source: scripts/generate-translit-tab @ 1279:088d2f519391

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
    Build Translit-Tab
    ~~~~~~~~~~~~~~~~~~

    Generates a conversion of a transtab map.  Used by `zine.utils.text`.
    You will need a version of transtab which you can get for example
    here: http://www.bitbucket.org/jek/translitcodec/

    :copyright: (c) 2010 by the Zine Team, see AUTHORS for more details.
    :license: BSD, see LICENSE for more details.
"""
import os
import sys
import csv
from optparse import OptionParser


import _init_zine
import zine
sys.path.append(os.path.dirname(__file__))
csv.register_dialect('transtab', delimiter=';')


def read_table(path):
    long, short, single = {}, {}, {}

    t = open(path)
    for line in t.readlines():
        if not line.startswith('<'):
            continue
        from_spec, raw_to = line.strip().split(' ', 1)
        from_ord = int(from_spec[2:-1], 16)

        raw = csv.reader([raw_to], 'transtab').next()
        long_char = _unpack_uchrs(raw[0])
        if len(raw) < 2:
            short_char = long_char
        else:
            short_char = _unpack_uchrs(raw[1])

        long[from_ord] = long_char
        short[from_ord] = short_char
        if len(short_char) == 1:
            single[from_ord] = short_char
    return long, short, single


def _unpack_uchrs(packed):
    chunks = packed.replace('<U', ' ').strip().split()
    return ''.join(unichr(int(spec[:-1], 16)) for spec in chunks)


def update_mapping(long, short, single, path):
    src = open(path)
    try:
        data = src.read()
        pos = 0
        for x in xrange(2):
            pos = data.find('"""', pos) + 1
        preamble = data[:pos + 3]
    finally:
        src.close()

    rewrite = open(path, 'wb')
    try:
        rewrite.writelines(preamble)
        _dump_dict(rewrite, 'LONG_TABLE', long)
        _dump_dict(rewrite, 'SHORT_TABLE', short)
        _dump_dict(rewrite, 'SINGLE_TABLE', single)
    finally:
        rewrite.close()


def _dump_dict(fh, name, data):
    fh.write('\n%s = {\n' % name)
    for pair in sorted(data.items()):
        fh.write('    %r: %r,\n' % pair)
    fh.write('}\n')


def main():
    global parser
    parser = OptionParser(usage='%prog [path/to/transtab]')
    options, args = parser.parse_args()
    if len(args) != 1:
        parser.error('incorrect number of arguments')

    mapping_file = os.path.join(os.path.dirname(zine.__file__),
                                '_dynamic', 'translit_tab.py')
    table = read_table(os.path.join(args[0], 'transtab'))
    update_mapping(path=mapping_file, *table)
    print 'All done.'


if __name__ == '__main__':
    main()