path: root/maintenance/language/zhtable/Makefile.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @author Philip
import tarfile as tf
import zipfile as zf
import os, re, shutil, sys, platform

pyversion = platform.python_version()
islinux = platform.system().lower() == 'linux'

if pyversion[:3] in ['2.6', '2.7']:
    import urllib as urllib_request
    import codecs
    open = codecs.open
    _unichr = unichr
    if sys.maxunicode < 0x10000:
        def unichr(i):
            if i < 0x10000:
                return _unichr(i)
            else:
                return _unichr( 0xD7C0 + ( i>>10 ) ) + _unichr( 0xDC00 + ( i & 0x3FF ) )
elif pyversion[:2] == '3.':
    import urllib.request as urllib_request
    unichr = chr

def unichr2( *args ):
    return [unichr( int( i.split('<')[0][2:], 16 ) ) for i in args]

def unichr3( *args ):
    return [unichr( int( i[2:7], 16 ) ) for i in args if i[2:7]]

# DEFINE
UNIHAN_VER = '6.2.0'
SF_MIRROR = 'dfn'
SCIM_TABLES_VER = '0.5.11'
SCIM_PINYIN_VER = '0.5.92'
LIBTABE_VER = '0.2.3'
# END OF DEFINE

def download( url, dest ):
    if os.path.isfile( dest ):
        print( 'File %s is up to date.' % dest )
        return
    global islinux
    if islinux:
        # we use wget instead urlretrieve under Linux, 
        # because wget could display details like download progress
        os.system( 'wget %s -O %s' % ( url, dest ) )
    else:
        print( 'Downloading from [%s] ...' % url )
        urllib_request.urlretrieve( url, dest )
        print( 'Download complete.\n' )
    return

def uncompress( fp, member, encoding = 'U8' ):
    name = member.rsplit( '/', 1 )[-1]
    print( 'Extracting %s ...' % name )
    fp.extract( member )
    shutil.move( member, name )
    if '/' in member:
        shutil.rmtree( member.split( '/', 1 )[0] )
    return open( name, 'rb', encoding, 'ignore' )

unzip = lambda path, member, encoding = 'U8': \
        uncompress( zf.ZipFile( path ), member, encoding )

untargz = lambda path, member, encoding = 'U8': \
        uncompress( tf.open( path, 'r:gz' ), member, encoding )

def parserCore( fp, pos, beginmark = None, endmark = None ):
    if beginmark and endmark:
        start = False
    else: start = True
    mlist = set()
    for line in fp:
        if beginmark and line.startswith( beginmark ):
            start = True
            continue
        elif endmark and line.startswith( endmark ):
            break
        if start and not line.startswith( '#' ):
            elems = line.split()
            if len( elems ) < 2:
                continue
            elif len( elems[0] ) > 1 and \
                len( elems[pos] ) > 1: # words only
                mlist.add( elems[pos] )
    return mlist

def tablesParser( path, name ):
    """ Read file from scim-tables and parse it. """
    global SCIM_TABLES_VER
    src = 'scim-tables-%s/tables/zh/%s' % ( SCIM_TABLES_VER, name )
    fp = untargz( path, src, 'U8' )
    return parserCore( fp, 1, 'BEGIN_TABLE', 'END_TABLE' )

ezbigParser = lambda path: tablesParser( path, 'EZ-Big.txt.in' )
wubiParser = lambda path: tablesParser( path, 'Wubi.txt.in' )
zrmParser = lambda path: tablesParser( path, 'Ziranma.txt.in' )

def phraseParser( path ):
    """ Read phrase_lib.txt and parse it. """
    global SCIM_PINYIN_VER
    src = 'scim-pinyin-%s/data/phrase_lib.txt' % SCIM_PINYIN_VER
    dst = 'phrase_lib.txt'
    fp = untargz( path, src, 'U8' )
    return parserCore( fp, 0 )

