openmedialibrary_platform/Darwin/lib/python3.5/gettext.py

540 lines
18 KiB
Python
Raw Normal View History

2013-10-11 17:28:32 +00:00
"""Internationalization and localization support.
This module provides internationalization (I18N) and localization (L10N)
support for your Python programs by providing an interface to the GNU gettext
message catalog library.
I18N refers to the operation by which a program is made aware of multiple
languages. L10N refers to the adaptation of your program, once
internationalized, to the local language and cultural habits.
"""
# This module represents the integration of work, contributions, feedback, and
# suggestions from the following people:
#
# Martin von Loewis, who wrote the initial implementation of the underlying
# C-based libintlmodule (later renamed _gettext), along with a skeletal
# gettext.py implementation.
#
# Peter Funk, who wrote fintl.py, a fairly complete wrapper around intlmodule,
# which also included a pure-Python implementation to read .mo files if
# intlmodule wasn't available.
#
# James Henstridge, who also wrote a gettext.py module, which has some
# interesting, but currently unsupported experimental features: the notion of
# a Catalog class and instances, and the ability to add to a catalog file via
# a Python API.
#
# Barry Warsaw integrated these modules, wrote the .install() API and code,
# and conformed all C and Python code to Python's coding standards.
#
# Francois Pinard and Marc-Andre Lemburg also contributed valuably to this
# module.
#
# J. David Ibanez implemented plural forms. Bruno Haible fixed some bugs.
#
# TODO:
# - Lazy loading of .mo files. Currently the entire catalog is loaded into
# memory, but that's probably bad for large translated programs. Instead,
# the lexical sort of original strings in GNU .mo files should be exploited
# to do binary searches and lazy initializations. Or you might want to use
# the undocumented double-hash algorithm for .mo files with hash tables, but
# you'll need to study the GNU gettext code to do this.
#
# - Support Solaris .mo file formats. Unfortunately, we've been unable to
# find this format documented anywhere.
2014-09-30 16:15:32 +00:00
import locale, copy, io, os, re, struct, sys
2013-10-11 17:28:32 +00:00
from errno import ENOENT
__all__ = ['NullTranslations', 'GNUTranslations', 'Catalog',
'find', 'translation', 'install', 'textdomain', 'bindtextdomain',
2015-11-25 01:25:01 +00:00
'bind_textdomain_codeset',
'dgettext', 'dngettext', 'gettext', 'lgettext', 'ldgettext',
'ldngettext', 'lngettext', 'ngettext',
2013-10-11 17:28:32 +00:00
]
2014-09-30 16:15:32 +00:00
_default_localedir = os.path.join(sys.base_prefix, 'share', 'locale')
2013-10-11 17:28:32 +00:00
def c2py(plural):
"""Gets a C expression as used in PO files for plural forms and returns a
Python lambda function that implements an equivalent expression.
"""
# Security check, allow only the "n" identifier
import token, tokenize
2014-09-30 16:15:32 +00:00
tokens = tokenize.generate_tokens(io.StringIO(plural).readline)
2013-10-11 17:28:32 +00:00
try:
danger = [x for x in tokens if x[0] == token.NAME and x[1] != 'n']
except tokenize.TokenError:
2014-09-30 16:15:32 +00:00
raise ValueError('plural forms expression error, maybe unbalanced parenthesis')
2013-10-11 17:28:32 +00:00
else:
if danger:
2014-09-30 16:15:32 +00:00
raise ValueError('plural forms expression could be dangerous')
2013-10-11 17:28:32 +00:00
# Replace some C operators by their Python equivalents
plural = plural.replace('&&', ' and ')
plural = plural.replace('||', ' or ')
expr = re.compile(r'\!([^=])')
plural = expr.sub(' not \\1', plural)
