Version 2.0.0

• Big Change: New Dataset class

  • This new Dataset class is designed to fully replace the old system of loading via the BPt_ML class.

  • This new system is designed to be more flexible than the old system.

  • Most of the old functionality is ported over, and various new functionality added.

• Big Change: New evaluate method

  • Instead of previous BPt_ML.Evaluate, now various function are provided instead e.g., evaluate()

  • Introduce new returned flexible class from evaluate() called BPtEvaluator.

• Updates to the project documentation

  • With replacing the BPt_ML class, large amounts of the documentation was replaced and updated.

  • Moved to new pandas style documentation.

• Class name changes

  • Input classes with _’s have been replaced with versions without underscores, e.g., Problem_Spec to ProblemSpec.

• Better support with sklearn

  • Now a number of functions can be passed also in native scikit-learn style syntax.

• More tests

  • More tests across all pieces of the project.

  • Changed from nosetests to pytest.

  • Added github actions CI to automatically run tests.

• Removed Values_Subset class

  • Now the ValueSubset class replaces this functionality, and more generally the feature of creating overlaps of column values on the fly is depreciated.

• Removed AutoGluon support

  • Too difficult to manage dependencies, removed for now this previous feature.

• More!

  • Given the scope of this update, not all of the changes and fixes can be listed here.