cmlenz@1: # -*- coding: utf-8 -*-
cmlenz@1: #
cmlenz@1: # Copyright (C) 2007 Edgewall Software
cmlenz@1: # All rights reserved.
cmlenz@1: #
cmlenz@1: # This software is licensed as described in the file COPYING, which
cmlenz@1: # you should have received as part of this distribution. The terms
cmlenz@1: # are also available at http://babel.edgewall.org/wiki/License.
cmlenz@1: #
cmlenz@1: # This software consists of voluntary contributions made by many
cmlenz@1: # individuals. For the exact contribution history, see the revision
cmlenz@1: # history and logs, available at http://babel.edgewall.org/log/.
cmlenz@1: 
cmlenz@1: """Reading and writing of files in the ``gettext`` PO (portable object)
cmlenz@1: format.
cmlenz@1: 
cmlenz@1: :see: `The Format of PO Files
cmlenz@1:        <http://www.gnu.org/software/gettext/manual/gettext.html#PO-Files>`_
cmlenz@1: """
cmlenz@1: 
cmlenz@5: from datetime import date, datetime
cmlenz@134: import os
cmlenz@1: import re
cmlenz@1: 
cmlenz@1: from babel import __version__ as VERSION
cmlenz@199: from babel.messages.catalog import Catalog, Message
fschwarz@507: from babel.util import set, wraptext, LOCALTZ
cmlenz@1: 
cmlenz@178: __all__ = ['read_po', 'write_po']
cmlenz@161: __docformat__ = 'restructuredtext en'
cmlenz@158: 
cmlenz@158: def unescape(string):
cmlenz@158:     r"""Reverse `escape` the given string.
palgarvio@200: 
cmlenz@158:     >>> print unescape('"Say:\\n  \\"hello, world!\\"\\n"')
cmlenz@158:     Say:
cmlenz@158:       "hello, world!"
cmlenz@158:     <BLANKLINE>
palgarvio@200: 
cmlenz@158:     :param string: the string to unescape
cmlenz@158:     :return: the unescaped string
cmlenz@158:     :rtype: `str` or `unicode`
cmlenz@158:     """
cmlenz@158:     return string[1:-1].replace('\\\\', '\\') \
cmlenz@158:                        .replace('\\t', '\t') \
cmlenz@158:                        .replace('\\r', '\r') \
cmlenz@158:                        .replace('\\n', '\n') \
cmlenz@158:                        .replace('\\"', '\"')
cmlenz@158: 
cmlenz@158: def denormalize(string):
cmlenz@158:     r"""Reverse the normalization done by the `normalize` function.
palgarvio@200: 
cmlenz@158:     >>> print denormalize(r'''""
cmlenz@158:     ... "Say:\n"
cmlenz@158:     ... "  \"hello, world!\"\n"''')
cmlenz@158:     Say:
cmlenz@158:       "hello, world!"
cmlenz@158:     <BLANKLINE>
palgarvio@200: 
cmlenz@158:     >>> print denormalize(r'''""
cmlenz@158:     ... "Say:\n"
cmlenz@158:     ... "  \"Lorem ipsum dolor sit "
cmlenz@158:     ... "amet, consectetur adipisicing"
cmlenz@158:     ... " elit, \"\n"''')
cmlenz@158:     Say:
cmlenz@158:       "Lorem ipsum dolor sit amet, consectetur adipisicing elit, "
cmlenz@158:     <BLANKLINE>
palgarvio@200: 
cmlenz@158:     :param string: the string to denormalize
cmlenz@158:     :return: the denormalized string
cmlenz@158:     :rtype: `unicode` or `str`
cmlenz@158:     """
cmlenz@158:     if string.startswith('""'):
cmlenz@158:         lines = []
cmlenz@158:         for line in string.splitlines()[1:]:
cmlenz@158:             lines.append(unescape(line))
cmlenz@158:         return ''.join(lines)
cmlenz@158:     else:
cmlenz@158:         return unescape(string)
cmlenz@1: 
cmlenz@199: def read_po(fileobj, locale=None, domain=None, ignore_obsolete=False):
cmlenz@6:     """Read messages from a ``gettext`` PO (portable object) file from the given
cmlenz@64:     file-like object and return a `Catalog`.