# Regular expression and replacement function used to transform
2014-09-30 16:15:32 +00:00
# "a?b:c" to "b if a else c".
2013-10-11 17:28:32 +00:00
expr = re.compile(r'(.*?)\?(.*?):(.*)')
def repl(x):
2014-09-30 16:15:32 +00:00
return "(%s if %s else %s)" % (x.group(2), x.group(1),
expr.sub(repl, x.group(3)))
2013-10-11 17:28:32 +00:00
# Code to transform the plural expression, taking care of parentheses
stack = ['']
for c in plural:
if c == '(':
stack.append('')
elif c == ')':
if len(stack) == 1:
# Actually, we never reach this code, because unbalanced
# parentheses get caught in the security check at the
# beginning.
2014-09-30 16:15:32 +00:00
raise ValueError('unbalanced parenthesis in plural form')
2013-10-11 17:28:32 +00:00
s = expr.sub(repl, stack.pop())
stack[-1] += '(%s)' % s
else:
stack[-1] += c
plural = expr.sub(repl, stack.pop())
return eval('lambda n: int(%s)' % plural)
2014-09-30 16:15:32 +00:00
def _expand_lang(loc):
loc = locale.normalize(loc)
2013-10-11 17:28:32 +00:00
COMPONENT_CODESET = 1 << 0
COMPONENT_TERRITORY = 1 << 1
COMPONENT_MODIFIER = 1 << 2
# split up the locale into its base components
mask = 0
2014-09-30 16:15:32 +00:00
pos = loc.find('@')
2013-10-11 17:28:32 +00:00
if pos >= 0:
2014-09-30 16:15:32 +00:00
modifier = loc[pos:]
loc = loc[:pos]
2013-10-11 17:28:32 +00:00
mask |= COMPONENT_MODIFIER
else:
modifier = ''
2014-09-30 16:15:32 +00:00
pos = loc.find('.')
2013-10-11 17:28:32 +00:00
if pos >= 0:
2014-09-30 16:15:32 +00:00
codeset = loc[pos:]
loc = loc[:pos]
2013-10-11 17:28:32 +00:00
mask |= COMPONENT_CODESET
else:
codeset = ''
2014-09-30 16:15:32 +00:00
pos = loc.find('_')
2013-10-11 17:28:32 +00:00
if pos >= 0:
2014-09-30 16:15:32 +00:00
territory = loc[pos:]
loc = loc[:pos]
2013-10-11 17:28:32 +00:00
mask |= COMPONENT_TERRITORY
else:
territory = ''
2014-09-30 16:15:32 +00:00
language = loc
2013-10-11 17:28:32 +00:00
ret = []
for i in range(mask+1):
if not (i & ~mask): # if all components for this combo exist ...
val = language
if i & COMPONENT_TERRITORY: val += territory
if i & COMPONENT_CODESET: val += codeset
if i & COMPONENT_MODIFIER: val += modifier
ret.append(val)
ret.reverse()
return ret
class NullTranslations:
def __init__(self, fp=None):
self._info = {}
self._charset = None
self._output_charset = None
self._fallback = None
if fp is not None:
self._parse(fp)
def _parse(self, fp):
pass
def add_fallback(self, fallback):
if self._fallback:
self._fallback.add_fallback(fallback)
else:
self._fallback = fallback
def gettext(self, message):
if self._fallback:
return self._fallback.gettext(message)
return message
def lgettext(self, message):
if self._fallback:
return self._fallback.lgettext(message)
return message
def ngettext(self, msgid1, msgid2, n):
if self._fallback:
return self._fallback.ngettext(msgid1, msgid2, n)
if n == 1:
return msgid1
else:
return msgid2
def lngettext(self, msgid1, msgid2, n):
if self._fallback:
return self._fallback.lngettext(msgid1, msgid2, n)
if n == 1:
return msgid1
else:
return msgid2
def info(self):
return self._info
def charset(self):
return self._charset
def output_charset(self):
return self._output_charset
def set_output_charset(self, charset):
