anonfunc/JellyPy
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Initial Commit

Browse Source
This commit is contained in:
Tim
2015-02-22 18:32:50 +02:00
commit 88daa3fb91
1311 changed files with 256240 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
lib/mutagen/README.rst Normal file
View File

@@ -0,0 +1,58 @@
Mutagen
=======
Mutagen is a Python module to handle audio metadata. It supports ASF, FLAC,
M4A, Monkey's Audio, MP3, Musepack, Ogg Opus, Ogg FLAC, Ogg Speex, Ogg
Theora, Ogg Vorbis, True Audio, WavPack, OptimFROG, and AIFF audio files.
All versions of ID3v2 are supported, and all standard ID3v2.4 frames are
parsed. It can read Xing headers to accurately calculate the bitrate and
length of MP3s. ID3 and APEv2 tags can be edited regardless of audio
format. It can also manipulate Ogg streams on an individual packet/page
level.
Mutagen works on Python 2.6, 2.7, 3.3, 3.4 (CPython and PyPy) and has no
dependencies outside the Python standard library.
Installing
----------
$ ./setup.py build
$ su -c "./setup.py install"
Documentation
-------------
The primary documentation for Mutagen is the doc strings found in
the source code and the sphinx documentation in the docs/ directory.
To build the docs (needs sphinx):
$ ./setup.py build_sphinx
The tools/ directory contains several useful examples.
The docs are also hosted on readthedocs.org:
http://mutagen.readthedocs.org
Testing the Module
------------------
To test Mutagen's MP3 reading support, run
$ tools/mutagen-pony <your top-level MP3 directory here>
Mutagen will try to load all of them, and report any errors.
To look at the tags in files, run
$ tools/mutagen-inspect filename ...
To run our test suite,
$ ./setup.py test
Compatibility/Bugs
------------------
See docs/bugs.rst

41
lib/mutagen/__init__.py Normal file
View File

@@ -0,0 +1,41 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005 Michael Urman
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""Mutagen aims to be an all purpose multimedia tagging library.
::
import mutagen.[format]
metadata = mutagen.[format].Open(filename)
`metadata` acts like a dictionary of tags in the file. Tags are generally a
list of string-like values, but may have additional methods available
depending on tag or format. They may also be entirely different objects
for certain keys, again depending on format.
"""
from mutagen._util import MutagenError
from mutagen._file import FileType, StreamInfo, File
from mutagen._tags import Metadata
version = (1, 27)
"""Version tuple."""
version_string = ".".join(map(str, version))
"""Version string."""
MutagenError
FileType
StreamInfo
File
Metadata

84
lib/mutagen/_compat.py Normal file
View File

@@ -0,0 +1,84 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2013 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
import sys
PY2 = sys.version_info[0] == 2
PY3 = not PY2
if PY2:
from StringIO import StringIO
BytesIO = StringIO
from cStringIO import StringIO as cBytesIO
long_ = long
integer_types = (int, long)
string_types = (str, unicode)
text_type = unicode
xrange = xrange
cmp = cmp
chr_ = chr
def endswith(text, end):
return text.endswith(end)
iteritems = lambda d: d.iteritems()
itervalues = lambda d: d.itervalues()
iterkeys = lambda d: d.iterkeys()
iterbytes = lambda b: iter(b)
exec("def reraise(tp, value, tb):\n raise tp, value, tb")
def swap_to_string(cls):
if "__str__" in cls.__dict__:
cls.__unicode__ = cls.__str__
if "__bytes__" in cls.__dict__:
cls.__str__ = cls.__bytes__
return cls
elif PY3:
from io import StringIO
StringIO = StringIO
from io import BytesIO
cBytesIO = BytesIO
long_ = int
integer_types = (int,)
string_types = (str,)
text_type = str
xrange = range
cmp = lambda a, b: (a > b) - (a < b)
chr_ = lambda x: bytes([x])
def endswith(text, end):
# usefull for paths which can be both, str and bytes
if isinstance(text, str):
if not isinstance(end, str):
end = end.decode("ascii")
else:
if not isinstance(end, bytes):
end = end.encode("ascii")
return text.endswith(end)
iteritems = lambda d: iter(d.items())
itervalues = lambda d: iter(d.values())
iterkeys = lambda d: iter(d.keys())
iterbytes = lambda b: (bytes([v]) for v in b)
def reraise(tp, value, tb):
raise tp(value).with_traceback(tb)
def swap_to_string(cls):
return cls

199
lib/mutagen/_constants.py Normal file
View File

@@ -0,0 +1,199 @@
# -*- coding: utf-8 -*-
"""Constants used by Mutagen."""
GENRES = [
u"Blues",
u"Classic Rock",
u"Country",
u"Dance",
u"Disco",
u"Funk",
u"Grunge",
u"Hip-Hop",
u"Jazz",
u"Metal",
u"New Age",
u"Oldies",
u"Other",
u"Pop",
u"R&B",
u"Rap",
u"Reggae",
u"Rock",
u"Techno",
u"Industrial",
u"Alternative",
u"Ska",
u"Death Metal",
u"Pranks",
u"Soundtrack",
u"Euro-Techno",
u"Ambient",
u"Trip-Hop",
u"Vocal",
u"Jazz+Funk",
u"Fusion",
u"Trance",
u"Classical",
u"Instrumental",
u"Acid",
u"House",
u"Game",
u"Sound Clip",
u"Gospel",
u"Noise",
u"Alt. Rock",
u"Bass",
u"Soul",
u"Punk",
u"Space",
u"Meditative",
u"Instrumental Pop",
u"Instrumental Rock",
u"Ethnic",
u"Gothic",
u"Darkwave",
u"Techno-Industrial",
u"Electronic",
u"Pop-Folk",
u"Eurodance",
u"Dream",
u"Southern Rock",
u"Comedy",
u"Cult",
u"Gangsta Rap",
u"Top 40",
u"Christian Rap",
u"Pop/Funk",
u"Jungle",
u"Native American",
u"Cabaret",
u"New Wave",
u"Psychedelic",
u"Rave",
u"Showtunes",
u"Trailer",
u"Lo-Fi",
u"Tribal",
u"Acid Punk",
u"Acid Jazz",
u"Polka",
u"Retro",
u"Musical",
u"Rock & Roll",
u"Hard Rock",
u"Folk",
u"Folk-Rock",
u"National Folk",
u"Swing",
u"Fast-Fusion",
u"Bebop",
u"Latin",
u"Revival",
u"Celtic",
u"Bluegrass",
u"Avantgarde",
u"Gothic Rock",
u"Progressive Rock",
u"Psychedelic Rock",
u"Symphonic Rock",
u"Slow Rock",
u"Big Band",
u"Chorus",
u"Easy Listening",
u"Acoustic",
u"Humour",
u"Speech",
u"Chanson",
u"Opera",
u"Chamber Music",
u"Sonata",
u"Symphony",
u"Booty Bass",
u"Primus",
u"Porn Groove",
u"Satire",
u"Slow Jam",
u"Club",
u"Tango",
u"Samba",
u"Folklore",
u"Ballad",
u"Power Ballad",
u"Rhythmic Soul",
u"Freestyle",
u"Duet",
u"Punk Rock",
u"Drum Solo",
u"A Cappella",
u"Euro-House",
u"Dance Hall",
u"Goa",
u"Drum & Bass",
u"Club-House",
u"Hardcore",
u"Terror",
u"Indie",
u"BritPop",
u"Afro-Punk",
u"Polsk Punk",
u"Beat",
u"Christian Gangsta Rap",
u"Heavy Metal",
u"Black Metal",
u"Crossover",
u"Contemporary Christian",
u"Christian Rock",
u"Merengue",
u"Salsa",
u"Thrash Metal",
u"Anime",
u"JPop",
u"Synthpop",
u"Abstract",
u"Art Rock",
u"Baroque",
u"Bhangra",
u"Big Beat",
u"Breakbeat",
u"Chillout",
u"Downtempo",
u"Dub",
u"EBM",
u"Eclectic",
u"Electro",
u"Electroclash",
u"Emo",
u"Experimental",
u"Garage",
u"Global",
u"IDM",
u"Illbient",
u"Industro-Goth",
u"Jam Band",
u"Krautrock",
u"Leftfield",
u"Lounge",
u"Math Rock",
u"New Romantic",
u"Nu-Breakz",
u"Post-Punk",
u"Post-Rock",
u"Psytrance",
u"Shoegaze",
u"Space Rock",
u"Trop Rock",
u"World Music",
u"Neoclassical",
u"Audiobook",
u"Audio Theatre",
u"Neue Deutsche Welle",
u"Podcast",
u"Indie Rock",
u"G-Funk",
u"Dubstep",
u"Garage Rock",
u"Psybient",
]
"""The ID3v1 genre list."""

237
lib/mutagen/_file.py Normal file
View File

@@ -0,0 +1,237 @@
# Copyright (C) 2005 Michael Urman
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
import warnings
from mutagen._util import DictMixin
class FileType(DictMixin):
"""An abstract object wrapping tags and audio stream information.
Attributes:
* info -- stream information (length, bitrate, sample rate)
* tags -- metadata tags, if any
Each file format has different potential tags and stream
information.
FileTypes implement an interface very similar to Metadata; the
dict interface, save, load, and delete calls on a FileType call
the appropriate methods on its tag data.
"""
__module__ = "mutagen"
info = None
tags = None
filename = None
_mimes = ["application/octet-stream"]
def __init__(self, filename=None, *args, **kwargs):
if filename is None:
warnings.warn("FileType constructor requires a filename",
DeprecationWarning)
else:
self.load(filename, *args, **kwargs)
def load(self, filename, *args, **kwargs):
raise NotImplementedError
def __getitem__(self, key):
"""Look up a metadata tag key.
If the file has no tags at all, a KeyError is raised.
"""
if self.tags is None:
raise KeyError(key)
else:
return self.tags[key]
def __setitem__(self, key, value):
"""Set a metadata tag.
If the file has no tags, an appropriate format is added (but
not written until save is called).
"""
if self.tags is None:
self.add_tags()
self.tags[key] = value
def __delitem__(self, key):
"""Delete a metadata tag key.
If the file has no tags at all, a KeyError is raised.
"""
if self.tags is None:
raise KeyError(key)
else:
del(self.tags[key])
def keys(self):
"""Return a list of keys in the metadata tag.
If the file has no tags at all, an empty list is returned.
"""
if self.tags is None:
return []
else:
return self.tags.keys()
def delete(self, filename=None):
"""Remove tags from a file."""
if self.tags is not None:
if filename is None:
filename = self.filename
else:
warnings.warn(
"delete(filename=...) is deprecated, reload the file",
DeprecationWarning)
return self.tags.delete(filename)
def save(self, filename=None, **kwargs):
"""Save metadata tags."""
if filename is None:
filename = self.filename
else:
warnings.warn(
"save(filename=...) is deprecated, reload the file",
DeprecationWarning)
if self.tags is not None:
return self.tags.save(filename, **kwargs)
else:
raise ValueError("no tags in file")
def pprint(self):
"""Print stream information and comment key=value pairs."""
stream = "%s (%s)" % (self.info.pprint(), self.mime[0])
try:
tags = self.tags.pprint()
except AttributeError:
return stream
else:
return stream + ((tags and "\n" + tags) or "")
def add_tags(self):
"""Adds new tags to the file.
Raises if tags already exist.
"""
raise NotImplementedError
@property
def mime(self):
"""A list of mime types"""
mimes = []
for Kind in type(self).__mro__:
for mime in getattr(Kind, '_mimes', []):
if mime not in mimes:
mimes.append(mime)
return mimes
@staticmethod
def score(filename, fileobj, header):
raise NotImplementedError
class StreamInfo(object):
"""Abstract stream information object.
Provides attributes for length, bitrate, sample rate etc.
See the implementations for details.
"""
__module__ = "mutagen"
def pprint(self):
"""Print stream information"""
raise NotImplementedError
def File(filename, options=None, easy=False):
"""Guess the type of the file and try to open it.
The file type is decided by several things, such as the first 128
bytes (which usually contains a file type identifier), the
filename extension, and the presence of existing tags.
If no appropriate type could be found, None is returned.
:param options: Sequence of :class:`FileType` implementations, defaults to
all included ones.
:param easy: If the easy wrappers should be returnd if available.
For example :class:`EasyMP3 <mp3.EasyMP3>` instead
of :class:`MP3 <mp3.MP3>`.
"""
if options is None:
from mutagen.asf import ASF
from mutagen.apev2 import APEv2File
from mutagen.flac import FLAC
if easy:
from mutagen.easyid3 import EasyID3FileType as ID3FileType
else:
from mutagen.id3 import ID3FileType
if easy:
from mutagen.mp3 import EasyMP3 as MP3
else:
from mutagen.mp3 import MP3
from mutagen.oggflac import OggFLAC
from mutagen.oggspeex import OggSpeex
from mutagen.oggtheora import OggTheora
from mutagen.oggvorbis import OggVorbis
from mutagen.oggopus import OggOpus
if easy:
from mutagen.trueaudio import EasyTrueAudio as TrueAudio
else:
from mutagen.trueaudio import TrueAudio
from mutagen.wavpack import WavPack
if easy:
from mutagen.easymp4 import EasyMP4 as MP4
else:
from mutagen.mp4 import MP4
from mutagen.musepack import Musepack
from mutagen.monkeysaudio import MonkeysAudio
from mutagen.optimfrog import OptimFROG
from mutagen.aiff import AIFF
from mutagen.aac import AAC
options = [MP3, TrueAudio, OggTheora, OggSpeex, OggVorbis, OggFLAC,
FLAC, AIFF, APEv2File, MP4, ID3FileType, WavPack,
Musepack, MonkeysAudio, OptimFROG, ASF, OggOpus, AAC]
if not options:
return None
fileobj = open(filename, "rb")
try:
header = fileobj.read(128)
# Sort by name after score. Otherwise import order affects
# Kind sort order, which affects treatment of things with
# equals scores.
results = [(Kind.score(filename, fileobj, header), Kind.__name__)
for Kind in options]
finally:
fileobj.close()
results = list(zip(results, options))
results.sort()
(score, name), Kind = results[-1]
if score > 0:
return Kind(filename)
else:
return None

