Null values not behaving as expected in _get_intervals method

[fealho]

Below is the code of the _get_intervals method:

RDT/rdt/transformers/categorical.py

Lines 92 to 108 in adc2b1a

	data = data.fillna(np.nan)
	frequencies = data.value_counts(dropna=False)
	start = 0
	end = 0
	elements = len(data)
	intervals = {}
	means = []
	starts = []
	for value, frequency in frequencies.items():
	prob = frequency / elements
	end = start + prob
	mean = start + prob / 2
	std = prob / 6
	if pd.isna(value):
	value = np.nan

Since we start by converting all null data to np.nan in line 92, we would expect lines 107-108 to be unnecessary. If we remove them, however, we run into multiple errors, which can be found in this branch, where the change above was implemented.

We should figure out why this error is happening, to make sure the code works as intended.