OneHotEncoder doesn't support dtype 'category'

[npatki]

Environment Details

• RDT version: 1.9.0

Error Description

The OneHotEncoder transformer crashes if the dtype is 'category' similar to the LabelEncoder as described in #617. Certain synthesizers -- such as CTGANSynthesizer and TVAESynthesizer -- use One Hot Encoding, so this transformer error prevents them from being able to model this type of data.

Steps to reproduce

import pandas as pd
from rdt.transformers.categorical import OneHotEncoder

# create some discrete data and store is as type 'category'
test_data = pd.DataFrame(data={
    'A': ['Yes', 'No', 'Yes', 'Maybe', 'No']
})
test_data['A'] = test_data['A'].astype('category')

# try to use one hot encoder
transformer = OneHotEncoder()
transformed_data = transformer.fit_transform(test_data, column='A')

Output:

TypeError: Cannot interpret 'CategoricalDtype(categories=['Maybe', 'No', 'Yes'], ordered=False)' as a data type

Stack Trace:
stack_trace.txt