31
lib/mutagen/_tags.py Normal file
View File

@@ -0,0 +1,31 @@
# Copyright (C) 2005 Michael Urman
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
class Metadata(object):
"""An abstract dict-like object.
Metadata is the base class for many of the tag objects in Mutagen.
"""
__module__ = "mutagen"
def __init__(self, *args, **kwargs):
if args or kwargs:
self.load(*args, **kwargs)
def load(self, *args, **kwargs):
raise NotImplementedError
def save(self, filename=None):
"""Save changes to a file."""
raise NotImplementedError
def delete(self, filename=None):
"""Remove tags from a file."""
raise NotImplementedError

603
lib/mutagen/_util.py Normal file
View File

@@ -0,0 +1,603 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Utility classes for Mutagen.
You should not rely on the interfaces here being stable. They are
intended for internal use in Mutagen only.
"""
import struct
import codecs
from fnmatch import fnmatchcase
from ._compat import chr_, text_type, PY2, iteritems, iterbytes, \
integer_types, xrange
class MutagenError(Exception):
"""Base class for all custom exceptions in mutagen
.. versionadded:: 1.25
"""
def total_ordering(cls):
assert "__eq__" in cls.__dict__
assert "__lt__" in cls.__dict__
cls.__le__ = lambda self, other: self == other or self < other
cls.__gt__ = lambda self, other: not (self == other or self < other)
cls.__ge__ = lambda self, other: not self < other
cls.__ne__ = lambda self, other: not self.__eq__(other)
return cls
def hashable(cls):
"""Makes sure the class is hashable.
Needs a working __eq__ and __hash__ and will add a __ne__.
"""
# py2
assert "__hash__" in cls.__dict__
# py3
assert cls.__dict__["__hash__"] is not None
assert "__eq__" in cls.__dict__
cls.__ne__ = lambda self, other: not self.__eq__(other)
return cls
def enum(cls):
assert cls.__bases__ == (object,)
d = dict(cls.__dict__)
new_type = type(cls.__name__, (int,), d)
new_type.__module__ = cls.__module__
map_ = {}
for key, value in iteritems(d):
if key.upper() == key and isinstance(value, integer_types):
value_instance = new_type(value)
setattr(new_type, key, value_instance)
map_[value] = key
def repr_(self):
if self in map_:
return "%s.%s" % (type(self).__name__, map_[self])
else:
return "%s(%s)" % (type(self).__name__, self)
setattr(new_type, "__repr__", repr_)
return new_type
@total_ordering
class DictMixin(object):
"""Implement the dict API using keys() and __*item__ methods.
Similar to UserDict.DictMixin, this takes a class that defines
__getitem__, __setitem__, __delitem__, and keys(), and turns it
into a full dict-like object.
UserDict.DictMixin is not suitable for this purpose because it's
an old-style class.
This class is not optimized for very large dictionaries; many
functions have linear memory requirements. I recommend you
override some of these functions if speed is required.
"""
def __iter__(self):
return iter(self.keys())
def __has_key(self, key):
try:
self[key]
except KeyError:
return False
else:
return True
if PY2:
has_key = __has_key
__contains__ = __has_key
if PY2:
iterkeys = lambda self: iter(self.keys())
def values(self):
return [self[k] for k in self.keys()]
if PY2:
itervalues = lambda self: iter(self.values())
def items(self):
return list(zip(self.keys(), self.values()))
if PY2:
iteritems = lambda s: iter(s.items())
def clear(self):
for key in list(self.keys()):
self.__delitem__(key)
def pop(self, key, *args):
if len(args) > 1:
raise TypeError("pop takes at most two arguments")
try:
value = self[key]
except KeyError:
if args:
return args[0]
else:
raise
del(self[key])
return value
def popitem(self):
for key in self.keys():
break
else:
raise KeyError("dictionary is empty")
return key, self.pop(key)
def update(self, other=None, **kwargs):
if other is None:
self.update(kwargs)
other = {}
try:
for key, value in other.items():
self.__setitem__(key, value)
except AttributeError:
for key, value in other:
self[key] = value
def setdefault(self, key, default=None):
try:
return self[key]
except KeyError:
self[key] = default
return default
def get(self, key, default=None):
try:
return self[key]
except KeyError:
return default
def __repr__(self):
return repr(dict(self.items()))
def __eq__(self, other):
return dict(self.items()) == other
def __lt__(self, other):
return dict(self.items()) < other
__hash__ = object.__hash__
def __len__(self):
return len(self.keys())
class DictProxy(DictMixin):
def __init__(self, *args, **kwargs):
self.__dict = {}
super(DictProxy, self).__init__(*args, **kwargs)
def __getitem__(self, key):
return self.__dict[key]
def __setitem__(self, key, value):
self.__dict[key] = value
def __delitem__(self, key):
del(self.__dict[key])
def keys(self):
return self.__dict.keys()
def _fill_cdata(cls):
"""Add struct pack/unpack functions"""
funcs = {}
for key, name in [("b", "char"), ("h", "short"),
("i", "int"), ("q", "longlong")]:
for echar, esuffix in [("<", "le"), (">", "be")]:
esuffix = "_" + esuffix
for unsigned in [True, False]:
s = struct.Struct(echar + (key.upper() if unsigned else key))
get_wrapper = lambda f: lambda *a, **k: f(*a, **k)[0]
unpack = get_wrapper(s.unpack)
unpack_from = get_wrapper(s.unpack_from)
def get_unpack_from(s):
def unpack_from(data, offset=0):
return s.unpack_from(data, offset)[0], offset + s.size
return unpack_from
unpack_from = get_unpack_from(s)
pack = s.pack
prefix = "u" if unsigned else ""
if s.size == 1:
esuffix = ""
bits = str(s.size * 8)
funcs["%s%s%s" % (prefix, name, esuffix)] = unpack
funcs["%sint%s%s" % (prefix, bits, esuffix)] = unpack
funcs["%s%s%s_from" % (prefix, name, esuffix)] = unpack_from
funcs["%sint%s%s_from" % (prefix, bits, esuffix)] = unpack_from
funcs["to_%s%s%s" % (prefix, name, esuffix)] = pack
funcs["to_%sint%s%s" % (prefix, bits, esuffix)] = pack
for key, func in iteritems(funcs):
setattr(cls, key, staticmethod(func))
class cdata(object):
"""C character buffer to Python numeric type conversions.
For each size/sign/endianness:
uint32_le(data)/to_uint32_le(num)/uint32_le_from(data, offset=0)
"""
from struct import error
error = error
bitswap = b''.join(
chr_(sum(((val >> i) & 1) << (7 - i) for i in range(8)))
for val in range(256))
test_bit = staticmethod(lambda value, n: bool((value >> n) & 1))
_fill_cdata(cdata)
def lock(fileobj):
"""Lock a file object 'safely'.
That means a failure to lock because the platform doesn't
support fcntl or filesystem locks is not considered a
failure. This call does block.
Returns whether or not the lock was successful, or
raises an exception in more extreme circumstances (full
lock table, invalid file).
"""
try:
import fcntl
except ImportError:
return False
else:
try:
fcntl.lockf(fileobj, fcntl.LOCK_EX)
except IOError:
# FIXME: There's possibly a lot of complicated
# logic that needs to go here in case the IOError
# is EACCES or EAGAIN.
return False
else:
return True
def unlock(fileobj):
"""Unlock a file object.
Don't call this on a file object unless a call to lock()
returned true.
"""
# If this fails there's a mismatched lock/unlock pair,
# so we definitely don't want to ignore errors.
import fcntl
fcntl.lockf(fileobj, fcntl.LOCK_UN)
def insert_bytes(fobj, size, offset, BUFFER_SIZE=2 ** 16):
"""Insert size bytes of empty space starting at offset.
fobj must be an open file object, open rb+ or
equivalent. Mutagen tries to use mmap to resize the file, but
falls back to a significantly slower method if mmap fails.
"""
assert 0 < size
assert 0 <= offset
locked = False
fobj.seek(0, 2)
filesize = fobj.tell()
movesize = filesize - offset
fobj.write(b'\x00' * size)
fobj.flush()
try:
try:
import mmap
file_map = mmap.mmap(fobj.fileno(), filesize + size)
try:
file_map.move(offset + size, offset, movesize)
finally:
file_map.close()
except (ValueError, EnvironmentError, ImportError):
# handle broken mmap scenarios
locked = lock(fobj)
fobj.truncate(filesize)
fobj.seek(0, 2)
padsize = size
# Don't generate an enormous string if we need to pad
# the file out several megs.
while padsize:
addsize = min(BUFFER_SIZE, padsize)
fobj.write(b"\x00" * addsize)
padsize -= addsize
fobj.seek(filesize, 0)
while movesize:
# At the start of this loop, fobj is pointing at the end
# of the data we need to move, which is of movesize length.
thismove = min(BUFFER_SIZE, movesize)
# Seek back however much we're going to read this frame.
fobj.seek(-thismove, 1)
nextpos = fobj.tell()
# Read it, so we're back at the end.
data = fobj.read(thismove)
# Seek back to where we need to write it.
fobj.seek(-thismove + size, 1)
# Write it.
fobj.write(data)
# And seek back to the end of the unmoved data.
fobj.seek(nextpos)
movesize -= thismove
fobj.flush()
finally:
if locked:
unlock(fobj)
def delete_bytes(fobj, size, offset, BUFFER_SIZE=2 ** 16):
"""Delete size bytes of empty space starting at offset.
fobj must be an open file object, open rb+ or
equivalent. Mutagen tries to use mmap to resize the file, but
falls back to a significantly slower method if mmap fails.
"""
locked = False
assert 0 < size
assert 0 <= offset
fobj.seek(0, 2)
filesize = fobj.tell()
movesize = filesize - offset - size
assert 0 <= movesize
try:
if movesize > 0:
fobj.flush()
try:
import mmap
file_map = mmap.mmap(fobj.fileno(), filesize)
try:
file_map.move(offset, offset + size, movesize)
finally:
file_map.close()
except (ValueError, EnvironmentError, ImportError):
# handle broken mmap scenarios
locked = lock(fobj)
fobj.seek(offset + size)
buf = fobj.read(BUFFER_SIZE)
while buf:
fobj.seek(offset)
fobj.write(buf)
offset += len(buf)
fobj.seek(offset + size)
buf = fobj.read(BUFFER_SIZE)
fobj.truncate(filesize - size)
fobj.flush()
finally:
if locked:
unlock(fobj)
def dict_match(d, key, default=None):
"""Like __getitem__ but works as if the keys() are all filename patterns.
Returns the value of any dict key that matches the passed key.
"""
if key in d and "[" not in key:
return d[key]
else:
for pattern, value in iteritems(d):
if fnmatchcase(key, pattern):
return value
return default
def decode_terminated(data, encoding, strict=True):
"""Returns the decoded data until the first NULL terminator
and all data after it.
In case the data can't be decoded raises UnicodeError.
In case the encoding is not found raises LookupError.
In case the data isn't null terminated (even if it is encoded correctly)
raises ValueError except if strict is False, then the decoded string
will be returned anyway.
"""
codec_info = codecs.lookup(encoding)
# normalize encoding name so we can compare by name
encoding = codec_info.name
# fast path
if encoding in ("utf-8", "iso8859-1"):
index = data.find(b"\x00")
if index == -1:
# make sure we raise UnicodeError first, like in the slow path
res = data.decode(encoding), b""
if strict:
raise ValueError("not null terminated")
else:
return res
return data[:index].decode(encoding), data[index + 1:]
# slow path
decoder = codec_info.incrementaldecoder()
r = []
for i, b in enumerate(iterbytes(data)):
c = decoder.decode(b)
if c == u"\x00":
return u"".join(r), data[i + 1:]
r.append(c)
else:
# make sure the decoder is finished
r.append(decoder.decode(b"", True))
if strict:
raise ValueError("not null terminated")
return u"".join(r), b""
def split_escape(string, sep, maxsplit=None, escape_char="\\"):
"""Like unicode/str/bytes.split but allows for the separator to be escaped
If passed unicode/str/bytes will only return list of unicode/str/bytes.
"""
assert len(sep) == 1
assert len(escape_char) == 1
if isinstance(string, bytes):
if isinstance(escape_char, text_type):
escape_char = escape_char.encode("ascii")
iter_ = iterbytes
else:
iter_ = iter
if maxsplit is None:
maxsplit = len(string)
empty = string[:0]
result = []
current = empty
escaped = False
for char in iter_(string):
if escaped:
if char != escape_char and char != sep:
current += escape_char
current += char
escaped = False
else:
if char == escape_char:
escaped = True
elif char == sep and len(result) < maxsplit:
result.append(current)
current = empty
else:
current += char
result.append(current)
return result
class BitReaderError(Exception):
pass
class BitReader(object):
def __init__(self, fileobj):
self._fileobj = fileobj
self._buffer = 0
self._bits = 0
self._pos = fileobj.tell()
def bits(self, count):
"""Reads `count` bits and returns an uint, MSB read first.
May raise BitReaderError if not enough data could be read or
IOError by the underlying file object.
"""
if count < 0:
raise ValueError
if count > self._bits:
n_bytes = (count - self._bits + 7) // 8
data = self._fileobj.read(n_bytes)
if len(data) != n_bytes:
raise BitReaderError("not enough data")
for b in bytearray(data):
self._buffer = (self._buffer << 8) | b
self._bits += n_bytes * 8
self._bits -= count
value = self._buffer >> self._bits
self._buffer &= (1 << self._bits) - 1
assert self._bits < 8
return value
def bytes(self, count):
"""Returns a bytearray of length `count`. Works unaligned."""
if count < 0:
raise ValueError
# fast path
if self._bits == 0:
data = self._fileobj.read(count)
if len(data) != count:
raise BitReaderError("not enough data")
return data
return bytes(bytearray(self.bits(8) for _ in xrange(count)))
def skip(self, count):
"""Skip `count` bits.
Might raise BitReaderError if there wasn't enough data to skip,
but might also fail on the next bits() instead.
"""
if count < 0:
raise ValueError
if count <= self._bits:
self.bits(count)
else:
count -= self.align()
n_bytes = count // 8
self._fileobj.seek(n_bytes, 1)
count -= n_bytes * 8
self.bits(count)
def get_position(self):
"""Returns the amount of bits read or skipped so far"""
return (self._fileobj.tell() - self._pos) * 8 - self._bits
def align(self):
"""Align to the next byte, returns the amount of bits skipped"""
bits = self._bits
self._buffer = 0
self._bits = 0
return bits
def is_aligned(self):
"""If we are currently aligned to bytes and nothing is buffered"""
return self._bits == 0

