KeyError: nan - CategoricalTransformer fails on numerical + nan data only

[csala]

• Reversible Data Transforms version: 0.2.8

Description

When the CategoricalTransformer is used with data that contains only numerical data only NaN values, it crashes with a KeyError when trying to transform the NaNs.

What I Did

In [1]: import numpy as np

In [2]: import pandas as pd

In [3]: data = pd.Series([1, 2, np.nan])

In [4]: from rdt.transformers import CategoricalTransformer

In [5]: ct = CategoricalTransformer()

In [6]: ct.fit_transform(data)
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-6-9f958bc8253f> in <module>
----> 1 ct.fit_transform(data)

~/Projects/MIT/RDT/rdt/transformers/base.py in fit_transform(self, data)
     44         """
     45         self.fit(data)
---> 46         return self.transform(data)
     47 
     48     def reverse_transform(self, data):

~/Projects/MIT/RDT/rdt/transformers/categorical.py in transform(self, data)
    172             data = data.map(MAPS[id(self)])
    173 
--> 174         return data.fillna(np.nan).apply(self._get_value).to_numpy()
    175 
    176     def _normalize(self, data):

~/.virtualenvs/RDT/lib/python3.8/site-packages/pandas/core/series.py in apply(self, func, convert_dtype, args, **kwds)
   4210             else:
   4211                 values = self.astype(object)._values
-> 4212                 mapped = lib.map_infer(values, f, convert=convert_dtype)
   4213 
   4214         if len(mapped) and isinstance(mapped[0], Series):

pandas/_libs/lib.pyx in pandas._libs.lib.map_infer()

~/Projects/MIT/RDT/rdt/transformers/categorical.py in _get_value(self, category)
    145     def _get_value(self, category):
    146         """Get the value that represents this category."""
--> 147         mean, std = self.intervals[category][2:]
    148         if self.fuzzy:
    149             return norm.rvs(mean, std)

KeyError: nan