Fix TypeError when computing p-values with missing categorical values

[tompollard]

This PR fixes a TypeError that occurs when computing p-values for categorical variables containing missing values. The error arises because missing values are replaced with the string 'None', leading to mixed-type categories (e.g.int and str) which cannot be sorted during internal processing (ref #161 and #160).

Previously, the following code would raise TypeError: '<' not supported between instances of 'str' and 'int':

from tableone import TableOne
import pandas as pd
import numpy as np

df = pd.DataFrame({
    'group': ['A', 'B', 'A', 'B', 'A'],
    'numeric_cat': [1, 2, np.nan, 2, 1]
})

t1 = TableOne(df, columns=['numeric_cat'], categorical=['numeric_cat'], groupby='group', pval=True)
print(t1.tableone)

The fixes are:

• Modified handle_categorical_nulls() to convert entire columns to string before replacing nulls with 'None', avoiding mixed-type category issues.
• Added a key=str argument when sorting categories to prevent sorting errors in edge cases.