def tsiParser( path ):
    """ Read tsi.src and parse it. """
    src = 'libtabe/tsi-src/tsi.src'
    dst = 'tsi.src'
    fp = untargz( path, src, 'big5hkscs' )
    return parserCore( fp, 0 )

def unihanParser( path ):
    """ Read Unihan_Variants.txt and parse it. """
    fp = unzip( path, 'Unihan_Variants.txt', 'U8' )
    t2s = dict()
    s2t = dict()
    for line in fp:
        if line.startswith( '#' ):
            continue
        else:
            elems = line.split()
            if len( elems ) < 3:
                continue
            type = elems.pop( 1 )
            elems = unichr2( *elems )
            if type == 'kTraditionalVariant':
                s2t[elems[0]] = elems[1:]
            elif type == 'kSimplifiedVariant':
                t2s[elems[0]] = elems[1:]
    fp.close()
    return ( t2s, s2t )

def applyExcludes( mlist, path ):
    """ Apply exclude rules from path to mlist. """
    excludes = open( path, 'rb', 'U8' ).read().split()
    excludes = [word.split( '#' )[0].strip() for word in excludes]
    excludes = '|'.join( excludes )
    excptn = re.compile( '.*(?:%s).*' % excludes )
    diff = [mword for mword in mlist if excptn.search( mword )]
    mlist.difference_update( diff )
    return mlist

def charManualTable( path ):
    fp = open( path, 'rb', 'U8' )
    ret = {}
    for line in fp:
        elems = line.split( '#' )[0].split( '|' )
        elems = unichr3( *elems )
        if len( elems ) > 1:
            ret[elems[0]] = elems[1:]
    return ret
        
def toManyRules( src_table ):
    tomany = set()
    for ( f, t ) in src_table.iteritems():
        for i in range( 1, len( t ) ):
            tomany.add( t[i] )
    return tomany

def removeRules( path, table ):
    fp = open( path, 'rb', 'U8' )
    texc = list()
    for line in fp:
        elems = line.split( '=>' )
        f = t = elems[0].strip()
        if len( elems ) == 2:
            t = elems[1].strip()
        f = f.strip('"').strip("'")
        t = t.strip('"').strip("'")
        if f:
            try:
                table.pop( f )
            except:
                pass
        if t:
            texc.append( t )
    texcptn = re.compile( '^(?:%s)$' % '|'.join( texc ) )
    for (tmp_f, tmp_t) in table.copy().iteritems():
        if texcptn.match( tmp_t ):
            table.pop( tmp_f )
    return table

def customRules( path ):
    fp = open( path, 'rb', 'U8' )
    ret = dict()
    for line in fp:
        elems = line.split( '#' )[0].split()
        if len( elems ) > 1:
            ret[elems[0]] = elems[1]
    return ret

def dictToSortedList( src_table, pos ):
    return sorted( src_table.items(), key = lambda m: m[pos] )

def translate( text, conv_table ):
    i = 0
    while i < len( text ):
        for j in range( len( text ) - i, 0, -1 ):
            f = text[i:][:j]
            t = conv_table.get( f )
            if t:
                text = text[:i] + t + text[i:][j:]
                i += len(t) - 1
                break
        i += 1
    return text

def manualWordsTable( path, conv_table, reconv_table ):
    fp = open( path, 'rb', 'U8' )
    reconv_table = {}
    wordlist = [line.split( '#' )[0].strip() for line in fp]
    wordlist = list( set( wordlist ) )
    wordlist.sort( key = len, reverse = True )
    while wordlist:
        word = wordlist.pop()
        new_word = translate( word, conv_table )
        rcv_word = translate( word, reconv_table )
        if word != rcv_word:
            reconv_table[word] = word
        reconv_table[new_word] = word
    return reconv_table