palgarvio@200: 
cmlenz@6:     >>> from StringIO import StringIO
cmlenz@106:     >>> buf = StringIO('''
cmlenz@6:     ... #: main.py:1
cmlenz@6:     ... #, fuzzy, python-format
cmlenz@6:     ... msgid "foo %(name)s"
cmlenz@6:     ... msgstr ""
palgarvio@21:     ...
palgarvio@94:     ... # A user comment
palgarvio@94:     ... #. An auto comment
cmlenz@6:     ... #: main.py:3
cmlenz@6:     ... msgid "bar"
cmlenz@6:     ... msgid_plural "baz"
cmlenz@6:     ... msgstr[0] ""
cmlenz@6:     ... msgstr[1] ""
cmlenz@6:     ... ''')
cmlenz@64:     >>> catalog = read_po(buf)
cmlenz@104:     >>> catalog.revision_date = datetime(2007, 04, 01)
palgarvio@200: 
cmlenz@64:     >>> for message in catalog:
cmlenz@67:     ...     if message.id:
cmlenz@67:     ...         print (message.id, message.string)
palgarvio@105:     ...         print ' ', (message.locations, message.flags)
palgarvio@105:     ...         print ' ', (message.user_comments, message.auto_comments)
cmlenz@149:     (u'foo %(name)s', '')
cmlenz@149:       ([(u'main.py', 1)], set([u'fuzzy', u'python-format']))
palgarvio@105:       ([], [])
cmlenz@149:     ((u'bar', u'baz'), ('', ''))
cmlenz@149:       ([(u'main.py', 3)], set([]))
cmlenz@149:       ([u'A user comment'], [u'An auto comment'])
palgarvio@200: 
cmlenz@1:     :param fileobj: the file-like object to read the PO file from
cmlenz@196:     :param locale: the locale identifier or `Locale` object, or `None`
cmlenz@196:                    if the catalog is not bound to a locale (which basically
cmlenz@196:                    means it's a template)
cmlenz@196:     :param domain: the message domain
cmlenz@227:     :param ignore_obsolete: whether to ignore obsolete messages in the input
cmlenz@1:     :return: an iterator over ``(message, translation, location)`` tuples
cmlenz@1:     :rtype: ``iterator``
cmlenz@1:     """
cmlenz@196:     catalog = Catalog(locale=locale, domain=domain)
cmlenz@64: 
cmlenz@196:     counter = [0]
cmlenz@220:     offset = [0]
cmlenz@6:     messages = []
cmlenz@6:     translations = []
cmlenz@6:     locations = []
cmlenz@6:     flags = []
palgarvio@105:     user_comments = []
palgarvio@105:     auto_comments = []
cmlenz@199:     obsolete = [False]
cmlenz@199:     in_msgid = [False]
cmlenz@199:     in_msgstr = [False]
cmlenz@6: 
cmlenz@64:     def _add_message():
cmlenz@6:         translations.sort()
cmlenz@64:         if len(messages) > 1:
cmlenz@106:             msgid = tuple([denormalize(m) for m in messages])
cmlenz@64:         else:
cmlenz@106:             msgid = denormalize(messages[0])
cmlenz@379:         if isinstance(msgid, (list, tuple)):
cmlenz@379:             string = []
cmlenz@379:             for idx in range(catalog.num_plurals):
cmlenz@379:                 try:
cmlenz@379:                     string.append(translations[idx])
cmlenz@379:                 except IndexError:
cmlenz@379:                     string.append((idx, ''))
cmlenz@379:             string = tuple([denormalize(t[1]) for t in string])
cmlenz@64:         else:
cmlenz@106:             string = denormalize(translations[0][1])
cmlenz@199:         message = Message(msgid, string, list(locations), set(flags),
cmlenz@227:                           auto_comments, user_comments, lineno=offset[0] + 1)
cmlenz@199:         if obsolete[0]:
cmlenz@199:             if not ignore_obsolete:
cmlenz@199:                 catalog.obsolete[msgid] = message
cmlenz@199:         else:
cmlenz@199:             catalog[msgid] = message
cmlenz@84:         del messages[:]; del translations[:]; del locations[:];
palgarvio@105:         del flags[:]; del auto_comments[:]; del user_comments[:]
cmlenz@199:         obsolete[0] = False
cmlenz@196:         counter[0] += 1
cmlenz@6: 
cmlenz@220:     def _process_message_line(lineno, line):
cmlenz@199:         if line.startswith('msgid_plural'):
cmlenz@199:             in_msgid[0] = True
cmlenz@199:             msg = line[12:].lstrip()