327
lib/mutagen/_vorbis.py Normal file
View File

@@ -0,0 +1,327 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005-2006 Joe Wreschnig
# 2013 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""Read and write Vorbis comment data.
Vorbis comments are freeform key/value pairs; keys are
case-insensitive ASCII and values are Unicode strings. A key may have
multiple values.
The specification is at http://www.xiph.org/vorbis/doc/v-comment.html.
"""
import sys
import mutagen
from ._compat import reraise, BytesIO, text_type, xrange, PY3, PY2
from mutagen._util import DictMixin, cdata
def is_valid_key(key):
"""Return true if a string is a valid Vorbis comment key.
Valid Vorbis comment keys are printable ASCII between 0x20 (space)
and 0x7D ('}'), excluding '='.
Takes str/unicode in Python 2, unicode in Python 3
"""
if PY3 and isinstance(key, bytes):
raise TypeError("needs to be str not bytes")
for c in key:
if c < " " or c > "}" or c == "=":
return False
else:
return bool(key)
istag = is_valid_key
class error(IOError):
pass
class VorbisUnsetFrameError(error):
pass
class VorbisEncodingError(error):
pass
class VComment(mutagen.Metadata, list):
"""A Vorbis comment parser, accessor, and renderer.
All comment ordering is preserved. A VComment is a list of
key/value pairs, and so any Python list method can be used on it.
Vorbis comments are always wrapped in something like an Ogg Vorbis
bitstream or a FLAC metadata block, so this loads string data or a
file-like object, not a filename.
Attributes:
* vendor -- the stream 'vendor' (i.e. writer); default 'Mutagen'
"""
vendor = u"Mutagen " + mutagen.version_string
def __init__(self, data=None, *args, **kwargs):
# Collect the args to pass to load, this lets child classes
# override just load and get equivalent magic for the
# constructor.
if data is not None:
if isinstance(data, bytes):
data = BytesIO(data)
elif not hasattr(data, 'read'):
raise TypeError("VComment requires bytes or a file-like")
self.load(data, *args, **kwargs)
def load(self, fileobj, errors='replace', framing=True):
"""Parse a Vorbis comment from a file-like object.
Keyword arguments:
* errors:
'strict', 'replace', or 'ignore'. This affects Unicode decoding
and how other malformed content is interpreted.
* framing -- if true, fail if a framing bit is not present
Framing bits are required by the Vorbis comment specification,
but are not used in FLAC Vorbis comment blocks.
"""
try:
vendor_length = cdata.uint_le(fileobj.read(4))
self.vendor = fileobj.read(vendor_length).decode('utf-8', errors)
count = cdata.uint_le(fileobj.read(4))
for i in xrange(count):
length = cdata.uint_le(fileobj.read(4))
try:
string = fileobj.read(length).decode('utf-8', errors)
except (OverflowError, MemoryError):
raise error("cannot read %d bytes, too large" % length)
try:
tag, value = string.split('=', 1)
except ValueError as err:
if errors == "ignore":
continue
elif errors == "replace":
tag, value = u"unknown%d" % i, string
else:
reraise(VorbisEncodingError, err, sys.exc_info()[2])
try:
tag = tag.encode('ascii', errors)
except UnicodeEncodeError:
raise VorbisEncodingError("invalid tag name %r" % tag)
else:
# string keys in py3k
if PY3:
tag = tag.decode("ascii")
if is_valid_key(tag):
self.append((tag, value))
if framing and not bytearray(fileobj.read(1))[0] & 0x01:
raise VorbisUnsetFrameError("framing bit was unset")
except (cdata.error, TypeError):
raise error("file is not a valid Vorbis comment")
def validate(self):
"""Validate keys and values.
Check to make sure every key used is a valid Vorbis key, and
that every value used is a valid Unicode or UTF-8 string. If
any invalid keys or values are found, a ValueError is raised.
In Python 3 all keys and values have to be a string.
"""
if not isinstance(self.vendor, text_type):
if PY3:
raise ValueError("vendor needs to be str")
try:
self.vendor.decode('utf-8')
except UnicodeDecodeError:
raise ValueError
for key, value in self:
try:
if not is_valid_key(key):
raise ValueError
except TypeError:
raise ValueError("%r is not a valid key" % key)
if not isinstance(value, text_type):
if PY3:
raise ValueError("%r needs to be str" % key)
try:
value.decode("utf-8")
except:
raise ValueError("%r is not a valid value" % value)
return True
def clear(self):
"""Clear all keys from the comment."""
for i in list(self):
self.remove(i)
def write(self, framing=True):
"""Return a string representation of the data.
Validation is always performed, so calling this function on
invalid data may raise a ValueError.
Keyword arguments:
* framing -- if true, append a framing bit (see load)
"""
self.validate()
def _encode(value):
if not isinstance(value, bytes):
return value.encode('utf-8')
return value
f = BytesIO()
vendor = _encode(self.vendor)
f.write(cdata.to_uint_le(len(vendor)))
f.write(vendor)
f.write(cdata.to_uint_le(len(self)))
for tag, value in self:
tag = _encode(tag)
value = _encode(value)
comment = tag + b"=" + value
f.write(cdata.to_uint_le(len(comment)))
f.write(comment)
if framing:
f.write(b"\x01")
return f.getvalue()
def pprint(self):
def _decode(value):
if not isinstance(value, text_type):
return value.decode('utf-8', 'replace')
return value
tags = [u"%s=%s" % (_decode(k), _decode(v)) for k, v in self]
return u"\n".join(tags)
class VCommentDict(VComment, DictMixin):
"""A VComment that looks like a dictionary.
This object differs from a dictionary in two ways. First,
len(comment) will still return the number of values, not the
number of keys. Secondly, iterating through the object will
iterate over (key, value) pairs, not keys. Since a key may have
multiple values, the same value may appear multiple times while
iterating.
Since Vorbis comment keys are case-insensitive, all keys are
normalized to lowercase ASCII.
"""
def __getitem__(self, key):
"""A list of values for the key.
This is a copy, so comment['title'].append('a title') will not
work.
"""
# PY3 only
if isinstance(key, slice):
return VComment.__getitem__(self, key)
if not is_valid_key(key):
raise ValueError
key = key.lower()
values = [value for (k, value) in self if k.lower() == key]
if not values:
raise KeyError(key)
else:
return values
def __delitem__(self, key):
"""Delete all values associated with the key."""
# PY3 only
if isinstance(key, slice):
return VComment.__delitem__(self, key)
if not is_valid_key(key):
raise ValueError
key = key.lower()
to_delete = [x for x in self if x[0].lower() == key]
if not to_delete:
raise KeyError(key)
else:
for item in to_delete:
self.remove(item)
def __contains__(self, key):
"""Return true if the key has any values."""
if not is_valid_key(key):
raise ValueError
key = key.lower()
for k, value in self:
if k.lower() == key:
return True
else:
return False
def __setitem__(self, key, values):
"""Set a key's value or values.
Setting a value overwrites all old ones. The value may be a
list of Unicode or UTF-8 strings, or a single Unicode or UTF-8
string.
"""
# PY3 only
if isinstance(key, slice):
return VComment.__setitem__(self, key, values)
if not is_valid_key(key):
raise ValueError
if not isinstance(values, list):
values = [values]
try:
del(self[key])
except KeyError:
pass
if PY2:
key = key.encode('ascii')
for value in values:
self.append((key, value))
def keys(self):
"""Return all keys in the comment."""
return list(set([k.lower() for k, v in self]))
def as_dict(self):
"""Return a copy of the comment data in a real dict."""
return dict([(key, self[key]) for key in self.keys()])

