Texy typoška

texytypo.py

import re

# sada regularnich vyrazu pro lepsi typosku
# zdroj: texy-2.1/texy/modules/TexyTypographyModule.php
TEXY_CHARS = ur'A-Za-z\u00C0-\u02FF\u0370-\u1EFF'

TEXY_TYPO_PATTERNS = [
    (re.compile(ur'(?<![.\u2026])\.{3,4}(?![.\u2026])', re.M | re.U), u"\u2026"),             # ellipsis  ...
    (re.compile(ur'(?<=[\d ])-(?=[\d ]|$)'), u"\u2013"),                                      # en dash 123-123
    (re.compile(ur'(?<=[^!*+,/:;<=>@\\\\_|-])--(?=[^!*+,/:;<=>@\\\\_|-])'), u"\u2013"),       # en dash alphanum--alphanum
    (re.compile(ur',-'), ur",\u2013"),                                                        # en dash ,-
    (re.compile(ur'(?<!\d)(\d{1,2}\.) (\d{1,2}\.) (\d\d)'), ur"\1\u00a0\2\u00a0\3"),          # date 23. 1. 1978
    (re.compile(ur'(?<!\d)(\d{1,2}\.) (\d{1,2}\.)'), ur"\1\u00a0\2"),                         # date 23. 1.
    (re.compile(ur' --- '), u"\u00a0\u2014 "),                                                # em dash ---
    (re.compile(ur' ([\u2013\u2014])', re.U), ur"\u00a0\1"),                                  # &nbsp; behind dash (dash stays at line end)
    (re.compile(ur' <-{1,2}> '), u" \u2194 "),                                                # left right arrow <-->
    (re.compile(ur'-{1,}> '), u" \u2192 "),                                                   # right arrow -->
    (re.compile(ur' <-{1,}'), u" \u2190 "),                                                   # left arrow <--
    (re.compile(ur'={1,}> '), u" \u21d2 "),                                                   # right arrow ==>
    (re.compile(ur'\\+-'), ur"\u00b1"),                                                       # +-

    (re.compile(ur'(\d+)( ?)x\\2(?=\d)'), ur"\1\u00d7"),                                      # dimension sign 123 x 123...
    (re.compile(ur'(?<=\d)x(?= |,|.|$)', re.M), ur"\u00d7"),                                  # dimension sign 123x
    (re.compile(ur'(\S ?)\(TM\)', re.I), ur"\1\u2122"),                                       # trademark (TM)
    (re.compile(ur'(\S ?)\(R\)', re.I), ur"\1\u00ae"),                                        # registered (R)
    (re.compile(ur'\(C\)( ?\S)', re.I), ur"\u00a9\1"),                                        # copyright (C)
    (re.compile(ur'\(EUR\)'), ur"\u20ac"),                                                    # Euro (EUR)
    (re.compile(ur'(\d) (?=\d{3})'), ur"\1\u00a0"),                                           # (phone) number 1 123 123 123...

    (re.compile(ur'(?<=[^\s\x17])\s+([\x17-\x1F]+)(?=\s)', re.U), r"\1"),                     # remove intermarkup space phase 1
    (re.compile(ur'(?<=\s)([\x17-\x1F]+)\s+', re.U), r"\1"),                                  # remove intermarkup space phase 2

    # space between preposition and word
    (re.compile(ur"(?<=[^0-9%s])([\x17-\x1F]*[ksvzouiaKSVZOUIA][\x17-\x1F]*)\s+(?=[\x17-\x1F]*[0-9%s])" % (TEXY_CHARS, TEXY_CHARS), re.M | re.U | re.S), ur'\1\u00a0'),
    # nbsp space between number (optionally followed by dot) and word, symbol, punctation, currency symbol
    (re.compile(ur'(?<= |\.|,|-|\+|\x16|\()([\x17-\x1F]*\d+\.?[\x17-\x1F]*)\s+(?=[\x17-\x1F]*[%' + TEXY_CHARS + r'\u00B0-\u00be\u2020-\u214f])', re.M | re.U), ur'\1\u00a0'),
    (re.compile(ur'(?<=\d\u00A0)([\x17-\x1F]*\d+\.?[\x17-\x1F]*)\s+(?=[\x17-\x1F]*[%' + TEXY_CHARS + r'\u00B0-\u00be\u2020-\u214f])', re.M | re.U), ur'\1\u00a0'),
    # space before last short word
    (re.compile(ur'(?<=.{50})\s+(?=[\x17-\x1F]*\S{1,6}[\x17-\x1F]*$)', re.S | re.U), u'\u00a0'),

    (re.compile('(?<!"|\w)"(?!\ |")(.+)(?<!\ |")"(?!")()', re.U), ur'\u201E\1\u201C'),        # double ""
    (re.compile('(?<!\'|\w)\'(?!\ |\')(.+?)(?<!\ |\')\'(?!\')()', re.U), ur'\u201A\1\u2018'), # single ''
]

def typotexy(text):
    """
    Aplikuje na zadany text typograficka pravidla z Texy.
    """
    for patt in TEXY_TYPO_PATTERNS:
        text = patt[0].sub(patt[1], text)
    return text