def defaultWordsTable( src_wordlist, src_tomany, char_conv_table, char_reconv_table ):
    wordlist = list( src_wordlist )
    wordlist.sort( key = len, reverse = True )
    word_conv_table = {}
    word_reconv_table = {}
    conv_table = char_conv_table.copy()
    reconv_table = char_reconv_table.copy()
    tomanyptn = re.compile( '(?:%s)' % '|'.join( src_tomany ) )
    while wordlist:
        conv_table.update( word_conv_table )
        reconv_table.update( word_reconv_table )
        word = wordlist.pop()
        new_word_len = word_len = len( word )
        while new_word_len == word_len:
            add = False
            test_word = translate( word, reconv_table )
            new_word = translate( word, conv_table )
            if not reconv_table.get( new_word ) \
               and ( test_word != word \
               or ( tomanyptn.search( word ) \
               and word != translate( new_word, reconv_table ) ) ):
                word_conv_table[word] = new_word
                word_reconv_table[new_word] = word
            try:
                word = wordlist.pop()
            except IndexError:
                break
            new_word_len = len(word)
    return word_reconv_table

def PHPArray( table ):
    lines = ['\'%s\' => \'%s\',' % (f, t) for (f, t) in table if f and t]
    return '\n'.join(lines)

def main():
    #Get Unihan.zip:
    url = 'http://www.unicode.org/Public/%s/ucd/Unihan.zip' % UNIHAN_VER
    han_dest = 'Unihan.zip'
    download( url, han_dest )
    
    # Get scim-tables-$(SCIM_TABLES_VER).tar.gz:
    url  = 'http://%s.dl.sourceforge.net/sourceforge/scim/scim-tables-%s.tar.gz' % ( SF_MIRROR, SCIM_TABLES_VER )
    tbe_dest = 'scim-tables-%s.tar.gz' % SCIM_TABLES_VER
    download( url, tbe_dest )
    
    # Get scim-pinyin-$(SCIM_PINYIN_VER).tar.gz:
    url  = 'http://%s.dl.sourceforge.net/sourceforge/scim/scim-pinyin-%s.tar.gz' % ( SF_MIRROR, SCIM_PINYIN_VER )
    pyn_dest = 'scim-pinyin-%s.tar.gz' % SCIM_PINYIN_VER
    download( url, pyn_dest )
    
    # Get libtabe-$(LIBTABE_VER).tgz:
    url  = 'http://%s.dl.sourceforge.net/sourceforge/libtabe/libtabe-%s.tgz' % ( SF_MIRROR, LIBTABE_VER )
    lbt_dest = 'libtabe-%s.tgz' % LIBTABE_VER
    download( url, lbt_dest )
    
    # Unihan.txt
    ( t2s_1tomany, s2t_1tomany ) = unihanParser( han_dest )

    t2s_1tomany.update( charManualTable( 'trad2simp.manual' ) )
    s2t_1tomany.update( charManualTable( 'simp2trad.manual' ) )
    
    t2s_1to1 = dict( [( f, t[0] ) for ( f, t ) in t2s_1tomany.iteritems()] )
    s2t_1to1 = dict( [( f, t[0] ) for ( f, t ) in s2t_1tomany.iteritems()] )
    
    s_tomany = toManyRules( t2s_1tomany )
    t_tomany = toManyRules( s2t_1tomany )

    # noconvert rules
    t2s_1to1 = removeRules( 'trad2simp_noconvert.manual', t2s_1to1 )
    s2t_1to1 = removeRules( 'simp2trad_noconvert.manual', s2t_1to1 )
    
    # the supper set for word to word conversion
    t2s_1to1_supp = t2s_1to1.copy()
    s2t_1to1_supp = s2t_1to1.copy()
    t2s_1to1_supp.update( customRules( 'trad2simp_supp_set.manual' ) )
    s2t_1to1_supp.update( customRules( 'simp2trad_supp_set.manual' ) )
    