407
lib/mutagen/aac.py Normal file
View File

@@ -0,0 +1,407 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2014 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""
* ADTS - Audio Data Transport Stream
* ADIF - Audio Data Interchange Format
* See ISO/IEC 13818-7 / 14496-03
"""
from mutagen import StreamInfo
from mutagen._file import FileType
from mutagen._util import BitReader, BitReaderError, MutagenError
from mutagen._compat import endswith, xrange
_FREQS = [
96000, 88200, 64000, 48000,
44100, 32000, 24000, 22050,
16000, 12000, 11025, 8000,
7350,
]
class _ADTSStream(object):
"""Represents a series of frames belonging to the same stream"""
parsed_frames = 0
"""Number of successfully parsed frames"""
offset = 0
"""offset in bytes at which the stream starts (the first sync word)"""
@classmethod
def find_stream(cls, fileobj, max_bytes):
"""Returns a possibly valid _ADTSStream or None.
Args:
max_bytes (int): maximum bytes to read
"""
r = BitReader(fileobj)
stream = cls(r)
if stream.sync(max_bytes):
stream.offset = (r.get_position() - 12) // 8
return stream
def sync(self, max_bytes):
"""Find the next sync.
Returns True if found."""
# at least 2 bytes for the sync
max_bytes = max(max_bytes, 2)
r = self._r
r.align()
while max_bytes > 0:
try:
b = r.bytes(1)
if b == b"\xff":
if r.bits(4) == 0xf:
return True
r.align()
max_bytes -= 2
else:
max_bytes -= 1
except BitReaderError:
return False
return False
def __init__(self, r):
"""Use _ADTSStream.find_stream to create a stream"""
self._fixed_header_key = None
self._r = r
self.offset = -1
self.parsed_frames = 0
self._samples = 0
self._payload = 0
self._start = r.get_position() / 8
self._last = self._start
@property
def bitrate(self):
"""Bitrate of the raw aac blocks, excluding framing/crc"""
assert self.parsed_frames, "no frame parsed yet"
if self._samples == 0:
return 0
return (8 * self._payload * self.frequency) // self._samples
@property
def samples(self):
"""samples so far"""
assert self.parsed_frames, "no frame parsed yet"
return self._samples
@property
def size(self):
"""bytes read in the stream so far (including framing)"""
assert self.parsed_frames, "no frame parsed yet"
return self._last - self._start
@property
def channels(self):
"""0 means unknown"""
assert self.parsed_frames, "no frame parsed yet"
b_index = self._fixed_header_key[6]
if b_index == 7:
return 8
elif b_index > 7:
return 0
else:
return b_index
@property
def frequency(self):
"""0 means unknown"""
assert self.parsed_frames, "no frame parsed yet"
f_index = self._fixed_header_key[4]
try:
return _FREQS[f_index]
except IndexError:
return 0
def parse_frame(self):
"""True if parsing was successful.
Fails either because the frame wasn't valid or the stream ended.
"""
try:
return self._parse_frame()
except BitReaderError:
return False
def _parse_frame(self):
r = self._r
# start == position of sync word
start = r.get_position() - 12
# adts_fixed_header
id_ = r.bits(1)
layer = r.bits(2)
protection_absent = r.bits(1)
profile = r.bits(2)
sampling_frequency_index = r.bits(4)
private_bit = r.bits(1)
# TODO: if 0 we could parse program_config_element()
channel_configuration = r.bits(3)
original_copy = r.bits(1)
home = r.bits(1)
# the fixed header has to be the same for every frame in the stream
fixed_header_key = (
id_, layer, protection_absent, profile, sampling_frequency_index,
private_bit, channel_configuration, original_copy, home,
)
if self._fixed_header_key is None:
self._fixed_header_key = fixed_header_key
else:
if self._fixed_header_key != fixed_header_key:
return False
# adts_variable_header
r.skip(2) # copyright_identification_bit/start
frame_length = r.bits(13)
r.skip(11) # adts_buffer_fullness
nordbif = r.bits(2)
# adts_variable_header end
crc_overhead = 0
if not protection_absent:
crc_overhead += (nordbif + 1) * 16
if nordbif != 0:
crc_overhead *= 2
left = (frame_length * 8) - (r.get_position() - start)
if left < 0:
return False
r.skip(left)
assert r.is_aligned()
self._payload += (left - crc_overhead) / 8
self._samples += (nordbif + 1) * 1024
self._last = r.get_position() / 8
self.parsed_frames += 1
return True
class ProgramConfigElement(object):
element_instance_tag = None
object_type = None
sampling_frequency_index = None
channels = None
def __init__(self, r):
"""Reads the program_config_element()
Raises BitReaderError
"""
self.element_instance_tag = r.bits(4)
self.object_type = r.bits(2)
self.sampling_frequency_index = r.bits(4)
num_front_channel_elements = r.bits(4)
num_side_channel_elements = r.bits(4)
num_back_channel_elements = r.bits(4)
num_lfe_channel_elements = r.bits(2)
num_assoc_data_elements = r.bits(3)
num_valid_cc_elements = r.bits(4)
mono_mixdown_present = r.bits(1)
if mono_mixdown_present == 1:
r.skip(4)
stereo_mixdown_present = r.bits(1)
if stereo_mixdown_present == 1:
r.skip(4)
matrix_mixdown_idx_present = r.bits(1)
if matrix_mixdown_idx_present == 1:
r.skip(3)
elms = num_front_channel_elements + num_side_channel_elements + \
num_back_channel_elements
channels = 0
for i in xrange(elms):
channels += 1
element_is_cpe = r.bits(1)
if element_is_cpe:
channels += 1
r.skip(4)
channels += num_lfe_channel_elements
self.channels = channels
r.skip(4 * num_lfe_channel_elements)
r.skip(4 * num_assoc_data_elements)
r.skip(5 * num_valid_cc_elements)
r.align()
comment_field_bytes = r.bits(8)
r.skip(8 * comment_field_bytes)
class AACError(MutagenError):
pass
class AACInfo(StreamInfo):
"""AAC stream information.
Attributes:
* channels -- number of audio channels
* length -- file length in seconds, as a float
* sample_rate -- audio sampling rate in Hz
* bitrate -- audio bitrate, in bits per second
The length of the stream is just a guess and might not be correct.
"""
channels = 0
length = 0
sample_rate = 0
bitrate = 0
def __init__(self, fileobj):
# skip id3v2 header
start_offset = 0
header = fileobj.read(10)
from mutagen.id3 import BitPaddedInt
if header.startswith(b"ID3"):
size = BitPaddedInt(header[6:])
start_offset = size + 10
fileobj.seek(start_offset)
adif = fileobj.read(4)
if adif == b"ADIF":
self._parse_adif(fileobj)
self._type = "ADIF"
else:
self._parse_adts(fileobj, start_offset)
self._type = "ADTS"
def _parse_adif(self, fileobj):
r = BitReader(fileobj)
try:
copyright_id_present = r.bits(1)
if copyright_id_present:
r.skip(72) # copyright_id
r.skip(1 + 1) # original_copy, home
bitstream_type = r.bits(1)
self.bitrate = r.bits(23)
npce = r.bits(4)
if bitstream_type == 0:
r.skip(20) # adif_buffer_fullness
pce = ProgramConfigElement(r)
try:
self.sample_rate = _FREQS[pce.sampling_frequency_index]
except IndexError:
pass
self.channels = pce.channels
# other pces..
for i in xrange(npce):
ProgramConfigElement(r)
r.align()
except BitReaderError as e:
raise AACError(e)
# use bitrate + data size to guess length
start = fileobj.tell()
fileobj.seek(0, 2)
length = fileobj.tell() - start
if self.bitrate != 0:
self.length = (8.0 * length) / self.bitrate
def _parse_adts(self, fileobj, start_offset):
max_initial_read = 512
max_resync_read = 10
max_sync_tries = 10
frames_max = 100
frames_needed = 3
# Try up to X times to find a sync word and read up to Y frames.
# If more than Z frames are valid we assume a valid stream
offset = start_offset
for i in xrange(max_sync_tries):
fileobj.seek(offset)
s = _ADTSStream.find_stream(fileobj, max_initial_read)
if s is None:
raise AACError("sync not found")
# start right after the last found offset
offset += s.offset + 1
for i in xrange(frames_max):
if not s.parse_frame():
break
if not s.sync(max_resync_read):
break
if s.parsed_frames >= frames_needed:
break
else:
raise AACError(
"no valid stream found (only %d frames)" % s.parsed_frames)
self.sample_rate = s.frequency
self.channels = s.channels
self.bitrate = s.bitrate
# size from stream start to end of file
fileobj.seek(0, 2)
stream_size = fileobj.tell() - (offset + s.offset)
# approx
self.length = float(s.samples * stream_size) / (s.size * s.frequency)
def pprint(self):
return "AAC (%s), %d Hz, %.2f seconds, %d channel(s), %d bps" % (
self._type, self.sample_rate, self.length, self.channels,
self.bitrate)
class AAC(FileType):
"""Load ADTS or ADIF streams containing AAC.
Tagging is not supported.
Use the ID3/APEv2 classes directly instead.
"""
_mimes = ["audio/x-aac"]
def load(self, filename):
self.filename = filename
with open(filename, "rb") as h:
self.info = AACInfo(h)
@staticmethod
def score(filename, fileobj, header):
filename = filename.lower()
s = endswith(filename, ".aac") or endswith(filename, ".adts") or \
endswith(filename, ".adif")
s += b"ADIF" in header
return s
Open = AAC
error = AACError
__all__ = ["AAC", "Open"]

