maintenance/language/zhtable/Makefile.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3 # @author Philip
   4 import tarfile as tf
   5 import zipfile as zf
   6 import os, re, shutil, sys, platform
   7
   8 pyversion = platform.python_version()
   9 islinux = platform.system().lower() == 'linux'
  10
  11 if pyversion[:3] in ['2.6', '2.7']:
  12     import urllib as urllib_request
  13     import codecs
  14     open = codecs.open
  15     _unichr = unichr
  16     if sys.maxunicode < 0x10000:
  17         def unichr(i):
  18             if i < 0x10000:
  19                 return _unichr(i)
  20             else:
  21                 return _unichr( 0xD7C0 + ( i>>10 ) ) + _unichr( 0xDC00 + ( i & 0x3FF ) )
  22 elif pyversion[:2] == '3.':
  23     import urllib.request as urllib_request
  24     unichr = chr
  25
  26 def unichr2( *args ):
  27     return [unichr( int( i.split('<')[0][2:], 16 ) ) for i in args]
  28
  29 def unichr3( *args ):
  30     return [unichr( int( i[2:7], 16 ) ) for i in args if i[2:7]]
  31
  32 # DEFINE
  33 UNIHAN_VER = '6.3.0'
  34 SF_MIRROR = 'dfn'
  35 SCIM_TABLES_VER = '0.5.13'
  36 SCIM_PINYIN_VER = '0.5.92'
  37 LIBTABE_VER = '0.2.3'
  38 # END OF DEFINE
  39
  40 def download( url, dest ):
  41     if os.path.isfile( dest ):
  42         print( 'File %s is up to date.' % dest )
  43         return
  44     global islinux
  45     if islinux:
  46         # we use wget instead urlretrieve under Linux,
  47         # because wget could display details like download progress
  48         os.system( 'wget %s -O %s' % ( url, dest ) )
  49     else:
  50         print( 'Downloading from [%s] ...' % url )
  51         urllib_request.urlretrieve( url, dest )
  52         print( 'Download complete.\n' )
  53     return
  54
  55 def uncompress( fp, member, encoding = 'U8' ):
  56     name = member.rsplit( '/', 1 )[-1]
  57     print( 'Extracting %s ...' % name )
  58     fp.extract( member )
  59     shutil.move( member, name )
  60     if '/' in member:
  61         shutil.rmtree( member.split( '/', 1 )[0] )
  62     if pyversion[:1] in ['2']:
  63         fc = open( name, 'rb', encoding, 'ignore' )
  64     else:
  65         fc = open( name, 'r', encoding = encoding, errors = 'ignore' )
  66     return fc
  67
  68 unzip = lambda path, member, encoding = 'U8': \
  69         uncompress( zf.ZipFile( path ), member, encoding )
  70
  71 untargz = lambda path, member, encoding = 'U8': \
  72         uncompress( tf.open( path, 'r:gz' ), member, encoding )
  73
  74 def parserCore( fp, pos, beginmark = None, endmark = None ):
  75     if beginmark and endmark:
  76         start = False
  77     else: start = True
  78     mlist = set()
  79     for line in fp:
  80         if beginmark and line.startswith( beginmark ):
  81             start = True
  82             continue
  83         elif endmark and line.startswith( endmark ):
  84             break
  85         if start and not line.startswith( '#' ):
  86             elems = line.split()
  87             if len( elems ) < 2:
  88                 continue
  89             elif len( elems[0] ) > 1 and \
  90                 len( elems[pos] ) > 1: # words only
  91                 mlist.add( elems[pos] )
  92     return mlist
  93
  94 def tablesParser( path, name ):
