https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/preprocessing/_discretization.py#L234-L270 I think we should call `self._encoder.transform` instead of `self._encoder.fit_transform`.