362
lib/mutagen/aiff.py Normal file
View File

@@ -0,0 +1,362 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2014 Evan Purkhiser
# 2014 Ben Ockmore
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""AIFF audio stream information and tags."""
# NOTE from Ben Ockmore - according to the Py3k migration guidelines, AIFF
# chunk keys should be unicode in Py3k, and unicode or bytes in Py2k (ASCII).
# To make this easier, chunk keys should be stored internally as unicode.
import struct
from struct import pack
from ._compat import endswith, text_type, PY3
from mutagen import StreamInfo, FileType
from mutagen.id3 import ID3
from mutagen.id3._util import error as ID3Error
from mutagen._util import insert_bytes, delete_bytes, MutagenError
__all__ = ["AIFF", "Open", "delete"]
class error(MutagenError, RuntimeError):
pass
class InvalidChunk(error, IOError):
pass
# based on stdlib's aifc
_HUGE_VAL = 1.79769313486231e+308
def is_valid_chunk_id(id):
if not isinstance(id, text_type):
if PY3:
raise TypeError("AIFF chunk must be unicode")
try:
id = id.decode('ascii')
except UnicodeDecodeError:
return False
return ((len(id) <= 4) and (min(id) >= u' ') and
(max(id) <= u'~'))
def read_float(data): # 10 bytes
expon, himant, lomant = struct.unpack('>hLL', data)
sign = 1
if expon < 0:
sign = -1
expon = expon + 0x8000
if expon == himant == lomant == 0:
f = 0.0
elif expon == 0x7FFF:
f = _HUGE_VAL
else:
expon = expon - 16383
f = (himant * 0x100000000 + lomant) * pow(2.0, expon - 63)
return sign * f
class IFFChunk(object):
"""Representation of a single IFF chunk"""
# Chunk headers are 8 bytes long (4 for ID and 4 for the size)
HEADER_SIZE = 8
def __init__(self, fileobj, parent_chunk=None):
self.__fileobj = fileobj
self.parent_chunk = parent_chunk
self.offset = fileobj.tell()
header = fileobj.read(self.HEADER_SIZE)
if len(header) < self.HEADER_SIZE:
raise InvalidChunk()
self.id, self.data_size = struct.unpack('>4si', header)
if not isinstance(self.id, text_type):
self.id = self.id.decode('ascii')
if not is_valid_chunk_id(self.id):
raise InvalidChunk()
self.size = self.HEADER_SIZE + self.data_size
self.data_offset = fileobj.tell()
self.data = None
def read(self):
"""Read the chunks data"""
self.__fileobj.seek(self.data_offset)
self.data = self.__fileobj.read(self.data_size)
def delete(self):
"""Removes the chunk from the file"""
delete_bytes(self.__fileobj, self.size, self.offset)
if self.parent_chunk is not None:
self.parent_chunk.resize(self.parent_chunk.data_size - self.size)
def resize(self, data_size):
"""Update the size of the chunk"""
self.__fileobj.seek(self.offset + 4)
self.__fileobj.write(pack('>I', data_size))
if self.parent_chunk is not None:
size_diff = self.data_size - data_size
self.parent_chunk.resize(self.parent_chunk.data_size - size_diff)
self.data_size = data_size
self.size = data_size + self.HEADER_SIZE
class IFFFile(object):
"""Representation of a IFF file"""
def __init__(self, fileobj):
self.__fileobj = fileobj
self.__chunks = {}
# AIFF Files always start with the FORM chunk which contains a 4 byte
# ID before the start of other chunks
fileobj.seek(0)
self.__chunks[u'FORM'] = IFFChunk(fileobj)
# Skip past the 4 byte FORM id
fileobj.seek(IFFChunk.HEADER_SIZE + 4)
# Where the next chunk can be located. We need to keep track of this
# since the size indicated in the FORM header may not match up with the
# offset determined from the size of the last chunk in the file
self.__next_offset = fileobj.tell()
# Load all of the chunks
while True:
try:
chunk = IFFChunk(fileobj, self[u'FORM'])
except InvalidChunk:
break
self.__chunks[chunk.id.strip()] = chunk
# Calculate the location of the next chunk,
# considering the pad byte
self.__next_offset = chunk.offset + chunk.size
self.__next_offset += self.__next_offset % 2
fileobj.seek(self.__next_offset)
def __contains__(self, id_):
"""Check if the IFF file contains a specific chunk"""
if not isinstance(id_, text_type):
id_ = id_.decode('ascii')
if not is_valid_chunk_id(id_):
raise KeyError("AIFF key must be four ASCII characters.")
return id_ in self.__chunks
def __getitem__(self, id_):
"""Get a chunk from the IFF file"""
if not isinstance(id_, text_type):
id_ = id_.decode('ascii')
if not is_valid_chunk_id(id_):
raise KeyError("AIFF key must be four ASCII characters.")
try:
return self.__chunks[id_]
except KeyError:
raise KeyError(
"%r has no %r chunk" % (self.__fileobj.name, id_))
def __delitem__(self, id_):
"""Remove a chunk from the IFF file"""
if not isinstance(id_, text_type):
id_ = id_.decode('ascii')
if not is_valid_chunk_id(id_):
raise KeyError("AIFF key must be four ASCII characters.")
self.__chunks.pop(id_).delete()
def insert_chunk(self, id_):
"""Insert a new chunk at the end of the IFF file"""
if not isinstance(id_, text_type):
id_ = id_.decode('ascii')
if not is_valid_chunk_id(id_):
raise KeyError("AIFF key must be four ASCII characters.")
self.__fileobj.seek(self.__next_offset)
self.__fileobj.write(pack('>4si', id_.ljust(4).encode('ascii'), 0))
self.__fileobj.seek(self.__next_offset)
chunk = IFFChunk(self.__fileobj, self[u'FORM'])
self[u'FORM'].resize(self[u'FORM'].data_size + chunk.size)
self.__chunks[id_] = chunk
self.__next_offset = chunk.offset + chunk.size
class AIFFInfo(StreamInfo):
"""AIFF audio stream information.
Information is parsed from the COMM chunk of the AIFF file
Useful attributes:
* length -- audio length, in seconds
* bitrate -- audio bitrate, in bits per second
* channels -- The number of audio channels
* sample_rate -- audio sample rate, in Hz
* sample_size -- The audio sample size
"""
length = 0
bitrate = 0
channels = 0
sample_rate = 0
def __init__(self, fileobj):
iff = IFFFile(fileobj)
try:
common_chunk = iff[u'COMM']
except KeyError as e:
raise error(str(e))
common_chunk.read()
info = struct.unpack('>hLh10s', common_chunk.data[:18])
channels, frame_count, sample_size, sample_rate = info
self.sample_rate = int(read_float(sample_rate))
self.sample_size = sample_size
self.channels = channels
self.bitrate = channels * sample_size * self.sample_rate
self.length = frame_count / float(self.sample_rate)
def pprint(self):
return "%d channel AIFF @ %d bps, %s Hz, %.2f seconds" % (
self.channels, self.bitrate, self.sample_rate, self.length)
class _IFFID3(ID3):
"""A AIFF file with ID3v2 tags"""
def _load_header(self):
try:
self._fileobj.seek(IFFFile(self._fileobj)[u'ID3'].data_offset)
except (InvalidChunk, KeyError):
raise ID3Error()
super(_IFFID3, self)._load_header()
def save(self, filename=None, v2_version=4, v23_sep='/'):
"""Save ID3v2 data to the AIFF file"""
framedata = self._prepare_framedata(v2_version, v23_sep)
framesize = len(framedata)
if filename is None:
filename = self.filename
# Unlike the parent ID3.save method, we won't save to a blank file
# since we would have to construct a empty AIFF file
fileobj = open(filename, 'rb+')
iff_file = IFFFile(fileobj)
try:
if u'ID3' not in iff_file:
iff_file.insert_chunk(u'ID3')
chunk = iff_file[u'ID3']
fileobj.seek(chunk.data_offset)
header = fileobj.read(10)
header = self._prepare_id3_header(header, framesize, v2_version)
header, new_size, _ = header
data = header + framedata + (b'\x00' * (new_size - framesize))
# Include ID3 header size in 'new_size' calculation
new_size += 10
# Expand the chunk if necessary, including pad byte
if new_size > chunk.size:
insert_at = chunk.offset + chunk.size
insert_size = new_size - chunk.size + new_size % 2
insert_bytes(fileobj, insert_size, insert_at)
chunk.resize(new_size)
fileobj.seek(chunk.data_offset)
fileobj.write(data)
finally:
fileobj.close()
def delete(self, filename=None):
"""Completely removes the ID3 chunk from the AIFF file"""
if filename is None:
filename = self.filename
delete(filename)
self.clear()
def delete(filename):
"""Completely removes the ID3 chunk from the AIFF file"""
with open(filename, "rb+") as file_:
try:
del IFFFile(file_)[u'ID3']
except KeyError:
pass
class AIFF(FileType):
"""An AIFF audio file.
:ivar info: :class:`AIFFInfo`
:ivar tags: :class:`ID3`
"""
_mimes = ["audio/aiff", "audio/x-aiff"]
@staticmethod
def score(filename, fileobj, header):
filename = filename.lower()
return (header.startswith(b"FORM") * 2 + endswith(filename, b".aif") +
endswith(filename, b".aiff") + endswith(filename, b".aifc"))
def add_tags(self):
"""Add an empty ID3 tag to the file."""
if self.tags is None:
self.tags = _IFFID3()
else:
raise error("an ID3 tag already exists")
def load(self, filename, **kwargs):
"""Load stream and tag information from a file."""
self.filename = filename
try:
self.tags = _IFFID3(filename, **kwargs)
except ID3Error:
self.tags = None
try:
fileobj = open(filename, "rb")
self.info = AIFFInfo(fileobj)
finally:
fileobj.close()
Open = AIFF