  95     """ Read file from scim-tables and parse it. """
  96     global SCIM_TABLES_VER
  97     src = 'scim-tables-%s/tables/zh/%s' % ( SCIM_TABLES_VER, name )
  98     fp = untargz( path, src, 'U8' )
  99     return parserCore( fp, 1, 'BEGIN_TABLE', 'END_TABLE' )
 100
 101 ezbigParser = lambda path: tablesParser( path, 'EZ-Big.txt.in' )
 102 wubiParser = lambda path: tablesParser( path, 'Wubi.txt.in' )
 103 zrmParser = lambda path: tablesParser( path, 'Ziranma.txt.in' )
 104
 105 def phraseParser( path ):
 106     """ Read phrase_lib.txt and parse it. """
 107     global SCIM_PINYIN_VER
 108     src = 'scim-pinyin-%s/data/phrase_lib.txt' % SCIM_PINYIN_VER
 109     dst = 'phrase_lib.txt'
 110     fp = untargz( path, src, 'U8' )
 111     return parserCore( fp, 0 )
 112
 113 def tsiParser( path ):
 114     """ Read tsi.src and parse it. """
 115     src = 'libtabe/tsi-src/tsi.src'
 116     dst = 'tsi.src'
 117     fp = untargz( path, src, 'big5hkscs' )
 118     return parserCore( fp, 0 )
 119
 120 def unihanParser( path ):
 121     """ Read Unihan_Variants.txt and parse it. """
 122     fp = unzip( path, 'Unihan_Variants.txt', 'U8' )
 123     t2s = dict()
 124     s2t = dict()
 125     for line in fp:
 126         if line.startswith( '#' ):
 127             continue
 128         else:
 129             elems = line.split()
 130             if len( elems ) < 3:
 131                 continue
 132             type = elems.pop( 1 )
 133             elems = unichr2( *elems )
 134             if type == 'kTraditionalVariant':
 135                 s2t[elems[0]] = elems[1:]
 136             elif type == 'kSimplifiedVariant':
 137                 t2s[elems[0]] = elems[1:]
 138     fp.close()
 139     return ( t2s, s2t )
 140
 141 def applyExcludes( mlist, path ):
 142     """ Apply exclude rules from path to mlist. """
 143     if pyversion[:1] in ['2']:
 144         excludes = open( path, 'rb', 'U8' ).read().split()
 145     else:
 146         excludes = open( path, 'r', encoding = 'U8' ).read().split()
 147     excludes = [word.split( '#' )[0].strip() for word in excludes]
 148     excludes = '|'.join( excludes )
 149     excptn = re.compile( '.*(?:%s).*' % excludes )
 150     diff = [mword for mword in mlist if excptn.search( mword )]
 151     mlist.difference_update( diff )
 152     return mlist
 153
 154 def charManualTable( path ):
 155     fp = open( path, 'r', encoding = 'U8' )
 156     ret = {}
 157     for line in fp:
 158         elems = line.split( '#' )[0].split( '|' )
 159         elems = unichr3( *elems )
 160         if len( elems ) > 1:
 161             ret[elems[0]] = elems[1:]
 162     return ret
 163
 164 def toManyRules( src_table ):
 165     tomany = set()
 166     if pyversion[:1] in ['2']:
 167         for ( f, t ) in src_table.iteritems():
 168             for i in range( 1, len( t ) ):
 169                 tomany.add( t[i] )
 170     else:
 171         for ( f, t ) in src_table.items():
 172             for i in range( 1, len( t ) ):
 173                 tomany.add( t[i] )
 174     return tomany
 175
 176 def removeRules( path, table ):
 177     fp = open( path, 'r', encoding = 'U8' )
 178     texc = list()
 179     for line in fp:
 180         elems = line.split( '=>' )
 181         f = t = elems[0].strip()
 182         if len( elems ) == 2:
