convert/oo2po.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 Gettext PO localization files
  24
  25 See: http://translate.sourceforge.net/wiki/toolkit/oo2po for examples and
  26 usage instructions
  27 """
  28
  29 import os
  30 import sys
  31 from translate.storage import po
  32 from translate.storage import oo
  33
  34 # TODO: support using one GSI file as template, another as input (for when English is in one and translation in another)
  35
  36 class oo2po:
  37     def __init__(self, sourcelanguage, targetlanguage, blankmsgstr=False, long_keys=False):
  38         """construct an oo2po converter for the specified languages"""
  39         self.sourcelanguage = sourcelanguage
  40         self.targetlanguage = targetlanguage
  41         self.blankmsgstr = blankmsgstr
  42         self.long_keys = long_keys
  43
  44     def maketargetunit(self, part1, part2, translators_comment, key, subkey):
  45         """makes a base unit (.po or XLIFF) out of a subkey of two parts"""
  46         #TODO: Do better
  47         text1 = getattr(part1, subkey)
  48         if text1 == "":
  49             return None
  50         text2 = getattr(part2, subkey)
  51
  52         unit = po.pounit(text1.decode('utf-8'), encoding="UTF-8")
  53         unit.target = text2.decode('utf-8')
  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 = po.pofile()
 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         targetheader = thetargetfile.makeheader(charset="UTF-8", encoding="8bit", x_accelerator_marker="~", report_msgid_bugs_to=bug_url)
 111         targetheader.addnote("extracted from %s" % theoofile.filename, "developer")
 112         thetargetfile.addunit(targetheader)
 113         # go through the oo and convert each element
 114         for theoo in theoofile.units:
 115             unitlist = self.convertelement(theoo)
 116             for unit in unitlist:
 117                 thetargetfile.addunit(unit)
 118         thetargetfile.removeduplicates(duplicatestyle)
 119         return thetargetfile
 120
 121 def verifyoptions(options):
 122     """verifies the commandline options"""
 123     if not options.pot and not options.targetlanguage:
 124         raise ValueError("You must specify the target language unless generating POT files (-P)")
 125
 126 def convertoo(inputfile, outputfile, templates, pot=False, sourcelanguage=None, targetlanguage=None, duplicatestyle="msgid_comment", multifilestyle="single"):
 127     """reads in stdin using inputstore class, converts using convertorclass, writes to stdout"""
 128     inputstore = oo.oofile()
 129     if hasattr(inputfile, "filename"):
 130         inputfilename = inputfile.filename
 131     else:
 132         inputfilename = "(input file name not known)"
 133     inputstore.filename = inputfilename
 134     inputstore.parse(inputfile.read())
 135     if not sourcelanguage:
 136         testlangtype = targetlanguage or (inputstore and inputstore.languages[0]) or ""
 137         if testlangtype.isdigit():
 138             sourcelanguage = "01"
 139         else:
 140             sourcelanguage = "en-US"
 141     if not sourcelanguage in inputstore.languages:
 142         print >> sys.stderr, "Warning: sourcelanguage '%s' not found in inputfile '%s' (contains %s)" % (sourcelanguage, inputfilename, ", ".join(inputstore.languages))
 143     if targetlanguage and targetlanguage not in inputstore.languages:
 144         print >> sys.stderr, "Warning: targetlanguage '%s' not found in inputfile '%s' (contains %s)" % (targetlanguage, inputfilename, ", ".join(inputstore.languages))
 145     convertor = oo2po(sourcelanguage, targetlanguage, blankmsgstr=pot, long_keys=multifilestyle!="single")
 146     outputstore = convertor.convertstore(inputstore, duplicatestyle)
 147     if outputstore.isempty():
 148         return 0
 149     outputfile.write(str(outputstore))
 150     return 1
 151
 152 def main(argv=None):
 153     from translate.convert import convert
 154     formats = {"oo":("po", convertoo), "sdf":("po", convertoo)}
 155     # always treat the input as an archive unless it is a directory
 156     archiveformats = {(None, "input"): oo.oomultifile}
 157     parser = convert.ArchiveConvertOptionParser(formats, usepots=True, description=__doc__, archiveformats=archiveformats)
 158     parser.add_option("-l", "--language", dest="targetlanguage", default=None,
 159             help="set target language to extract from oo file (e.g. af-ZA)", metavar="LANG")
 160     parser.add_option("", "--source-language", dest="sourcelanguage", default=None,
 161             help="set source language code (default en-US)", metavar="LANG")
 162     parser.add_option("", "--nonrecursiveinput", dest="allowrecursiveinput", default=True, action="store_false", help="don't treat the input oo as a recursive store")
 163     parser.add_duplicates_option()
 164     parser.add_multifile_option()
 165     parser.passthrough.append("pot")
 166     parser.passthrough.append("sourcelanguage")
 167     parser.passthrough.append("targetlanguage")
 168     parser.verifyoptions = verifyoptions
 169     parser.run(argv)
 170
 171 if __name__ == '__main__':
 172     main()