714
lib/mutagen/apev2.py Normal file
View File

@@ -0,0 +1,714 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""APEv2 reading and writing.
The APEv2 format is most commonly used with Musepack files, but is
also the format of choice for WavPack and other formats. Some MP3s
also have APEv2 tags, but this can cause problems with many MP3
decoders and taggers.
APEv2 tags, like Vorbis comments, are freeform key=value pairs. APEv2
keys can be any ASCII string with characters from 0x20 to 0x7E,
between 2 and 255 characters long. Keys are case-sensitive, but
readers are recommended to be case insensitive, and it is forbidden to
multiple keys which differ only in case. Keys are usually stored
title-cased (e.g. 'Artist' rather than 'artist').
APEv2 values are slightly more structured than Vorbis comments; values
are flagged as one of text, binary, or an external reference (usually
a URI).
Based off the format specification found at
http://wiki.hydrogenaudio.org/index.php?title=APEv2_specification.
"""
__all__ = ["APEv2", "APEv2File", "Open", "delete"]
import sys
import struct
from collections import MutableSequence
from ._compat import (cBytesIO, PY3, text_type, PY2, reraise, swap_to_string,
xrange)
from mutagen import Metadata, FileType, StreamInfo
from mutagen._util import (DictMixin, cdata, delete_bytes, total_ordering,
MutagenError)
def is_valid_apev2_key(key):
if not isinstance(key, text_type):
if PY3:
raise TypeError("APEv2 key must be str")
try:
key = key.decode('ascii')
except UnicodeDecodeError:
return False
# PY26 - Change to set literal syntax (since set is faster than list here)
return ((2 <= len(key) <= 255) and (min(key) >= u' ') and
(max(key) <= u'~') and
(key not in [u"OggS", u"TAG", u"ID3", u"MP+"]))
# There are three different kinds of APE tag values.
# "0: Item contains text information coded in UTF-8
# 1: Item contains binary information
# 2: Item is a locator of external stored information [e.g. URL]
# 3: reserved"
TEXT, BINARY, EXTERNAL = range(3)
HAS_HEADER = 1 << 31
HAS_NO_FOOTER = 1 << 30
IS_HEADER = 1 << 29
class error(IOError, MutagenError):
pass
class APENoHeaderError(error, ValueError):
pass
class APEUnsupportedVersionError(error, ValueError):
pass
class APEBadItemError(error, ValueError):
pass
class _APEv2Data(object):
# Store offsets of the important parts of the file.
start = header = data = footer = end = None
# Footer or header; seek here and read 32 to get version/size/items/flags
metadata = None
# Actual tag data
tag = None
version = None
size = None
items = None
flags = 0
# The tag is at the start rather than the end. A tag at both
# the start and end of the file (i.e. the tag is the whole file)
# is not considered to be at the start.
is_at_start = False
def __init__(self, fileobj):
self.__find_metadata(fileobj)
if self.header is None:
self.metadata = self.footer
elif self.footer is None:
self.metadata = self.header
else:
self.metadata = max(self.header, self.footer)
if self.metadata is None:
return
self.__fill_missing(fileobj)
self.__fix_brokenness(fileobj)
if self.data is not None:
fileobj.seek(self.data)
self.tag = fileobj.read(self.size)
def __find_metadata(self, fileobj):
# Try to find a header or footer.
# Check for a simple footer.
try:
fileobj.seek(-32, 2)
except IOError:
fileobj.seek(0, 2)
return
if fileobj.read(8) == b"APETAGEX":
fileobj.seek(-8, 1)
self.footer = self.metadata = fileobj.tell()
return
# Check for an APEv2 tag followed by an ID3v1 tag at the end.
try:
fileobj.seek(-128, 2)
if fileobj.read(3) == b"TAG":
fileobj.seek(-35, 1) # "TAG" + header length
if fileobj.read(8) == b"APETAGEX":
fileobj.seek(-8, 1)
self.footer = fileobj.tell()
return
# ID3v1 tag at the end, maybe preceded by Lyrics3v2.
# (http://www.id3.org/lyrics3200.html)
# (header length - "APETAGEX") - "LYRICS200"
fileobj.seek(15, 1)
if fileobj.read(9) == b'LYRICS200':
fileobj.seek(-15, 1) # "LYRICS200" + size tag
try:
offset = int(fileobj.read(6))
except ValueError:
raise IOError
fileobj.seek(-32 - offset - 6, 1)
if fileobj.read(8) == b"APETAGEX":
fileobj.seek(-8, 1)
self.footer = fileobj.tell()
return
except IOError:
pass
# Check for a tag at the start.
fileobj.seek(0, 0)
if fileobj.read(8) == b"APETAGEX":
self.is_at_start = True
self.header = 0
def __fill_missing(self, fileobj):
fileobj.seek(self.metadata + 8)
self.version = fileobj.read(4)
self.size = cdata.uint_le(fileobj.read(4))
self.items = cdata.uint_le(fileobj.read(4))
self.flags = cdata.uint_le(fileobj.read(4))
if self.header is not None:
self.data = self.header + 32
# If we're reading the header, the size is the header
# offset + the size, which includes the footer.
self.end = self.data + self.size
fileobj.seek(self.end - 32, 0)
if fileobj.read(8) == b"APETAGEX":
self.footer = self.end - 32
elif self.footer is not None:
self.end = self.footer + 32
self.data = self.end - self.size
if self.flags & HAS_HEADER:
self.header = self.data - 32
else:
self.header = self.data
else:
raise APENoHeaderError("No APE tag found")
# exclude the footer from size
if self.footer is not None:
self.size -= 32
def __fix_brokenness(self, fileobj):
# Fix broken tags written with PyMusepack.
if self.header is not None:
start = self.header
else:
start = self.data
fileobj.seek(start)
while start > 0:
# Clean up broken writing from pre-Mutagen PyMusepack.
# It didn't remove the first 24 bytes of header.
try:
fileobj.seek(-24, 1)
except IOError:
break
else:
if fileobj.read(8) == b"APETAGEX":
fileobj.seek(-8, 1)
start = fileobj.tell()
else:
break
self.start = start
class _CIDictProxy(DictMixin):
def __init__(self, *args, **kwargs):
self.__casemap = {}
self.__dict = {}
super(_CIDictProxy, self).__init__(*args, **kwargs)
# Internally all names are stored as lowercase, but the case
# they were set with is remembered and used when saving. This
# is roughly in line with the standard, which says that keys
# are case-sensitive but two keys differing only in case are
# not allowed, and recommends case-insensitive
# implementations.
def __getitem__(self, key):
return self.__dict[key.lower()]
def __setitem__(self, key, value):
lower = key.lower()
self.__casemap[lower] = key
self.__dict[lower] = value
def __delitem__(self, key):
lower = key.lower()
del(self.__casemap[lower])
del(self.__dict[lower])
def keys(self):
return [self.__casemap.get(key, key) for key in self.__dict.keys()]
class APEv2(_CIDictProxy, Metadata):
"""A file with an APEv2 tag.
ID3v1 tags are silently ignored and overwritten.
"""
filename = None
def pprint(self):
"""Return tag key=value pairs in a human-readable format."""
items = sorted(self.items())
return u"\n".join(u"%s=%s" % (k, v.pprint()) for k, v in items)
def load(self, filename):
"""Load tags from a filename."""
self.filename = filename
fileobj = open(filename, "rb")
try:
data = _APEv2Data(fileobj)
finally:
fileobj.close()
if data.tag:
self.clear()
self.__parse_tag(data.tag, data.items)
else:
raise APENoHeaderError("No APE tag found")
def __parse_tag(self, tag, count):
fileobj = cBytesIO(tag)
for i in xrange(count):
size_data = fileobj.read(4)
# someone writes wrong item counts
if not size_data:
break
size = cdata.uint_le(size_data)
flags = cdata.uint_le(fileobj.read(4))
# Bits 1 and 2 bits are flags, 0-3
# Bit 0 is read/write flag, ignored
kind = (flags & 6) >> 1
if kind == 3:
raise APEBadItemError("value type must be 0, 1, or 2")
key = value = fileobj.read(1)
while key[-1:] != b'\x00' and value:
value = fileobj.read(1)
key += value
if key[-1:] == b"\x00":
key = key[:-1]
if PY3:
try:
key = key.decode("ascii")
except UnicodeError as err:
reraise(APEBadItemError, err, sys.exc_info()[2])
value = fileobj.read(size)
value = _get_value_type(kind)._new(value)
self[key] = value
def __getitem__(self, key):
if not is_valid_apev2_key(key):
raise KeyError("%r is not a valid APEv2 key" % key)
if PY2:
key = key.encode('ascii')
return super(APEv2, self).__getitem__(key)
def __delitem__(self, key):
if not is_valid_apev2_key(key):
raise KeyError("%r is not a valid APEv2 key" % key)
if PY2:
key = key.encode('ascii')
super(APEv2, self).__delitem__(key)
def __setitem__(self, key, value):
"""'Magic' value setter.
This function tries to guess at what kind of value you want to
store. If you pass in a valid UTF-8 or Unicode string, it
treats it as a text value. If you pass in a list, it treats it
as a list of string/Unicode values. If you pass in a string
that is not valid UTF-8, it assumes it is a binary value.
Python 3: all bytes will be assumed to be a byte value, even
if they are valid utf-8.
If you need to force a specific type of value (e.g. binary
data that also happens to be valid UTF-8, or an external
reference), use the APEValue factory and set the value to the
result of that::
from mutagen.apev2 import APEValue, EXTERNAL
tag['Website'] = APEValue('http://example.org', EXTERNAL)
"""
if not is_valid_apev2_key(key):
raise KeyError("%r is not a valid APEv2 key" % key)
if PY2:
key = key.encode('ascii')
if not isinstance(value, _APEValue):
# let's guess at the content if we're not already a value...
if isinstance(value, text_type):
# unicode? we've got to be text.
value = APEValue(value, TEXT)
elif isinstance(value, list):
items = []
for v in value:
if not isinstance(v, text_type):
if PY3:
raise TypeError("item in list not str")
v = v.decode("utf-8")
items.append(v)
# list? text.
value = APEValue(u"\0".join(items), TEXT)
else:
if PY3:
value = APEValue(value, BINARY)
else:
try:
value.decode("utf-8")
except UnicodeError:
# invalid UTF8 text, probably binary
value = APEValue(value, BINARY)
else:
# valid UTF8, probably text
value = APEValue(value, TEXT)
super(APEv2, self).__setitem__(key, value)
def save(self, filename=None):
"""Save changes to a file.
If no filename is given, the one most recently loaded is used.
Tags are always written at the end of the file, and include
a header and a footer.
"""
filename = filename or self.filename
try:
fileobj = open(filename, "r+b")
except IOError:
fileobj = open(filename, "w+b")
data = _APEv2Data(fileobj)
if data.is_at_start:
delete_bytes(fileobj, data.end - data.start, data.start)
elif data.start is not None:
fileobj.seek(data.start)
# Delete an ID3v1 tag if present, too.
fileobj.truncate()
fileobj.seek(0, 2)
tags = []
for key, value in self.items():
# Packed format for an item:
# 4B: Value length
# 4B: Value type
# Key name
# 1B: Null
# Key value
value_data = value._write()
if not isinstance(key, bytes):
key = key.encode("utf-8")
tag_data = bytearray()
tag_data += struct.pack("<2I", len(value_data), value.kind << 1)
tag_data += key + b"\0" + value_data
tags.append(bytes(tag_data))
# "APE tags items should be sorted ascending by size... This is
# not a MUST, but STRONGLY recommended. Actually the items should
# be sorted by importance/byte, but this is not feasible."
tags.sort(key=len)
num_tags = len(tags)
tags = b"".join(tags)
header = bytearray(b"APETAGEX")
# version, tag size, item count, flags
header += struct.pack("<4I", 2000, len(tags) + 32, num_tags,
HAS_HEADER | IS_HEADER)
header += b"\0" * 8
fileobj.write(header)
fileobj.write(tags)
footer = bytearray(b"APETAGEX")
footer += struct.pack("<4I", 2000, len(tags) + 32, num_tags,
HAS_HEADER)
footer += b"\0" * 8
fileobj.write(footer)
fileobj.close()
def delete(self, filename=None):
"""Remove tags from a file."""
filename = filename or self.filename
fileobj = open(filename, "r+b")
try:
data = _APEv2Data(fileobj)
if data.start is not None and data.size is not None:
delete_bytes(fileobj, data.end - data.start, data.start)
finally:
fileobj.close()
self.clear()
Open = APEv2
def delete(filename):
"""Remove tags from a file."""
try:
APEv2(filename).delete()
except APENoHeaderError:
pass
def _get_value_type(kind):
"""Returns a _APEValue subclass or raises ValueError"""
if kind == TEXT:
return APETextValue
elif kind == BINARY:
return APEBinaryValue
elif kind == EXTERNAL:
return APEExtValue
raise ValueError("unknown kind %r" % kind)
def APEValue(value, kind):
"""APEv2 tag value factory.
Use this if you need to specify the value's type manually. Binary
and text data are automatically detected by APEv2.__setitem__.
"""
try:
type_ = _get_value_type(kind)
except ValueError:
raise ValueError("kind must be TEXT, BINARY, or EXTERNAL")
else:
return type_(value)
class _APEValue(object):
kind = None
value = None
def __init__(self, value, kind=None):
# kind kwarg is for backwards compat
if kind is not None and kind != self.kind:
raise ValueError
self.value = self._validate(value)
@classmethod
def _new(cls, data):
instance = cls.__new__(cls)
instance._parse(data)
return instance
def _parse(self, data):
"""Sets value or raises APEBadItemError"""
raise NotImplementedError
def _write(self):
"""Returns bytes"""
raise NotImplementedError
def _validate(self, value):
"""Returns validated value or raises TypeError/ValueErrr"""
raise NotImplementedError
def __repr__(self):
return "%s(%r, %d)" % (type(self).__name__, self.value, self.kind)
@swap_to_string
@total_ordering
class _APEUtf8Value(_APEValue):
def _parse(self, data):
try:
self.value = data.decode("utf-8")
except UnicodeDecodeError as e:
reraise(APEBadItemError, e, sys.exc_info()[2])
def _validate(self, value):
if not isinstance(value, text_type):
if PY3:
raise TypeError("value not str")
else:
value = value.decode("utf-8")
return value
def _write(self):
return self.value.encode("utf-8")
def __len__(self):
return len(self.value)
def __bytes__(self):
return self._write()
def __eq__(self, other):
return self.value == other
def __lt__(self, other):
return self.value < other
def __str__(self):
return self.value
class APETextValue(_APEUtf8Value, MutableSequence):
"""An APEv2 text value.
Text values are Unicode/UTF-8 strings. They can be accessed like
strings (with a null separating the values), or arrays of strings.
"""
kind = TEXT
def __iter__(self):
"""Iterate over the strings of the value (not the characters)"""
return iter(self.value.split(u"\0"))
def __getitem__(self, index):
return self.value.split(u"\0")[index]
def __len__(self):
return self.value.count(u"\0") + 1
def __setitem__(self, index, value):
if not isinstance(value, text_type):
if PY3:
raise TypeError("value not str")
else:
value = value.decode("utf-8")
values = list(self)
values[index] = value
self.value = u"\0".join(values)
def insert(self, index, value):
if not isinstance(value, text_type):
if PY3:
raise TypeError("value not str")
else:
value = value.decode("utf-8")
values = list(self)
values.insert(index, value)
self.value = u"\0".join(values)
def __delitem__(self, index):
values = list(self)
del values[index]
self.value = u"\0".join(values)
def pprint(self):
return u" / ".join(self)
@swap_to_string
@total_ordering
class APEBinaryValue(_APEValue):
"""An APEv2 binary value."""
kind = BINARY
def _parse(self, data):
self.value = data
def _write(self):
return self.value
def _validate(self, value):
if not isinstance(value, bytes):
raise TypeError("value not bytes")
return bytes(value)
def __len__(self):
return len(self.value)
def __bytes__(self):
return self._write()
def __eq__(self, other):
return self.value == other
def __lt__(self, other):
return self.value < other
def pprint(self):
return u"[%d bytes]" % len(self)
class APEExtValue(_APEUtf8Value):
"""An APEv2 external value.
External values are usually URI or IRI strings.
"""
kind = EXTERNAL
def pprint(self):
return u"[External] %s" % self.value
class APEv2File(FileType):
class _Info(StreamInfo):
length = 0
bitrate = 0
def __init__(self, fileobj):
pass
@staticmethod
def pprint():
return u"Unknown format with APEv2 tag."
def load(self, filename):
self.filename = filename
self.info = self._Info(open(filename, "rb"))
try:
self.tags = APEv2(filename)
except APENoHeaderError:
self.tags = None
def add_tags(self):
if self.tags is None:
self.tags = APEv2()
else:
raise ValueError("%r already has tags: %r" % (self, self.tags))
@staticmethod
def score(filename, fileobj, header):
try:
fileobj.seek(-160, 2)
except IOError:
fileobj.seek(0)
footer = fileobj.read()
return ((b"APETAGEX" in footer) - header.startswith(b"ID3"))