 183             t = elems[1].strip()
 184         f = f.strip('"').strip("'")
 185         t = t.strip('"').strip("'")
 186         if f:
 187             try:
 188                 table.pop( f )
 189             except:
 190                 pass
 191         if t:
 192             texc.append( t )
 193     texcptn = re.compile( '^(?:%s)$' % '|'.join( texc ) )
 194     if pyversion[:1] in ['2']:
 195         for (tmp_f, tmp_t) in table.copy().iteritems():
 196             if texcptn.match( tmp_t ):
 197                 table.pop( tmp_f )
 198     else:
 199         for (tmp_f, tmp_t) in table.copy().items():
 200             if texcptn.match( tmp_t ):
 201                 table.pop( tmp_f )
 202     return table
 203
 204 def customRules( path ):
 205     fp = open( path, 'r', encoding = 'U8' )
 206     ret = dict()
 207     for line in fp:
 208         elems = line.split( '#' )[0].split()
 209         if len( elems ) > 1:
 210             ret[elems[0]] = elems[1]
 211     return ret
 212
 213 def dictToSortedList( src_table, pos ):
 214     return sorted( src_table.items(), key = lambda m: m[pos] )
 215
 216 def translate( text, conv_table ):
 217     i = 0
 218     while i < len( text ):
 219         for j in range( len( text ) - i, 0, -1 ):
 220             f = text[i:][:j]
 221             t = conv_table.get( f )
 222             if t:
 223                 text = text[:i] + t + text[i:][j:]
 224                 i += len(t) - 1
 225                 break
 226         i += 1
 227     return text
 228
 229 def manualWordsTable( path, conv_table, reconv_table ):
 230     fp = open( path, 'r', encoding = 'U8' )
 231     reconv_table = {}
 232     wordlist = [line.split( '#' )[0].strip() for line in fp]
 233     wordlist = list( set( wordlist ) )
 234     wordlist.sort( key = len, reverse = True )
 235     while wordlist:
 236         word = wordlist.pop()
 237         new_word = translate( word, conv_table )
 238         rcv_word = translate( word, reconv_table )
 239         if word != rcv_word:
 240             reconv_table[word] = word
 241         reconv_table[new_word] = word
 242     return reconv_table
 243
 244 def defaultWordsTable( src_wordlist, src_tomany, char_conv_table, char_reconv_table ):
 245     wordlist = list( src_wordlist )
 246     wordlist.sort( key = len, reverse = True )
 247     word_conv_table = {}
 248     word_reconv_table = {}
 249     conv_table = char_conv_table.copy()
 250     reconv_table = char_reconv_table.copy()
 251     tomanyptn = re.compile( '(?:%s)' % '|'.join( src_tomany ) )
 252     while wordlist:
 253         conv_table.update( word_conv_table )
 254         reconv_table.update( word_reconv_table )
 255         word = wordlist.pop()
 256         new_word_len = word_len = len( word )
 257         while new_word_len == word_len:
 258             add = False
 259             test_word = translate( word, reconv_table )
 260             new_word = translate( word, conv_table )
 261             if not reconv_table.get( new_word ) \
 262                and ( test_word != word \
 263                or ( tomanyptn.search( word ) \
 264                and word != translate( new_word, reconv_table ) ) ):
 265                 word_conv_table[word] = new_word
 266                 word_reconv_table[new_word] = word