self._output_charset = charset
2014-09-30 16:15:32 +00:00
def install(self, names=None):
import builtins
builtins.__dict__['_'] = self.gettext
2013-10-11 17:28:32 +00:00
if hasattr(names, "__contains__"):
if "gettext" in names:
2014-09-30 16:15:32 +00:00
builtins.__dict__['gettext'] = builtins.__dict__['_']
2013-10-11 17:28:32 +00:00
if "ngettext" in names:
2014-09-30 16:15:32 +00:00
builtins.__dict__['ngettext'] = self.ngettext
2013-10-11 17:28:32 +00:00
if "lgettext" in names:
2014-09-30 16:15:32 +00:00
builtins.__dict__['lgettext'] = self.lgettext
2013-10-11 17:28:32 +00:00
if "lngettext" in names:
2014-09-30 16:15:32 +00:00
builtins.__dict__['lngettext'] = self.lngettext
2013-10-11 17:28:32 +00:00
class GNUTranslations(NullTranslations):
# Magic number of .mo files
2014-09-30 16:15:32 +00:00
LE_MAGIC = 0x950412de
BE_MAGIC = 0xde120495
2013-10-11 17:28:32 +00:00
2015-11-25 01:25:01 +00:00
# Acceptable .mo versions
VERSIONS = (0, 1)
def _get_versions(self, version):
"""Returns a tuple of major version, minor version"""
return (version >> 16, version & 0xffff)
2013-10-11 17:28:32 +00:00
def _parse(self, fp):
"""Override this method to support alternative .mo formats."""
unpack = struct.unpack
filename = getattr(fp, 'name', '')
# Parse the .mo file header, which consists of 5 little endian 32
# bit words.
self._catalog = catalog = {}
self.plural = lambda n: int(n != 1) # germanic plural by default
buf = fp.read()
buflen = len(buf)
# Are we big endian or little endian?
magic = unpack('<I', buf[:4])[0]
if magic == self.LE_MAGIC:
version, msgcount, masteridx, transidx = unpack('<4I', buf[4:20])
ii = '<II'
elif magic == self.BE_MAGIC:
version, msgcount, masteridx, transidx = unpack('>4I', buf[4:20])
ii = '>II'
else:
2014-09-30 16:15:32 +00:00
raise OSError(0, 'Bad magic number', filename)
2015-11-25 01:25:01 +00:00
major_version, minor_version = self._get_versions(version)
if major_version not in self.VERSIONS:
raise OSError(0, 'Bad version number ' + str(major_version), filename)
2013-10-11 17:28:32 +00:00
# Now put all messages from the .mo file buffer into the catalog
# dictionary.
2014-09-30 16:15:32 +00:00
for i in range(0, msgcount):
2013-10-11 17:28:32 +00:00
mlen, moff = unpack(ii, buf[masteridx:masteridx+8])
mend = moff + mlen
tlen, toff = unpack(ii, buf[transidx:transidx+8])
tend = toff + tlen
if mend < buflen and tend < buflen:
msg = buf[moff:mend]
tmsg = buf[toff:tend]
else:
2014-09-30 16:15:32 +00:00
raise OSError(0, 'File is corrupt', filename)
2013-10-11 17:28:32 +00:00
# See if we're looking at GNU .mo conventions for metadata
if mlen == 0:
# Catalog description
2015-11-25 01:25:01 +00:00
lastk = None
2014-09-30 16:15:32 +00:00
for b_item in tmsg.split('\n'.encode("ascii")):
item = b_item.decode().strip()
2013-10-11 17:28:32 +00:00
if not item:
continue
2015-11-25 01:25:01 +00:00
k = v = None
2013-10-11 17:28:32 +00:00
if ':' in item:
k, v = item.split(':', 1)
k = k.strip().lower()
v = v.strip()
self._info[k] = v
lastk = k
elif lastk:
self._info[lastk] += '\n' + item
if k == 'content-type':
self._charset = v.split('charset=')[1]
elif k == 'plural-forms':
v = v.split(';')
plural = v[1].split('plural=')[1]
self.plural = c2py(plural)