862
lib/mutagen/asf.py Normal file
View File

@@ -0,0 +1,862 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005-2006 Joe Wreschnig
# Copyright (C) 2006-2007 Lukas Lalinsky
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write ASF (Window Media Audio) files."""
__all__ = ["ASF", "Open"]
import sys
import struct
from mutagen import FileType, Metadata, StreamInfo
from mutagen._util import (insert_bytes, delete_bytes, DictMixin,
total_ordering, MutagenError)
from ._compat import swap_to_string, text_type, PY2, string_types, reraise, \
xrange, long_, PY3
class error(IOError, MutagenError):
pass
class ASFError(error):
pass
class ASFHeaderError(error):
pass
class ASFInfo(StreamInfo):
"""ASF stream information."""
def __init__(self):
self.length = 0.0
self.sample_rate = 0
self.bitrate = 0
self.channels = 0
def pprint(self):
s = "Windows Media Audio %d bps, %s Hz, %d channels, %.2f seconds" % (
self.bitrate, self.sample_rate, self.channels, self.length)
return s
class ASFTags(list, DictMixin, Metadata):
"""Dictionary containing ASF attributes."""
def pprint(self):
return "\n".join("%s=%s" % (k, v) for k, v in self)
def __getitem__(self, key):
"""A list of values for the key.
This is a copy, so comment['title'].append('a title') will not
work.
"""
# PY3 only
if isinstance(key, slice):
return list.__getitem__(self, key)
values = [value for (k, value) in self if k == key]
if not values:
raise KeyError(key)
else:
return values
def __delitem__(self, key):
"""Delete all values associated with the key."""
# PY3 only
if isinstance(key, slice):
return list.__delitem__(self, key)
to_delete = [x for x in self if x[0] == key]
if not to_delete:
raise KeyError(key)
else:
for k in to_delete:
self.remove(k)
def __contains__(self, key):
"""Return true if the key has any values."""
for k, value in self:
if k == key:
return True
else:
return False
def __setitem__(self, key, values):
"""Set a key's value or values.
Setting a value overwrites all old ones. The value may be a
list of Unicode or UTF-8 strings, or a single Unicode or UTF-8
string.
"""
# PY3 only
if isinstance(key, slice):
return list.__setitem__(self, key, values)
if not isinstance(values, list):
values = [values]
to_append = []
for value in values:
if not isinstance(value, ASFBaseAttribute):
if isinstance(value, string_types):
value = ASFUnicodeAttribute(value)
elif PY3 and isinstance(value, bytes):
value = ASFByteArrayAttribute(value)
elif isinstance(value, bool):
value = ASFBoolAttribute(value)
elif isinstance(value, int):
value = ASFDWordAttribute(value)
elif isinstance(value, long_):
value = ASFQWordAttribute(value)
else:
raise TypeError("Invalid type %r" % type(value))
to_append.append((key, value))
try:
del(self[key])
except KeyError:
pass
self.extend(to_append)
def keys(self):
"""Return all keys in the comment."""
return self and set(next(iter(zip(*self))))
def as_dict(self):
"""Return a copy of the comment data in a real dict."""
d = {}
for key, value in self:
d.setdefault(key, []).append(value)
return d
class ASFBaseAttribute(object):
"""Generic attribute."""
TYPE = None
def __init__(self, value=None, data=None, language=None,
stream=None, **kwargs):
self.language = language
self.stream = stream
if data:
self.value = self.parse(data, **kwargs)
else:
if value is None:
# we used to support not passing any args and instead assign
# them later, keep that working..
self.value = None
else:
self.value = self._validate(value)
def _validate(self, value):
"""Raises TypeError or ValueError in case the user supplied value
isn't valid.
"""
return value
def data_size(self):
raise NotImplementedError
def __repr__(self):
name = "%s(%r" % (type(self).__name__, self.value)
if self.language:
name += ", language=%d" % self.language
if self.stream:
name += ", stream=%d" % self.stream
name += ")"
return name
def render(self, name):
name = name.encode("utf-16-le") + b"\x00\x00"
data = self._render()
return (struct.pack("<H", len(name)) + name +
struct.pack("<HH", self.TYPE, len(data)) + data)
def render_m(self, name):
name = name.encode("utf-16-le") + b"\x00\x00"
if self.TYPE == 2:
data = self._render(dword=False)
else:
data = self._render()
return (struct.pack("<HHHHI", 0, self.stream or 0, len(name),
self.TYPE, len(data)) + name + data)
def render_ml(self, name):
name = name.encode("utf-16-le") + b"\x00\x00"
if self.TYPE == 2:
data = self._render(dword=False)
else:
data = self._render()
return (struct.pack("<HHHHI", self.language or 0, self.stream or 0,
len(name), self.TYPE, len(data)) + name + data)
@swap_to_string
@total_ordering
class ASFUnicodeAttribute(ASFBaseAttribute):
"""Unicode string attribute."""
TYPE = 0x0000
def parse(self, data):
try:
return data.decode("utf-16-le").strip("\x00")
except UnicodeDecodeError as e:
reraise(ASFError, e, sys.exc_info()[2])
def _validate(self, value):
if not isinstance(value, text_type):
if PY2:
return value.decode("utf-8")
else:
raise TypeError("%r not str" % value)
return value
def _render(self):
return self.value.encode("utf-16-le") + b"\x00\x00"
def data_size(self):
return len(self._render())
def __bytes__(self):
return self.value.encode("utf-16-le")
def __str__(self):
return self.value
def __eq__(self, other):
return text_type(self) == other
def __lt__(self, other):
return text_type(self) < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFByteArrayAttribute(ASFBaseAttribute):
"""Byte array attribute."""
TYPE = 0x0001
def parse(self, data):
assert isinstance(data, bytes)
return data
def _render(self):
assert isinstance(self.value, bytes)
return self.value
def _validate(self, value):
if not isinstance(value, bytes):
raise TypeError("must be bytes/str: %r" % value)
return value
def data_size(self):
return len(self.value)
def __bytes__(self):
return self.value
def __str__(self):
return "[binary data (%d bytes)]" % len(self.value)
def __eq__(self, other):
return self.value == other
def __lt__(self, other):
return self.value < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFBoolAttribute(ASFBaseAttribute):
"""Bool attribute."""
TYPE = 0x0002
def parse(self, data, dword=True):
if dword:
return struct.unpack("<I", data)[0] == 1
else:
return struct.unpack("<H", data)[0] == 1
def _render(self, dword=True):
if dword:
return struct.pack("<I", bool(self.value))
else:
return struct.pack("<H", bool(self.value))
def _validate(self, value):
return bool(value)
def data_size(self):
return 4
def __bool__(self):
return bool(self.value)
def __bytes__(self):
return text_type(self.value).encode('utf-8')
def __str__(self):
return text_type(self.value)
def __eq__(self, other):
return bool(self.value) == other
def __lt__(self, other):
return bool(self.value) < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFDWordAttribute(ASFBaseAttribute):
"""DWORD attribute."""
TYPE = 0x0003
def parse(self, data):
return struct.unpack("<L", data)[0]
def _render(self):
return struct.pack("<L", self.value)
def _validate(self, value):
value = int(value)
if not 0 <= value <= 2 ** 32 - 1:
raise ValueError("Out of range")
return value
def data_size(self):
return 4
def __int__(self):
return self.value
def __bytes__(self):
return text_type(self.value).encode('utf-8')
def __str__(self):
return text_type(self.value)
def __eq__(self, other):
return int(self.value) == other
def __lt__(self, other):
return int(self.value) < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFQWordAttribute(ASFBaseAttribute):
"""QWORD attribute."""
TYPE = 0x0004
def parse(self, data):
return struct.unpack("<Q", data)[0]
def _render(self):
return struct.pack("<Q", self.value)
def _validate(self, value):
value = int(value)
if not 0 <= value <= 2 ** 64 - 1:
raise ValueError("Out of range")
return value
def data_size(self):
return 8
def __int__(self):
return self.value
def __bytes__(self):
return text_type(self.value).encode('utf-8')
def __str__(self):
return text_type(self.value)
def __eq__(self, other):
return int(self.value) == other
def __lt__(self, other):
return int(self.value) < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFWordAttribute(ASFBaseAttribute):
"""WORD attribute."""
TYPE = 0x0005
def parse(self, data):
return struct.unpack("<H", data)[0]
def _render(self):
return struct.pack("<H", self.value)
def _validate(self, value):
value = int(value)
if not 0 <= value <= 2 ** 16 - 1:
raise ValueError("Out of range")
return value
def data_size(self):
return 2
def __int__(self):
return self.value
def __bytes__(self):
return text_type(self.value).encode('utf-8')
def __str__(self):
return text_type(self.value)
def __eq__(self, other):
return int(self.value) == other
def __lt__(self, other):
return int(self.value) < other
__hash__ = ASFBaseAttribute.__hash__
@swap_to_string
@total_ordering
class ASFGUIDAttribute(ASFBaseAttribute):
"""GUID attribute."""
TYPE = 0x0006
def parse(self, data):
assert isinstance(data, bytes)
return data
def _render(self):
assert isinstance(self.value, bytes)
return self.value
def _validate(self, value):
if not isinstance(value, bytes):
raise TypeError("must be bytes/str: %r" % value)
return value
def data_size(self):
return len(self.value)
def __bytes__(self):
return self.value
def __str__(self):
return repr(self.value)
def __eq__(self, other):
return self.value == other
def __lt__(self, other):
return self.value < other
__hash__ = ASFBaseAttribute.__hash__
UNICODE = ASFUnicodeAttribute.TYPE
BYTEARRAY = ASFByteArrayAttribute.TYPE
BOOL = ASFBoolAttribute.TYPE
DWORD = ASFDWordAttribute.TYPE
QWORD = ASFQWordAttribute.TYPE
WORD = ASFWordAttribute.TYPE
GUID = ASFGUIDAttribute.TYPE
def ASFValue(value, kind, **kwargs):
try:
attr_type = _attribute_types[kind]
except KeyError:
raise ValueError("Unknown value type %r" % kind)
else:
return attr_type(value=value, **kwargs)
_attribute_types = {
ASFUnicodeAttribute.TYPE: ASFUnicodeAttribute,
ASFByteArrayAttribute.TYPE: ASFByteArrayAttribute,
ASFBoolAttribute.TYPE: ASFBoolAttribute,
ASFDWordAttribute.TYPE: ASFDWordAttribute,
ASFQWordAttribute.TYPE: ASFQWordAttribute,
ASFWordAttribute.TYPE: ASFWordAttribute,
ASFGUIDAttribute.TYPE: ASFGUIDAttribute,
}
class BaseObject(object):
"""Base ASF object."""