 267             try:
 268                 word = wordlist.pop()
 269             except IndexError:
 270                 break
 271             new_word_len = len(word)
 272     return word_reconv_table
 273
 274 def PHPArray( table ):
 275     lines = ['\'%s\' => \'%s\',' % (f, t) for (f, t) in table if f and t]
 276     return '\n'.join(lines)
 277
 278 def main():
 279     #Get Unihan.zip:
 280     url = 'http://www.unicode.org/Public/%s/ucd/Unihan.zip' % UNIHAN_VER
 281     han_dest = 'Unihan.zip'
 282     download( url, han_dest )
 283
 284     # Get scim-tables-$(SCIM_TABLES_VER).tar.gz:
 285     url  = 'http://%s.dl.sourceforge.net/sourceforge/scim/scim-tables-%s.tar.gz' % ( SF_MIRROR, SCIM_TABLES_VER )
 286     tbe_dest = 'scim-tables-%s.tar.gz' % SCIM_TABLES_VER
 287     download( url, tbe_dest )
 288
 289     # Get scim-pinyin-$(SCIM_PINYIN_VER).tar.gz:
 290     url  = 'http://%s.dl.sourceforge.net/sourceforge/scim/scim-pinyin-%s.tar.gz' % ( SF_MIRROR, SCIM_PINYIN_VER )
 291     pyn_dest = 'scim-pinyin-%s.tar.gz' % SCIM_PINYIN_VER
 292     download( url, pyn_dest )
 293
 294     # Get libtabe-$(LIBTABE_VER).tgz:
 295     url  = 'http://%s.dl.sourceforge.net/sourceforge/libtabe/libtabe-%s.tgz' % ( SF_MIRROR, LIBTABE_VER )
 296     lbt_dest = 'libtabe-%s.tgz' % LIBTABE_VER
 297     download( url, lbt_dest )
 298
 299     # Unihan.txt
 300     ( t2s_1tomany, s2t_1tomany ) = unihanParser( han_dest )
 301
 302     t2s_1tomany.update( charManualTable( 'trad2simp.manual' ) )
 303     s2t_1tomany.update( charManualTable( 'simp2trad.manual' ) )
 304
 305     if pyversion[:1] in ['2']:
 306       t2s_1to1 = dict( [( f, t[0] ) for ( f, t ) in t2s_1tomany.iteritems()] )
 307       s2t_1to1 = dict( [( f, t[0] ) for ( f, t ) in s2t_1tomany.iteritems()] )
 308     else:
 309       t2s_1to1 = dict( [( f, t[0] ) for ( f, t ) in t2s_1tomany.items()] )
 310       s2t_1to1 = dict( [( f, t[0] ) for ( f, t ) in s2t_1tomany.items()] )
 311
 312     s_tomany = toManyRules( t2s_1tomany )
 313     t_tomany = toManyRules( s2t_1tomany )
 314
 315     # noconvert rules
 316     t2s_1to1 = removeRules( 'trad2simp_noconvert.manual', t2s_1to1 )
 317     s2t_1to1 = removeRules( 'simp2trad_noconvert.manual', s2t_1to1 )
 318
 319     # the supper set for word to word conversion
 320     t2s_1to1_supp = t2s_1to1.copy()
 321     s2t_1to1_supp = s2t_1to1.copy()
 322     t2s_1to1_supp.update( customRules( 'trad2simp_supp_set.manual' ) )
 323     s2t_1to1_supp.update( customRules( 'simp2trad_supp_set.manual' ) )
 324
 325     # word to word manual rules
 326     t2s_word2word_manual = manualWordsTable( 'simpphrases.manual', s2t_1to1_supp, t2s_1to1_supp )
 327     t2s_word2word_manual.update( customRules( 'toSimp.manual' ) )
 328     s2t_word2word_manual = manualWordsTable( 'tradphrases.manual', t2s_1to1_supp, s2t_1to1_supp )
 329     s2t_word2word_manual.update( customRules( 'toTrad.manual' ) )
 330
 331     # word to word rules from input methods
 332     t_wordlist = set()
 333     s_wordlist = set()
 334     t_wordlist.update( ezbigParser( tbe_dest ),
 335                        tsiParser( lbt_dest ) )
 336     s_wordlist.update( wubiParser( tbe_dest ),
 337                        zrmParser( tbe_dest ),
 338                        phraseParser( pyn_dest ) )
 339
 340     # exclude
