generate-unicode-category-regex.py

import sys
import unicodedata

def get_regex(category: str) -> str:
    start = None
    regex = '['
    for i in range(sys.maxunicode):
        if unicodedata.category(chr(i)) == category:
            if start is None:
                start = i
        elif start is not None:
            if start == i - 1:
                regex += escape(i - 1)
            else:
                regex += escape(start) + '-' + escape(i - 1)
            start = None
    return regex + ']'

def escape(char: int) -> str:
    return chr(char).encode('unicode-escape').decode('utf-8')

Lt = get_regex('Lt')
Lu = get_regex('Lu')
Lo = get_regex('Lo')
Nl = get_regex('Nl')
Ll = get_regex('Ll')
Sm = get_regex('Sm')
So = get_regex('So')

print("Lt = u'" + Lt + "'")
print("Lu = u'" + Lu + "'")
print("Lo = u'" + Lo + "'")
print("Nl = u'" + Nl + "'")
print("Ll = u'" + Ll + "'")
print("Sm = u'" + Sm + "'")
print("So = u'" + So + "'")