sc/workben/celltrans/parse.py

   1 #!/usr/bin/env python
   2 #***********************************************************************
   3 #
   4 # DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5 #
   6 # Copyright 2008 by Sun Microsystems, Inc.
   7 #
   8 # OpenOffice.org - a multi-platform office productivity suite
   9 #
  10 # $RCSfile: parse.py,v $
  11 #
  12 # $Revision: 1.3 $
  13 #
  14 # This file is part of OpenOffice.org.
  15 #
  16 # OpenOffice.org is free software: you can redistribute it and/or modify
  17 # it under the terms of the GNU Lesser General Public License version 3
  18 # only, as published by the Free Software Foundation.
  19 #
  20 # OpenOffice.org is distributed in the hope that it will be useful,
  21 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  22 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23 # GNU Lesser General Public License version 3 for more details
  24 # (a copy is included in the LICENSE file that accompanied this code).
  25 #
  26 # You should have received a copy of the GNU Lesser General Public License
  27 # version 3 along with OpenOffice.org.  If not, see
  28 # <http://www.openoffice.org/license.html>
  29 # for a copy of the LGPLv3 License.
  30 #
  31 #***********************************************************************
  32
  33 import sys
  34
  35 localeNames = {'fr': 'French', 'hu': 'Hungarian', 'de': 'German'}
  36 def getLocaleName (code):
  37     global localeNames
  38     if localeNames.has_key(code):
  39         return localeNames[code]
  40     else:
  41         return "(unknown locale)"
  42
  43 def getAscii (ords):
  44     ascii = ''
  45     for c in ords:
  46         ascii += chr(c)
  47     return ascii
  48
  49 class LocaleData(object):
  50     def __init__ (self, locale):
  51         self.locale = locale
  52         self.funcList = {}
  53
  54     def addKeywordMap (self, funcName, localeName, engName):
  55         if not self.funcList.has_key(funcName):
  56             self.funcList[funcName] = []
  57
  58         self.funcList[funcName].append([localeName, engName])
  59
  60     def getLocaleFuncVarName (self, func, pair):
  61         return func.lower() + "_" + getAscii(pair[1]).lower() + "_" + self.locale
  62
  63     def dumpCode (self):
  64         chars = ""
  65
  66         # locale output
  67         chars += "// " + "-"*75 + "\n"
  68         chars += "// %s language locale (automatically generated)\n"%getLocaleName(self.locale)
  69         chars += "// " + "-"*75 + "\n"
  70         chars += "static const Locale a" + self.locale.capitalize() + "(OUString::createFromAscii(\""
  71         chars += self.locale
  72         chars += "\"), OUString(), OUString());\n\n"
  73
  74         # pre instantiations of localized function names.
  75         funcs = self.funcList.keys()
  76         funcs.sort()
  77         chars += "// pre instantiations of localized function names\n"
  78         for func in funcs:
  79             for item in self.funcList[func]:
  80                 chars += "static const sal_Unicode " + self.getLocaleFuncVarName(func, item) + "[] = {\n"
  81                 chars += "    "
  82                 isFirst = True
  83                 # Dump the UTF-16 bytes.
  84                 for uval in item[0]:
  85                     if isFirst:
  86                         isFirst = False
  87                     else:
  88                         chars += ", "
  89                     chars += "0x%.4X"%uval
  90
  91                 # Don't forget to null-terminate the string.
  92                 if not isFirst:
  93                     chars += ", "
  94                 chars += "0x0000"
  95
  96                 chars += "};\n"
  97
  98         # map item instantiations
  99         chars += "\n"
 100         chars += "static const TransItem p" + self.locale.capitalize() + "[] = {\n"
 101         for func in funcs:
 102             for item in self.funcList[func]:
 103                 chars += "    "
 104                 chars += "{%s, \"%s\", %s},\n"%(self.getLocaleFuncVarName(func, item),
 105                                                 getAscii(item[1]),
 106                                                 "oc"+func.capitalize())
 107
 108         chars += "    {NULL, NULL, ocNone}\n"
 109         chars += "};\n\n"
 110
 111         # addToMap call
 112         chars += "addToMap(%s, %s);\n"%(
 113             "p"+self.locale.capitalize(), "a"+self.locale.capitalize())
 114
 115         return chars
 116
 117 class Parser(object):
 118
 119     def __init__ (self, args):
 120         # default input & output files.
 121         self.infile = "./keywords_utf16.txt"
 122         self.outfile = "../../source/core/tool/cellkeywords.inl"
 123
 124         if len(args) >= 2:
 125             self.infile = args[1]
 126         if len(args) >= 3:
 127             self.outfile = args[2]
 128
 129     def getDByte (self):
 130         # Assume little endian.
 131         bh = ord(self.bytes[self.i])
 132         bl = ord(self.bytes[self.i+1])
 133         dbyte = bl*256 + bh
 134         self.i += 2
 135         return dbyte
 136
 137     def parseLine (self):
 138         buf = []
 139         while self.i < self.size:
 140             dbyte = self.getDByte()
 141             if dbyte == 0x000A:
 142                 break
 143             buf.append(dbyte)
 144         return buf
 145
 146     def dumpBuf (self, buf, linefeed=True):
 147         for item in buf:
 148             sys.stdout.write(chr(item))
 149         if linefeed:
 150             print ''
 151
 152     def parse (self):
 153
 154         file = open(self.infile, 'r')
 155         self.bytes = file.read()
 156         file.close()
 157
 158         self.size = len(self.bytes)
 159         self.i = 0
 160
 161         localeList = []  # stores an array of locale data objects.
 162         funcName = None
 163         word = []
 164         wordPair = []
 165
 166         while self.i < self.size:
 167             dbyte = self.getDByte()
 168             if dbyte == 0xFEFF and self.i == 2:
 169                 # unicode signature - ignore it.
 170                 pass
 171             elif dbyte == 0x0024:
 172                 # $ - locale name
 173                 buf = self.parseLine()
 174                 locale = getAscii(buf)
 175                 localeList.append(LocaleData(locale))
 176
 177             elif dbyte == 0x0040:
 178                 # @ - function name
 179                 buf = self.parseLine()
 180                 funcName = getAscii(buf)
 181
 182             elif dbyte == 0x002C:
 183                 # , - comma separator
 184                 if len(word) > 0:
 185                     wordPair.append(word)
 186                     word = []
 187             elif dbyte == 0x000A:
 188                 # linefeed
 189                 if len(word) > 0:
 190                     wordPair.append(word)
 191                     word = []
 192                 if len(wordPair) >= 2:
 193                     localeList[-1].addKeywordMap(funcName, wordPair[0], wordPair[1])
 194                 wordPair = []
 195             elif dbyte in [0x0009, 0x0020]:
 196                 # whitespace - ignore it.
 197                 pass
 198             else:
 199                 word.append(dbyte)
 200
 201         chars = "// This file has been automatically generated.  Do not hand-edit this!\n"
 202         for obj in localeList:
 203             chars += "\n" + obj.dumpCode()
 204
 205         # Write to output file.
 206         file = open(self.outfile, 'w')
 207         file.write(chars)
 208         file.close()
 209
 210 if __name__=='__main__':
 211     parser = Parser(sys.argv)
 212     parser.parse()
 213