convert/oo2xliff.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3 #
   4 # Copyright 2003-2008 Zuza Software Foundation
   5 #
   6 # This file is part of translate.
   7 #
   8 # translate is free software; you can redistribute it and/or modify
   9 # it under the terms of the GNU General Public License as published by
  10 # the Free Software Foundation; either version 2 of the License, or
  11 # (at your option) any later version.
  12 #
  13 # translate is distributed in the hope that it will be useful,
  14 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 # GNU General Public License for more details.
  17 #
  18 # You should have received a copy of the GNU General Public License
  19 # along with translate; if not, write to the Free Software
  20 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  21 #
  22
  23 """convert an OpenOffice.org (SDF) localization file to XLIFF localization files
  24
  25 User documentation: http://translate.sourceforge.net/wiki/toolkit/oo2po
  26 """
  27
  28 import os
  29 import sys
  30 from translate.storage import xliff
  31 from translate.storage import oo
  32
  33 # TODO: support using one GSI file as template, another as input (for when English is in one and translation in another)
  34
  35 class oo2xliff:
  36     def __init__(self, sourcelanguage, targetlanguage, blankmsgstr=False, long_keys=False):
  37         """construct an oo2xliff converter for the specified languages"""
  38         self.sourcelanguage = sourcelanguage
  39         self.targetlanguage = targetlanguage
  40         self.blankmsgstr = blankmsgstr
  41         self.long_keys = long_keys
  42
  43     def maketargetunit(self, part1, part2, translators_comment, key, subkey):
  44         """makes a base unit (.po or XLIFF) out of a subkey of two parts"""
  45         #TODO: Do better
  46         text1 = getattr(part1, subkey)
  47         if text1 == "":
  48             return None
  49         text2 = getattr(part2, subkey)
  50
  51         unit = xliff.xliffunit(text1)
  52         unit.target = text2
  53         unit.markfuzzy(False)
  54         unit.addlocation(key + "." + subkey)
  55         if getattr(translators_comment, subkey).strip() != "":
  56             unit.addnote(getattr(translators_comment, subkey), origin="developer")
  57         return unit
  58
  59     def makekey(self, ookey):
  60         """converts an oo key tuple into a key identifier for the base class file (.po or XLIFF)"""
  61         project, sourcefile, resourcetype, groupid, localid, platform = ookey
  62         sourcefile = sourcefile.replace('\\','/')
  63         if self.long_keys:
  64             sourcebase = os.path.join(project, sourcefile)
  65         else:
  66             sourceparts = sourcefile.split('/')
  67             sourcebase = "".join(sourceparts[-1:])
  68         if (groupid) == 0 or len(localid) == 0:
  69             ooid = groupid + localid
  70         else:
  71             ooid = groupid + "." + localid
  72         if resourcetype:
  73             ooid = ooid + "." + resourcetype
  74         key = "%s#%s" % (sourcebase, ooid)
  75         return oo.normalizefilename(key)
  76
  77     def convertelement(self, theoo):
  78         """convert an oo element into a list of base units (.po or XLIFF)"""
  79         if self.sourcelanguage in theoo.languages:
  80             part1 = theoo.languages[self.sourcelanguage]
  81         else:
  82             print >> sys.stderr, "/".join(theoo.lines[0].getkey()), "language not found: %s" % (self.sourcelanguage)
  83             return []
  84         if self.blankmsgstr:
  85             # use a blank part2
  86             part2 = oo.ooline()
  87         else:
  88             if self.targetlanguage in theoo.languages:
  89                 part2 = theoo.languages[self.targetlanguage]
  90             else:
  91                 # if the language doesn't exist, the translation is missing ... so make it blank
  92                 part2 = oo.ooline()
  93         if "x-comment" in theoo.languages:
  94             translators_comment = theoo.languages["x-comment"]