# Note: we unconditionally convert both msgids and msgstrs to
# Unicode using the character encoding specified in the charset
# parameter of the Content-Type header. The gettext documentation
# strongly encourages msgids to be us-ascii, but some applications
# require alternative encodings (e.g. Zope's ZCML and ZPT). For
# traditional gettext applications, the msgid conversion will
# cause no problems since us-ascii should always be a subset of
# the charset encoding. We may want to fall back to 8-bit msgids
# if the Unicode conversion fails.
2014-09-30 16:15:32 +00:00
charset = self._charset or 'ascii'
if b'\x00' in msg:
2013-10-11 17:28:32 +00:00
# Plural forms
2014-09-30 16:15:32 +00:00
msgid1, msgid2 = msg.split(b'\x00')
tmsg = tmsg.split(b'\x00')
msgid1 = str(msgid1, charset)
for i, x in enumerate(tmsg):
catalog[(msgid1, i)] = str(x, charset)
2013-10-11 17:28:32 +00:00
else:
2014-09-30 16:15:32 +00:00
catalog[str(msg, charset)] = str(tmsg, charset)
2013-10-11 17:28:32 +00:00
# advance to next entry in the seek tables
masteridx += 8
transidx += 8
def lgettext(self, message):
missing = object()
tmsg = self._catalog.get(message, missing)
if tmsg is missing:
if self._fallback:
return self._fallback.lgettext(message)
return message
if self._output_charset:
return tmsg.encode(self._output_charset)
return tmsg.encode(locale.getpreferredencoding())
def lngettext(self, msgid1, msgid2, n):
try:
tmsg = self._catalog[(msgid1, self.plural(n))]
if self._output_charset:
return tmsg.encode(self._output_charset)
return tmsg.encode(locale.getpreferredencoding())
except KeyError:
if self._fallback:
return self._fallback.lngettext(msgid1, msgid2, n)
if n == 1:
return msgid1
else:
return msgid2
2014-09-30 16:15:32 +00:00
def gettext(self, message):
2013-10-11 17:28:32 +00:00
missing = object()
tmsg = self._catalog.get(message, missing)
if tmsg is missing:
if self._fallback:
2014-09-30 16:15:32 +00:00
return self._fallback.gettext(message)
return message
2013-10-11 17:28:32 +00:00
return tmsg
2014-09-30 16:15:32 +00:00
def ngettext(self, msgid1, msgid2, n):
2013-10-11 17:28:32 +00:00
try:
tmsg = self._catalog[(msgid1, self.plural(n))]
except KeyError:
if self._fallback:
2014-09-30 16:15:32 +00:00
return self._fallback.ngettext(msgid1, msgid2, n)
2013-10-11 17:28:32 +00:00
if n == 1:
2014-09-30 16:15:32 +00:00
tmsg = msgid1
2013-10-11 17:28:32 +00:00
else:
2014-09-30 16:15:32 +00:00
tmsg = msgid2
2013-10-11 17:28:32 +00:00
return tmsg
# Locate a .mo file using the gettext strategy
2014-09-30 16:15:32 +00:00
def find(domain, localedir=None, languages=None, all=False):
2013-10-11 17:28:32 +00:00
# Get some reasonable defaults for arguments that were not supplied
if localedir is None:
localedir = _default_localedir
if languages is None:
languages = []
for envar in ('LANGUAGE', 'LC_ALL', 'LC_MESSAGES', 'LANG'):
val = os.environ.get(envar)
if val:
languages = val.split(':')
break
if 'C' not in languages:
languages.append('C')
# now normalize and expand the languages
nelangs = []
for lang in languages:
for nelang in _expand_lang(lang):
if nelang not in nelangs:
nelangs.append(nelang)
# select a language
if all:
result = []
else:
result = None
for lang in nelangs:
if lang == 'C':
break
mofile = os.path.join(localedir, lang, 'LC_MESSAGES', '%s.mo' % domain)
if os.path.exists(mofile):
if all:
result.append(mofile)
else:
return mofile
return result
# a mapping between absolute .mo file path and Translation object
_translations = {}
def translation(domain, localedir=None, languages=None,
class_=None, fallback=False, codeset=None):
if class_ is None:
class_ = GNUTranslations
2014-09-30 16:15:32 +00:00
mofiles = find(domain, localedir, languages, all=True)
2013-10-11 17:28:32 +00:00
if not mofiles:
if fallback:
return NullTranslations()
2014-09-30 16:15:32 +00:00
raise OSError(ENOENT, 'No translation file found for domain', domain)