cmlenz@199:             messages.append(msg)
cmlenz@199:         elif line.startswith('msgid'):
cmlenz@199:             in_msgid[0] = True
cmlenz@220:             offset[0] = lineno
cmlenz@199:             txt = line[5:].lstrip()
cmlenz@199:             if messages:
cmlenz@199:                 _add_message()
cmlenz@199:             messages.append(txt)
cmlenz@199:         elif line.startswith('msgstr'):
cmlenz@199:             in_msgid[0] = False
cmlenz@199:             in_msgstr[0] = True
cmlenz@199:             msg = line[6:].lstrip()
cmlenz@199:             if msg.startswith('['):
jruigrok@448:                 idx, msg = msg[1:].split(']', 1)
cmlenz@199:                 translations.append([int(idx), msg.lstrip()])
cmlenz@199:             else:
cmlenz@199:                 translations.append([0, msg])
cmlenz@199:         elif line.startswith('"'):
cmlenz@199:             if in_msgid[0]:
cmlenz@199:                 messages[-1] += u'\n' + line.rstrip()
cmlenz@199:             elif in_msgstr[0]:
cmlenz@199:                 translations[-1][1] += u'\n' + line.rstrip()
cmlenz@199: 
cmlenz@220:     for lineno, line in enumerate(fileobj.readlines()):
fschwarz@507:         line = line.strip()
fschwarz@507:         if not isinstance(line, unicode):
fschwarz@507:             line = line.decode(catalog.charset)
cmlenz@1:         if line.startswith('#'):
cmlenz@199:             in_msgid[0] = in_msgstr[0] = False
cmlenz@199:             if messages and translations:
cmlenz@106:                 _add_message()
cmlenz@106:             if line[1:].startswith(':'):
cmlenz@106:                 for location in line[2:].lstrip().split():
cmlenz@369:                     pos = location.rfind(':')
cmlenz@369:                     if pos >= 0:
cmlenz@369:                         try:
cmlenz@369:                             lineno = int(location[pos + 1:])
cmlenz@369:                         except ValueError:
cmlenz@369:                             continue
cmlenz@369:                         locations.append((location[:pos], lineno))
cmlenz@106:             elif line[1:].startswith(','):
cmlenz@106:                 for flag in line[2:].lstrip().split(','):
cmlenz@106:                     flags.append(flag.strip())
cmlenz@199:             elif line[1:].startswith('~'):
cmlenz@199:                 obsolete[0] = True
cmlenz@220:                 _process_message_line(lineno, line[2:].lstrip())
cmlenz@106:             elif line[1:].startswith('.'):
cmlenz@106:                 # These are called auto-comments
cmlenz@106:                 comment = line[2:].strip()
cmlenz@199:                 if comment: # Just check that we're not adding empty comments
cmlenz@106:                     auto_comments.append(comment)
cmlenz@120:             else:
cmlenz@106:                 # These are called user comments
cmlenz@120:                 user_comments.append(line[1:].strip())
cmlenz@104:         else:
cmlenz@220:             _process_message_line(lineno, line)
cmlenz@6: 
cmlenz@6:     if messages:
cmlenz@64:         _add_message()
cmlenz@196: 
cmlenz@196:     # No actual messages found, but there was some info in comments, from which
cmlenz@196:     # we'll construct an empty header message
cmlenz@196:     elif not counter[0] and (flags or user_comments or auto_comments):
cmlenz@196:         messages.append(u'')
cmlenz@196:         translations.append([0, u''])
cmlenz@196:         _add_message()
cmlenz@196: 
cmlenz@64:     return catalog
cmlenz@1: 
cmlenz@24: WORD_SEP = re.compile('('
cmlenz@24:     r'\s+|'                                 # any whitespace
cmlenz@24:     r'[^\s\w]*\w+[a-zA-Z]-(?=\w+[a-zA-Z])|' # hyphenated words
cmlenz@24:     r'(?<=[\w\!\"\'\&\.\,\?])-{2,}(?=\w)'   # em-dash
cmlenz@24: ')')
cmlenz@24: 
cmlenz@24: def escape(string):
cmlenz@24:     r"""Escape the given string so that it can be included in double-quoted
cmlenz@24:     strings in ``PO`` files.