 341     s_wordlist = applyExcludes( s_wordlist, 'simpphrases_exclude.manual' )
 342     t_wordlist = applyExcludes( t_wordlist, 'tradphrases_exclude.manual' )
 343
 344     s2t_supp = s2t_1to1_supp.copy()
 345     s2t_supp.update( s2t_word2word_manual )
 346     t2s_supp = t2s_1to1_supp.copy()
 347     t2s_supp.update( t2s_word2word_manual )
 348
 349     # parse list to dict
 350     t2s_word2word = defaultWordsTable( s_wordlist, s_tomany, s2t_1to1_supp, t2s_supp )
 351     t2s_word2word.update( t2s_word2word_manual )
 352     s2t_word2word = defaultWordsTable( t_wordlist, t_tomany, t2s_1to1_supp, s2t_supp )
 353     s2t_word2word.update( s2t_word2word_manual )
 354
 355     # Final tables
 356     # sorted list toHans
 357     if pyversion[:1] in ['2']:
 358         t2s_1to1 = dict( [( f, t ) for ( f, t ) in t2s_1to1.iteritems() if f != t] )
 359     else:
 360         t2s_1to1 = dict( [( f, t ) for ( f, t ) in t2s_1to1.items() if f != t] )
 361     toHans = dictToSortedList( t2s_1to1, 0 ) + dictToSortedList( t2s_word2word, 1 )
 362     # sorted list toHant
 363     if pyversion[:1] in ['2']:
 364         s2t_1to1 = dict( [( f, t ) for ( f, t ) in s2t_1to1.iteritems() if f != t] )
 365     else:
 366         s2t_1to1 = dict( [( f, t ) for ( f, t ) in s2t_1to1.items() if f != t] )
 367     toHant = dictToSortedList( s2t_1to1, 0 ) + dictToSortedList( s2t_word2word, 1 )
 368     # sorted list toCN
 369     toCN = dictToSortedList( customRules( 'toCN.manual' ), 1 )
 370     # sorted list toHK
 371     toHK = dictToSortedList( customRules( 'toHK.manual' ), 1 )
 372     # sorted list toSG
 373     toSG = dictToSortedList( customRules( 'toSG.manual' ), 1 )
 374     # sorted list toTW
 375     toTW = dictToSortedList( customRules( 'toTW.manual' ), 1 )
 376
 377     # Get PHP Array
 378     php = '''<?php
 379 /**
 380  * Simplified / Traditional Chinese conversion tables
 381  *
 382  * Automatically generated using code and data in maintenance/language/zhtable/
 383  * Do not modify directly!
 384  *
 385  * @file
 386  */
 387
 388 $zh2Hant = array(\n'''
 389     php += PHPArray( toHant ) \
 390         +  '\n);\n\n$zh2Hans = array(\n' \
 391         +  PHPArray( toHans ) \
 392         +  '\n);\n\n$zh2TW = array(\n' \
 393         +  PHPArray( toTW ) \
 394         +  '\n);\n\n$zh2HK = array(\n' \
 395         +  PHPArray( toHK ) \
 396         +  '\n);\n\n$zh2CN = array(\n' \
 397         +  PHPArray( toCN ) \
 398         +  '\n);\n\n$zh2SG = array(\n' \
 399         +  PHPArray( toSG ) \
 400         +  '\n);\n'
 401
 402     if pyversion[:1] in ['2']:
 403         f = open( os.path.join( '..', '..', '..', 'includes', 'ZhConversion.php' ), 'wb', encoding = 'utf8' )
 404     else:
 405         f = open( os.path.join( '..', '..', '..', 'includes', 'ZhConversion.php' ), 'w', buffering = 4096, encoding = 'utf8' )
 406     print ('Writing ZhConversion.php ... ')
 407     f.write( php )
 408     f.close()
 409
 410     # Remove temporary files
 411     print ('Deleting temporary files ... ')
 412     os.remove('EZ-Big.txt.in')
 413     os.remove('phrase_lib.txt')
 414     os.remove('tsi.src')
 415     os.remove('Unihan_Variants.txt')
 416     os.remove('Wubi.txt.in')
 417     os.remove('Ziranma.txt.in')
 418
 419
 420 if __name__ == '__main__':
 421     main()