Types of changes
Remove deprecated metatransformer
Description
Remove metatransformer submodule and its documentation
Any other comments?
The metatransformer submodule was meant to be a simple sklearn compatible way of generating tda features. However, its transformers are now:
harder to understand: the documentation is large and does no longer match the standard of quality of the rest of the code
harder to maintain: a change in the documentation or in the api of one of its transformer will make them unusable
incompatible with sklearn: they do not return a 2d array and are therefore not usable in standard machine learning pipelines, which makes them pointless
unflexible: They are all based on VietorisRips and work only on point clouds while their name makes the user assume that they can take as arguments any data type.
Types of changes Remove deprecated metatransformer
Description Remove metatransformer submodule and its documentation
Any other comments? The metatransformer submodule was meant to be a simple sklearn compatible way of generating tda features. However, its transformers are now:
Checklist
flake8
to check my Python changes.pytest
to check this on Python tests.