palgarvio@200: 
cmlenz@24:     >>> escape('''Say:
cmlenz@24:     ...   "hello, world!"
cmlenz@24:     ... ''')
cmlenz@24:     '"Say:\\n  \\"hello, world!\\"\\n"'
palgarvio@200: 
cmlenz@24:     :param string: the string to escape
cmlenz@24:     :return: the escaped string
cmlenz@24:     :rtype: `str` or `unicode`
cmlenz@24:     """
cmlenz@24:     return '"%s"' % string.replace('\\', '\\\\') \
cmlenz@24:                           .replace('\t', '\\t') \
cmlenz@24:                           .replace('\r', '\\r') \
cmlenz@24:                           .replace('\n', '\\n') \
cmlenz@24:                           .replace('\"', '\\"')
cmlenz@24: 
cmlenz@190: def normalize(string, prefix='', width=76):
cmlenz@106:     r"""Convert a string into a format that is appropriate for .po files.
palgarvio@200: 
cmlenz@24:     >>> print normalize('''Say:
cmlenz@24:     ...   "hello, world!"
cmlenz@24:     ... ''', width=None)
cmlenz@24:     ""
cmlenz@24:     "Say:\n"
cmlenz@24:     "  \"hello, world!\"\n"
palgarvio@200: 
cmlenz@24:     >>> print normalize('''Say:
cmlenz@24:     ...   "Lorem ipsum dolor sit amet, consectetur adipisicing elit, "
cmlenz@24:     ... ''', width=32)
cmlenz@24:     ""
cmlenz@24:     "Say:\n"
cmlenz@24:     "  \"Lorem ipsum dolor sit "
cmlenz@24:     "amet, consectetur adipisicing"
cmlenz@24:     " elit, \"\n"
palgarvio@200: 
cmlenz@24:     :param string: the string to normalize
cmlenz@190:     :param prefix: a string that should be prepended to every line
cmlenz@24:     :param width: the maximum line width; use `None`, 0, or a negative number
cmlenz@24:                   to completely disable line wrapping
cmlenz@24:     :return: the normalized string
cmlenz@24:     :rtype: `unicode`
cmlenz@24:     """
cmlenz@24:     if width and width > 0:
cmlenz@190:         prefixlen = len(prefix)
cmlenz@24:         lines = []
cmlenz@24:         for idx, line in enumerate(string.splitlines(True)):
cmlenz@190:             if len(escape(line)) + prefixlen > width:
cmlenz@24:                 chunks = WORD_SEP.split(line)
cmlenz@24:                 chunks.reverse()
cmlenz@24:                 while chunks:
cmlenz@24:                     buf = []
cmlenz@24:                     size = 2
cmlenz@24:                     while chunks:
cmlenz@190:                         l = len(escape(chunks[-1])) - 2 + prefixlen
cmlenz@24:                         if size + l < width:
cmlenz@24:                             buf.append(chunks.pop())
cmlenz@24:                             size += l
cmlenz@24:                         else:
cmlenz@24:                             if not buf:
cmlenz@24:                                 # handle long chunks by putting them on a
cmlenz@24:                                 # separate line
cmlenz@24:                                 buf.append(chunks.pop())
cmlenz@24:                             break
cmlenz@24:                     lines.append(u''.join(buf))
cmlenz@24:             else:
cmlenz@24:                 lines.append(line)
cmlenz@24:     else:
cmlenz@24:         lines = string.splitlines(True)
cmlenz@24: 
cmlenz@67:     if len(lines) <= 1:
cmlenz@24:         return escape(string)
cmlenz@24: 
cmlenz@24:     # Remove empty trailing line
cmlenz@67:     if lines and not lines[-1]:
cmlenz@24:         del lines[-1]
cmlenz@24:         lines[-1] += '\n'
cmlenz@190:     return u'""\n' + u'\n'.join([(prefix + escape(l)) for l in lines])
cmlenz@24: 
cmlenz@104: def write_po(fileobj, catalog, width=76, no_location=False, omit_header=False,
palgarvio@200:              sort_output=False, sort_by_file=False, ignore_obsolete=False,
cmlenz@203:              include_previous=False):
cmlenz@56:     r"""Write a ``gettext`` PO (portable object) template file for a given
cmlenz@56:     message catalog to the provided file-like object.
