sort.py

# -*- coding: UTF-8 -*-
import unicodedata
import re

RE_ALPHANUMSPACE = re.compile(r'[^0-9A-Za-z\s]*')  # \W include "_" as does A-z

def normalize_sort_field(sort_field,
                         default_missing='~title unknown',
                         missing_equivalents=['title unknown']):
    
    #fold case
    sort_field = sort_field.lower()
    
    # fold diacritics https://stackoverflow.com/a/518232/1763984
    sort_field = ''.join(c for c in unicodedata.normalize('NFD', sort_field)
                  if unicodedata.category(c) != 'Mn')
    
    # remove punctuation
    sort_field = RE_ALPHANUMSPACE.sub('', sort_field)
    
    # remove English initial articles
    words = sort_field.split()
    if words:
        if words[0] in ('the', 'a', 'an'):
            sort_field = ' '.join(words[1:])
    if not sort_field or sort_field in missing_equivalents:
        sort_field = default_missing
        
    # normalize whitespace https://stackoverflow.com/a/46501496/1763984    
    sort_field = u' '.join(sort_field.split())
 
    return sort_field


print(normalize_sort_field(u'嶋巡月弓張 [Shimameguri tsuki no yumihari]'))
print(normalize_sort_field(u'$ 容競出入湊 [Sugatakurabe deiri no minato]'))
print(normalize_sort_field(u'菅原伝授手習鑑 [Sugawara denju tenarai kagami]'))
print(normalize_sort_field(u'$ 崇禅寺馬場 [Sōzenji bama]'))