2013-10-11 17:28:32 +00:00
# Avoid opening, reading, and parsing the .mo file after it's been done
# once.
result = None
for mofile in mofiles:
key = (class_, os.path.abspath(mofile))
t = _translations.get(key)
if t is None:
with open(mofile, 'rb') as fp:
t = _translations.setdefault(key, class_(fp))
# Copy the translation object to allow setting fallbacks and
# output charset. All other instance data is shared with the
# cached object.
t = copy.copy(t)
if codeset:
t.set_output_charset(codeset)
if result is None:
result = t
else:
result.add_fallback(t)
return result
2014-09-30 16:15:32 +00:00
def install(domain, localedir=None, codeset=None, names=None):
2013-10-11 17:28:32 +00:00
t = translation(domain, localedir, fallback=True, codeset=codeset)
2014-09-30 16:15:32 +00:00
t.install(names)
2013-10-11 17:28:32 +00:00
# a mapping b/w domains and locale directories
_localedirs = {}
# a mapping b/w domains and codesets
_localecodesets = {}
# current global domain, `messages' used for compatibility w/ GNU gettext
_current_domain = 'messages'
def textdomain(domain=None):
global _current_domain
if domain is not None:
_current_domain = domain
return _current_domain
def bindtextdomain(domain, localedir=None):
global _localedirs
if localedir is not None:
_localedirs[domain] = localedir
return _localedirs.get(domain, _default_localedir)
def bind_textdomain_codeset(domain, codeset=None):
global _localecodesets
if codeset is not None:
_localecodesets[domain] = codeset
return _localecodesets.get(domain)
def dgettext(domain, message):
try:
t = translation(domain, _localedirs.get(domain, None),
codeset=_localecodesets.get(domain))
2014-09-30 16:15:32 +00:00
except OSError:
2013-10-11 17:28:32 +00:00
return message
return t.gettext(message)
def ldgettext(domain, message):
try:
t = translation(domain, _localedirs.get(domain, None),
codeset=_localecodesets.get(domain))
2014-09-30 16:15:32 +00:00
except OSError:
2013-10-11 17:28:32 +00:00
return message
return t.lgettext(message)
def dngettext(domain, msgid1, msgid2, n):
try:
t = translation(domain, _localedirs.get(domain, None),
codeset=_localecodesets.get(domain))
2014-09-30 16:15:32 +00:00
except OSError:
2013-10-11 17:28:32 +00:00
if n == 1:
return msgid1
else:
return msgid2
return t.ngettext(msgid1, msgid2, n)
def ldngettext(domain, msgid1, msgid2, n):
try:
t = translation(domain, _localedirs.get(domain, None),
codeset=_localecodesets.get(domain))
2014-09-30 16:15:32 +00:00
except OSError:
2013-10-11 17:28:32 +00:00
if n == 1:
return msgid1
else:
return msgid2
return t.lngettext(msgid1, msgid2, n)
def gettext(message):
return dgettext(_current_domain, message)
def lgettext(message):
return ldgettext(_current_domain, message)
def ngettext(msgid1, msgid2, n):
return dngettext(_current_domain, msgid1, msgid2, n)
def lngettext(msgid1, msgid2, n):
return ldngettext(_current_domain, msgid1, msgid2, n)
# dcgettext() has been deemed unnecessary and is not implemented.
# James Henstridge's Catalog constructor from GNOME gettext. Documented usage
# was:
#
# import gettext
# cat = gettext.Catalog(PACKAGE, localedir=LOCALEDIR)
# _ = cat.gettext
# print _('Hello World')
# The resulting catalog object currently don't support access through a
# dictionary API, which was supported (but apparently unused) in GNOME
# gettext.
Catalog = translation