palgarvio@200: 
cmlenz@56:     >>> catalog = Catalog()
cmlenz@56:     >>> catalog.add(u'foo %(name)s', locations=[('main.py', 1)],
cmlenz@56:     ...             flags=('fuzzy',))
cmlenz@56:     >>> catalog.add((u'bar', u'baz'), locations=[('main.py', 3)])
cmlenz@1:     >>> from StringIO import StringIO
cmlenz@1:     >>> buf = StringIO()
cmlenz@104:     >>> write_po(buf, catalog, omit_header=True)
cmlenz@1:     >>> print buf.getvalue()
cmlenz@1:     #: main.py:1
cmlenz@6:     #, fuzzy, python-format
cmlenz@6:     msgid "foo %(name)s"
cmlenz@1:     msgstr ""
cmlenz@1:     <BLANKLINE>
cmlenz@1:     #: main.py:3
cmlenz@1:     msgid "bar"
cmlenz@1:     msgid_plural "baz"
cmlenz@1:     msgstr[0] ""
cmlenz@1:     msgstr[1] ""
cmlenz@1:     <BLANKLINE>
cmlenz@1:     <BLANKLINE>
palgarvio@200: 
cmlenz@1:     :param fileobj: the file-like object to write to
cmlenz@67:     :param catalog: the `Catalog` instance
cmlenz@24:     :param width: the maximum line width for the generated output; use `None`,
cmlenz@24:                   0, or a negative number to completely disable line wrapping
cmlenz@1:     :param no_location: do not emit a location comment for every message
cmlenz@1:     :param omit_header: do not include the ``msgid ""`` entry at the top of the
cmlenz@1:                         output
cmlenz@227:     :param sort_output: whether to sort the messages in the output by msgid
cmlenz@227:     :param sort_by_file: whether to sort the messages in the output by their
cmlenz@227:                          locations
cmlenz@227:     :param ignore_obsolete: whether to ignore obsolete messages and not include
cmlenz@227:                             them in the output; by default they are included as
cmlenz@227:                             comments
cmlenz@203:     :param include_previous: include the old msgid as a comment when
cmlenz@229:                              updating the catalog
cmlenz@1:     """
cmlenz@190:     def _normalize(key, prefix=''):
cmlenz@190:         return normalize(key, prefix=prefix, width=width) \
cmlenz@190:             .encode(catalog.charset, 'backslashreplace')
cmlenz@24: 
cmlenz@24:     def _write(text):
cmlenz@24:         if isinstance(text, unicode):
cmlenz@102:             text = text.encode(catalog.charset)
cmlenz@24:         fileobj.write(text)
cmlenz@1: 
cmlenz@181:     def _write_comment(comment, prefix=''):
jruigrok@476:         # xgettext always wraps comments even if --no-wrap is passed;
jruigrok@476:         # provide the same behaviour
cmlenz@181:         if width and width > 0:
jruigrok@476:             _width = width
jruigrok@476:         else:
jruigrok@476:             _width = 76
jruigrok@476:         for line in wraptext(comment, _width):
cmlenz@181:             _write('#%s %s\n' % (prefix, line.strip()))
cmlenz@181: 
cmlenz@181:     def _write_message(message, prefix=''):
cmlenz@181:         if isinstance(message.id, (list, tuple)):
cmlenz@190:             _write('%smsgid %s\n' % (prefix, _normalize(message.id[0], prefix)))
cmlenz@190:             _write('%smsgid_plural %s\n' % (
cmlenz@190:                 prefix, _normalize(message.id[1], prefix)
