Code we use to replace "dumb" vulgar fractions with their Unicode counterparts

vulgarfractions.py

# Map ASCII fractions to fancy fractions
FRACTION_MAP = {
    u'1/2' : u'\u00BD',
    u'1/4' : u'\u00BC',
    u'3/4' : u'\u00BE',
    u'1/3' : u'\u2153',
    u'2/3' : u'\u2154',
    u'1/5' : u'\u2155',
    u'2/5' : u'\u2156',
    u'3/5' : u'\u2157',
    u'4/5' : u'\u2158',
    u'1/6' : u'\u2159',
    u'5/6' : u'\u215A',
    u'1/8' : u'\u215B',
    u'3/8' : u'\u215C',
    u'5/8' : u'\u215D',
    u'7/8' : u'\u215E',
}

FRACRE = re.compile(ur'(\d+\s+)?(%s)\b' % u'|'.join(FRACTION_MAP))

def _substitute_fractions(m, lookup=FRACTION_MAP.get):
    whole, frac = m.groups()
    frac = lookup(frac)
    return whole.rstrip() + frac if whole else frac

def nicefractions(s, sub=FRACRE.sub):
    """Replaces dumb vulgar fractions (e.g. "1/2") with their nice Unicode
    counterparts (e.g. u"\u00BD")."""
    try:
        return sub(_substitute_fractions, s)
    except:
        return s

The actual compiled regular expression looks like this, after the keys from FRACTION_MAP are joined and subsituted in:

(\d+\s+)?(1/2|1/3|1/4|1/5|1/6|1/8|2/3|2/5|3/4|3/5|3/8|4/5|5/6|5/8|7/8)\b