  95         else:
  96             translators_comment = oo.ooline()
  97         key = self.makekey(part1.getkey())
  98         unitlist = []
  99         for subkey in ("text", "quickhelptext", "title"):
 100             unit = self.maketargetunit(part1, part2, translators_comment, key, subkey)
 101             if unit is not None:
 102                 unitlist.append(unit)
 103         return unitlist
 104
 105     def convertstore(self, theoofile, duplicatestyle="msgctxt"):
 106         """converts an entire oo file to a base class format (.po or XLIFF)"""
 107         thetargetfile = xliff.xlifffile()
 108         # create a header for the file
 109         bug_url = 'http://qa.openoffice.org/issues/enter_bug.cgi' + ('''?subcomponent=ui&comment=&short_desc=Localization issue in file: %(filename)s&component=l10n&form_name=enter_issue''' % {"filename": theoofile.filename}).replace(" ", "%20").replace(":", "%3A")
 110         # go through the oo and convert each element
 111         for theoo in theoofile.units:
 112             unitlist = self.convertelement(theoo)
 113             for unit in unitlist:
 114                 thetargetfile.addunit(unit)
 115         return thetargetfile
 116
 117 def verifyoptions(options):
 118     """verifies the commandline options"""
 119     if not options.targetlanguage:
 120         raise ValueError("You must specify the target language.")
 121
 122 def convertoo(inputfile, outputfile, templates, pot=False, sourcelanguage=None, targetlanguage=None, duplicatestyle="msgid_comment", multifilestyle="single"):
 123     """reads in stdin using inputstore class, converts using convertorclass, writes to stdout"""
 124     inputstore = oo.oofile()
 125     if hasattr(inputfile, "filename"):
 126         inputstore.filename = inputfile.filename
 127     inputstore.parse(inputfile.read())
 128     if not sourcelanguage:
 129         testlangtype = targetlanguage or (inputstore and inputstore.languages[0]) or ""
 130         if testlangtype.isdigit():
 131             sourcelanguage = "01"
 132         else:
 133             sourcelanguage = "en-US"
 134     if not sourcelanguage in inputstore.languages:
 135         print >> sys.stderr, "Warning: sourcelanguage %s not found in inputfile (contains %s)" % (sourcelanguage, ", ".join(inputstore.languages))
 136     if not pot and targetlanguage and targetlanguage not in inputstore.languages:
 137         print >> sys.stderr, "Warning: targetlanguage %s not found in inputfile (contains %s)" % (targetlanguage, ", ".join(inputstore.languages))
 138     convertor = oo2xliff(sourcelanguage, targetlanguage, blankmsgstr=pot, long_keys=multifilestyle!="single")
 139     outputstore = convertor.convertstore(inputstore, duplicatestyle)
 140     if outputstore.isempty():
 141         return 0
 142     outputfile.write(str(outputstore))
 143     return 1
 144
 145 def main(argv=None):
 146     from translate.convert import convert
 147     formats = {"oo":("xlf", convertoo), "sdf":("xlf", convertoo)}
 148     # always treat the input as an archive unless it is a directory
 149     archiveformats = {(None, "input"): oo.oomultifile}
 150     parser = convert.ArchiveConvertOptionParser(formats, usepots=False, description=__doc__, archiveformats=archiveformats)
 151     parser.add_option("-l", "--language", dest="targetlanguage", default=None,
 152             help="set target language to extract from oo file (e.g. af-ZA)", metavar="LANG")
 153     parser.add_option("", "--source-language", dest="sourcelanguage", default=None,
 154             help="set source language code (default en-US)", metavar="LANG")
 155     parser.add_option("", "--nonrecursiveinput", dest="allowrecursiveinput", default=True, action="store_false", help="don't treat the input oo as a recursive store")
 156     parser.add_duplicates_option()
 157     parser.add_multifile_option()
 158     parser.passthrough.append("sourcelanguage")
 159     parser.passthrough.append("targetlanguage")
 160     parser.verifyoptions = verifyoptions
 161     parser.run(argv)
 162
 163 if __name__ == '__main__':
 164     main()