    # word to word manual rules
    t2s_word2word_manual = manualWordsTable( 'simpphrases.manual', s2t_1to1_supp, t2s_1to1_supp )
    t2s_word2word_manual.update( customRules( 'toSimp.manual' ) )
    s2t_word2word_manual = manualWordsTable( 'tradphrases.manual', t2s_1to1_supp, s2t_1to1_supp )
    s2t_word2word_manual.update( customRules( 'toTrad.manual' ) )

    # word to word rules from input methods
    t_wordlist = set()
    s_wordlist = set()
    t_wordlist.update( ezbigParser( tbe_dest ),
                       tsiParser( lbt_dest ) )
    s_wordlist.update( wubiParser( tbe_dest ),
                       zrmParser( tbe_dest ),
                       phraseParser( pyn_dest ) )

    # exclude
    s_wordlist = applyExcludes( s_wordlist, 'simpphrases_exclude.manual' )
    t_wordlist = applyExcludes( t_wordlist, 'tradphrases_exclude.manual' )

    s2t_supp = s2t_1to1_supp.copy()
    s2t_supp.update( s2t_word2word_manual )
    t2s_supp = t2s_1to1_supp.copy()
    t2s_supp.update( t2s_word2word_manual )

    # parse list to dict
    t2s_word2word = defaultWordsTable( s_wordlist, s_tomany, s2t_1to1_supp, t2s_supp )
    t2s_word2word.update( t2s_word2word_manual )
    s2t_word2word = defaultWordsTable( t_wordlist, t_tomany, t2s_1to1_supp, s2t_supp )
    s2t_word2word.update( s2t_word2word_manual )
    
    # Final tables
    # sorted list toHans
    t2s_1to1 = dict( [( f, t ) for ( f, t ) in t2s_1to1.iteritems() if f != t] )
    toHans = dictToSortedList( t2s_1to1, 0 ) + dictToSortedList( t2s_word2word, 1 )
    # sorted list toHant
    s2t_1to1 = dict( [( f, t ) for ( f, t ) in s2t_1to1.iteritems() if f != t] )
    toHant = dictToSortedList( s2t_1to1, 0 ) + dictToSortedList( s2t_word2word, 1 )
    # sorted list toCN
    toCN = dictToSortedList( customRules( 'toCN.manual' ), 1 )
    # sorted list toHK
    toHK = dictToSortedList( customRules( 'toHK.manual' ), 1 )
    # sorted list toSG
    toSG = dictToSortedList( customRules( 'toSG.manual' ), 1 )
    # sorted list toTW
    toTW = dictToSortedList( customRules( 'toTW.manual' ), 1 )
    
    # Get PHP Array
    php = '''<?php
/**
 * Simplified / Traditional Chinese conversion tables
 *
 * Automatically generated using code and data in includes/zhtable/
 * Do not modify directly!
 *
 * @file
 */

$zh2Hant = array(\n'''
    php += PHPArray( toHant ) \
        +  '\n);\n\n$zh2Hans = array(\n' \
        +  PHPArray( toHans ) \
        +  '\n);\n\n$zh2TW = array(\n' \
        +  PHPArray( toTW ) \
        +  '\n);\n\n$zh2HK = array(\n' \
        +  PHPArray( toHK ) \
        +  '\n);\n\n$zh2CN = array(\n' \
        +  PHPArray( toCN ) \
        +  '\n);\n\n$zh2SG = array(\n' \
        +  PHPArray( toSG ) \
        +  '\n);\n'
    
    f = open( os.path.join( '..', '..', '..', 'includes', 'ZhConversion.php' ), 'wb', encoding = 'utf8' )
    print ('Writing ZhConversion.php ... ')
    f.write( php )
    f.close()
    
    # Remove temporary files
    print ('Deleting temporary files ... ')
    os.remove('EZ-Big.txt.in')
    os.remove('phrase_lib.txt')
    os.remove('tsi.src')
    os.remove('Unihan_Variants.txt')
    os.remove('Wubi.txt.in')
    os.remove('Ziranma.txt.in')
    

if __name__ == '__main__':
    main()