Hisel

This is the second release of `hisel`.

It contains the feature selection workflow from pandas dataframes to the selection results. In the workflow, the user can decide to pre-filter features using `sklearn.feature_selection.mutual_info_regression`.

This is the first release of `hisel`.
It provides the basic API to perform supervised feature selection through `hisel.select.Selector`.
The usage is demonstrated in the `notebooks/` of the [repository](https://github.com/transferwise/hisel).
The release includes two types of kernels, `KerneType.RBF` and `KernelType.DELTA`, that are used respectively with `FeatureType.CONT` and `FeatureType.DISCR`.