cmlenz@190:             ))
cmlenz@379: 
cmlenz@379:             for idx in range(catalog.num_plurals):
cmlenz@379:                 try:
cmlenz@379:                     string = message.string[idx]
cmlenz@379:                 except IndexError:
cmlenz@379:                     string = ''
cmlenz@190:                 _write('%smsgstr[%d] %s\n' % (
cmlenz@379:                     prefix, idx, _normalize(string, prefix)
cmlenz@190:                 ))
cmlenz@181:         else:
cmlenz@190:             _write('%smsgid %s\n' % (prefix, _normalize(message.id, prefix)))
cmlenz@190:             _write('%smsgstr %s\n' % (
cmlenz@190:                 prefix, _normalize(message.string or '', prefix)
cmlenz@190:             ))
cmlenz@181: 
cmlenz@104:     messages = list(catalog)
palgarvio@71:     if sort_output:
pjenvey@248:         messages.sort()
palgarvio@71:     elif sort_by_file:
palgarvio@71:         messages.sort(lambda x,y: cmp(x.locations, y.locations))
cmlenz@68: 
palgarvio@71:     for message in messages:
cmlenz@67:         if not message.id: # This is the header "message"
cmlenz@67:             if omit_header:
cmlenz@67:                 continue
cmlenz@104:             comment_header = catalog.header_comment
cmlenz@103:             if width and width > 0:
cmlenz@103:                 lines = []
cmlenz@104:                 for line in comment_header.splitlines():
cmlenz@316:                     lines += wraptext(line, width=width,
cmlenz@316:                                       subsequent_indent='# ')
cmlenz@104:                 comment_header = u'\n'.join(lines) + u'\n'
cmlenz@104:             _write(comment_header)
cmlenz@102: 
cmlenz@227:         for comment in message.user_comments:
cmlenz@181:             _write_comment(comment)
cmlenz@227:         for comment in message.auto_comments:
cmlenz@181:             _write_comment(comment, prefix='.')
cmlenz@1: 
cmlenz@1:         if not no_location:
cmlenz@134:             locs = u' '.join([u'%s:%d' % (filename.replace(os.sep, '/'), lineno)
cmlenz@134:                               for filename, lineno in message.locations])
cmlenz@181:             _write_comment(locs, prefix=':')
cmlenz@56:         if message.flags:
cmlenz@56:             _write('#%s\n' % ', '.join([''] + list(message.flags)))
cmlenz@24: 
cmlenz@203:         if message.previous_id and include_previous:
cmlenz@310:             _write_comment('msgid %s' % _normalize(message.previous_id[0]),
cmlenz@203:                            prefix='|')
cmlenz@203:             if len(message.previous_id) > 1:
cmlenz@310:                 _write_comment('msgid_plural %s' % _normalize(
cmlenz@203:                     message.previous_id[1]
cmlenz@203:                 ), prefix='|')
palgarvio@200: 
cmlenz@181:         _write_message(message)
cmlenz@24:         _write('\n')
cmlenz@181: 
cmlenz@191:     if not ignore_obsolete:
cmlenz@191:         for message in catalog.obsolete.values():
cmlenz@227:             for comment in message.user_comments:
cmlenz@191:                 _write_comment(comment)
cmlenz@191:             _write_message(message, prefix='#~ ')
cmlenz@191:             _write('\n')