Allow OrdinalEncoder to have unsorted numeric categories

[tamargrey]

• As a user, I wish I could use the Ordinal Encoder to encode Ordinal columns whose categories' orders are not in strictly increasing numeric order.

The following code raises the error ValueError: Unsorted categories are not supported for numerical categories

    X = pd.DataFrame(
        {
            "col_1": [2, 0, 1, 0, 0],
            "col_2": ["a", "b", "a", "c", "d"],
            "col_3": ["x", "x", "x", "y", "y"],
            "col_4": [1, 2, 2, 3, 1],
        },
    )
 # The order is 3, 2, 1 instead of 1, 2, 3
    X.ww.init(logical_types={"col_2": Ordinal(order=["a", "b", "c", "d"]), "col_4": Ordinal(order=[3, 2, 1])})

    encoder = OrdinalEncoder()
    encoder.fit(X)

This isn't a super uncommon situation where numbers have been assigned to ordered categories in the reverse order of the relative sizes of the categories. We should be able to handle that case.