-category stratification.
Improved the watex.cases.BaseSteps by adding the default pipeline watex.cases.prepare.default_pipeline() for processing the data out of the class in any other datasets.
Rewritten the watex.cases module to be a boilerplate module for any kind of dataset that objective is to predict the groundwater flow rate from the dedicated predictor [X and y].
Designed the watex.methods.electrical module to supplement the watex.methods.erp module that is purely focused on the auto-detection. The new module watex.methods.electrical is meaningful, efficient, and complete by henceforth implementing the constraints which are mostly encountered during the CDWS project and affected the drilling location positioning. An example of constraints can be found in the Case-history in Bagoue-region paper.

Bug fixes and adaptations#

Hashed the train and test split using the id methods to avoid an error when singular data is passed to BaseSteps.stratifydata(), issuing a warning instead.
Fixed a bug where values would be excluded from categorical plots from watex.view.EvalPlot when only y was a pandas Series with a non-default index.
Fixed a bug when passing a norm object and using color annotations in the pobj alias of watex.property.BasePlot.
Fixed a bug where data were not rearranged to match the X data in watex.cases.prepare.base_transform().
Standardized the output format of custom color palette functions.
Fixed a bug where legends for numerical variables in a relational plot could show a surprisingly large number of decimal places.
Improved robustness to missing values in watex.view.QuickPlot distribution plots.