Multiple reproducibility issues

[cottrell]

This is easily fixable but a bit frustrating that someone didn't just run pip freeze.

Traceback (most recent call last):
  File "run.py", line 12, in <module>
    from npt.column_encoding_dataset import ColumnEncodingDataset
  File "/home/cottrell/dev/non-parametric-transformers/npt/column_encoding_dataset.py", line 10, in <module>
    from npt.datasets.boston_housing import BostonHousingDataset
  File "/home/cottrell/dev/non-parametric-transformers/npt/datasets/boston_housing.py", line 2, in <module>
    from sklearn.datasets import load_boston
  File "/home/cottrell/anaconda3/envs/npt/lib/python3.8/site-packages/sklearn/datasets/__init__.py", line 157, in __getattr__
    raise ImportError(msg)
ImportError: 
`load_boston` has been removed from scikit-learn since version 1.2.

The documentation refers to what appear to be non-existant entities like:

$ git grep data_loaders
npt/model/npt.py:    expected to provide datasets as given in `npt.data_loaders`.)

The paper description appears to not entirely be consistent with the code, at least not obviously so.