GUID = None
def parse(self, asf, data, fileobj, size):
self.data = data
def render(self, asf):
data = self.GUID + struct.pack("<Q", len(self.data) + 24) + self.data
return data
class UnknownObject(BaseObject):
"""Unknown ASF object."""
def __init__(self, guid):
assert isinstance(guid, bytes)
self.GUID = guid
class HeaderObject(object):
"""ASF header."""
GUID = b"\x30\x26\xB2\x75\x8E\x66\xCF\x11\xA6\xD9\x00\xAA\x00\x62\xCE\x6C"
class ContentDescriptionObject(BaseObject):
"""Content description."""
GUID = b"\x33\x26\xB2\x75\x8E\x66\xCF\x11\xA6\xD9\x00\xAA\x00\x62\xCE\x6C"
NAMES = [
u"Title",
u"Author",
u"Copyright",
u"Description",
u"Rating",
]
def parse(self, asf, data, fileobj, size):
super(ContentDescriptionObject, self).parse(asf, data, fileobj, size)
asf.content_description_obj = self
lengths = struct.unpack("<HHHHH", data[:10])
texts = []
pos = 10
for length in lengths:
end = pos + length
if length > 0:
texts.append(data[pos:end].decode("utf-16-le").strip(u"\x00"))
else:
texts.append(None)
pos = end
for key, value in zip(self.NAMES, texts):
if value is not None:
value = ASFUnicodeAttribute(value=value)
asf._tags.setdefault(self.GUID, []).append((key, value))
def render(self, asf):
def render_text(name):
value = asf.to_content_description.get(name)
if value is not None:
return text_type(value).encode("utf-16-le") + b"\x00\x00"
else:
return b""
texts = [render_text(x) for x in self.NAMES]
data = struct.pack("<HHHHH", *map(len, texts)) + b"".join(texts)
return self.GUID + struct.pack("<Q", 24 + len(data)) + data
class ExtendedContentDescriptionObject(BaseObject):
"""Extended content description."""
GUID = b"\x40\xA4\xD0\xD2\x07\xE3\xD2\x11\x97\xF0\x00\xA0\xC9\x5E\xA8\x50"
def parse(self, asf, data, fileobj, size):
super(ExtendedContentDescriptionObject, self).parse(
asf, data, fileobj, size)
asf.extended_content_description_obj = self
num_attributes, = struct.unpack("<H", data[0:2])
pos = 2
for i in xrange(num_attributes):
name_length, = struct.unpack("<H", data[pos:pos + 2])
pos += 2
name = data[pos:pos + name_length]
name = name.decode("utf-16-le").strip("\x00")
pos += name_length
value_type, value_length = struct.unpack("<HH", data[pos:pos + 4])
pos += 4
value = data[pos:pos + value_length]
pos += value_length
attr = _attribute_types[value_type](data=value)
asf._tags.setdefault(self.GUID, []).append((name, attr))
def render(self, asf):
attrs = asf.to_extended_content_description.items()
data = b"".join(attr.render(name) for (name, attr) in attrs)
data = struct.pack("<QH", 26 + len(data), len(attrs)) + data
return self.GUID + data
class FilePropertiesObject(BaseObject):
"""File properties."""
GUID = b"\xA1\xDC\xAB\x8C\x47\xA9\xCF\x11\x8E\xE4\x00\xC0\x0C\x20\x53\x65"
def parse(self, asf, data, fileobj, size):
super(FilePropertiesObject, self).parse(asf, data, fileobj, size)
length, _, preroll = struct.unpack("<QQQ", data[40:64])
asf.info.length = (length / 10000000.0) - (preroll / 1000.0)
class StreamPropertiesObject(BaseObject):
"""Stream properties."""
GUID = b"\x91\x07\xDC\xB7\xB7\xA9\xCF\x11\x8E\xE6\x00\xC0\x0C\x20\x53\x65"
def parse(self, asf, data, fileobj, size):
super(StreamPropertiesObject, self).parse(asf, data, fileobj, size)
channels, sample_rate, bitrate = struct.unpack("<HII", data[56:66])
asf.info.channels = channels
asf.info.sample_rate = sample_rate
asf.info.bitrate = bitrate * 8
class HeaderExtensionObject(BaseObject):
"""Header extension."""
GUID = b"\xb5\x03\xbf_.\xa9\xcf\x11\x8e\xe3\x00\xc0\x0c Se"
def parse(self, asf, data, fileobj, size):
super(HeaderExtensionObject, self).parse(asf, data, fileobj, size)
asf.header_extension_obj = self
datasize, = struct.unpack("<I", data[18:22])
datapos = 0
self.objects = []
while datapos < datasize:
guid, size = struct.unpack(
"<16sQ", data[22 + datapos:22 + datapos + 24])
if guid in _object_types:
obj = _object_types[guid]()
else:
obj = UnknownObject(guid)
obj.parse(asf, data[22 + datapos + 24:22 + datapos + size],
fileobj, size)
self.objects.append(obj)
datapos += size
def render(self, asf):
data = b"".join(obj.render(asf) for obj in self.objects)
return (self.GUID + struct.pack("<Q", 24 + 16 + 6 + len(data)) +
b"\x11\xD2\xD3\xAB\xBA\xA9\xcf\x11" +
b"\x8E\xE6\x00\xC0\x0C\x20\x53\x65" +
b"\x06\x00" + struct.pack("<I", len(data)) + data)
class MetadataObject(BaseObject):
"""Metadata description."""
GUID = b"\xea\xcb\xf8\xc5\xaf[wH\x84g\xaa\x8cD\xfaL\xca"
def parse(self, asf, data, fileobj, size):
super(MetadataObject, self).parse(asf, data, fileobj, size)
asf.metadata_obj = self
num_attributes, = struct.unpack("<H", data[0:2])
pos = 2
for i in xrange(num_attributes):
(reserved, stream, name_length, value_type,
value_length) = struct.unpack("<HHHHI", data[pos:pos + 12])
pos += 12
name = data[pos:pos + name_length]
name = name.decode("utf-16-le").strip("\x00")
pos += name_length
value = data[pos:pos + value_length]
pos += value_length
args = {'data': value, 'stream': stream}
if value_type == 2:
args['dword'] = False
attr = _attribute_types[value_type](**args)
asf._tags.setdefault(self.GUID, []).append((name, attr))
def render(self, asf):
attrs = asf.to_metadata.items()
data = b"".join([attr.render_m(name) for (name, attr) in attrs])
return (self.GUID + struct.pack("<QH", 26 + len(data), len(attrs)) +
data)
class MetadataLibraryObject(BaseObject):
"""Metadata library description."""
GUID = b"\x94\x1c#D\x98\x94\xd1I\xa1A\x1d\x13NEpT"
def parse(self, asf, data, fileobj, size):
super(MetadataLibraryObject, self).parse(asf, data, fileobj, size)
asf.metadata_library_obj = self
num_attributes, = struct.unpack("<H", data[0:2])
pos = 2
for i in xrange(num_attributes):
(language, stream, name_length, value_type,
value_length) = struct.unpack("<HHHHI", data[pos:pos + 12])
pos += 12
name = data[pos:pos + name_length]
name = name.decode("utf-16-le").strip("\x00")
pos += name_length
value = data[pos:pos + value_length]
pos += value_length
args = {'data': value, 'language': language, 'stream': stream}
if value_type == 2:
args['dword'] = False
attr = _attribute_types[value_type](**args)
asf._tags.setdefault(self.GUID, []).append((name, attr))
def render(self, asf):
attrs = asf.to_metadata_library
data = b"".join([attr.render_ml(name) for (name, attr) in attrs])
return (self.GUID + struct.pack("<QH", 26 + len(data), len(attrs)) +
data)
_object_types = {
ExtendedContentDescriptionObject.GUID: ExtendedContentDescriptionObject,
ContentDescriptionObject.GUID: ContentDescriptionObject,
FilePropertiesObject.GUID: FilePropertiesObject,
StreamPropertiesObject.GUID: StreamPropertiesObject,
HeaderExtensionObject.GUID: HeaderExtensionObject,
MetadataLibraryObject.GUID: MetadataLibraryObject,
MetadataObject.GUID: MetadataObject,
}
class ASF(FileType):
"""An ASF file, probably containing WMA or WMV."""
_mimes = ["audio/x-ms-wma", "audio/x-ms-wmv", "video/x-ms-asf",
"audio/x-wma", "video/x-wmv"]
def load(self, filename):
self.filename = filename
with open(filename, "rb") as fileobj:
self.size = 0
self.size1 = 0
self.size2 = 0
self.offset1 = 0
self.offset2 = 0
self.num_objects = 0
self.info = ASFInfo()
self.tags = ASFTags()
self.__read_file(fileobj)
def save(self):
# Move attributes to the right objects
self.to_content_description = {}
self.to_extended_content_description = {}
self.to_metadata = {}
self.to_metadata_library = []
for name, value in self.tags:
library_only = (value.data_size() > 0xFFFF or value.TYPE == GUID)
can_cont_desc = value.TYPE == UNICODE
if library_only or value.language is not None:
self.to_metadata_library.append((name, value))
elif value.stream is not None:
if name not in self.to_metadata:
self.to_metadata[name] = value
else:
self.to_metadata_library.append((name, value))
elif name in ContentDescriptionObject.NAMES:
if name not in self.to_content_description and can_cont_desc:
self.to_content_description[name] = value
else:
self.to_metadata_library.append((name, value))
else:
if name not in self.to_extended_content_description:
self.to_extended_content_description[name] = value
else:
self.to_metadata_library.append((name, value))
# Add missing objects
if not self.content_description_obj:
self.content_description_obj = \
ContentDescriptionObject()
self.objects.append(self.content_description_obj)
if not self.extended_content_description_obj:
self.extended_content_description_obj = \
ExtendedContentDescriptionObject()
self.objects.append(self.extended_content_description_obj)
if not self.header_extension_obj:
self.header_extension_obj = \
HeaderExtensionObject()
self.objects.append(self.header_extension_obj)
if not self.metadata_obj:
self.metadata_obj = \
MetadataObject()
self.header_extension_obj.objects.append(self.metadata_obj)
if not self.metadata_library_obj:
self.metadata_library_obj = \
MetadataLibraryObject()
self.header_extension_obj.objects.append(self.metadata_library_obj)
# Render the header
data = b"".join([obj.render(self) for obj in self.objects])
data = (HeaderObject.GUID +
struct.pack("<QL", len(data) + 30, len(self.objects)) +
b"\x01\x02" + data)
with open(self.filename, "rb+") as fileobj:
size = len(data)
if size > self.size:
insert_bytes(fileobj, size - self.size, self.size)
if size < self.size:
delete_bytes(fileobj, self.size - size, 0)
fileobj.seek(0)
fileobj.write(data)
self.size = size
self.num_objects = len(self.objects)
def __read_file(self, fileobj):
header = fileobj.read(30)
if len(header) != 30 or header[:16] != HeaderObject.GUID:
raise ASFHeaderError("Not an ASF file.")
self.extended_content_description_obj = None
self.content_description_obj = None
self.header_extension_obj = None
self.metadata_obj = None
self.metadata_library_obj = None
self.size, self.num_objects = struct.unpack("<QL", header[16:28])
self.objects = []
self._tags = {}
for i in xrange(self.num_objects):
self.__read_object(fileobj)
for guid in [ContentDescriptionObject.GUID,
ExtendedContentDescriptionObject.GUID, MetadataObject.GUID,
MetadataLibraryObject.GUID]:
self.tags.extend(self._tags.pop(guid, []))
assert not self._tags
def __read_object(self, fileobj):
guid, size = struct.unpack("<16sQ", fileobj.read(24))
if guid in _object_types:
obj = _object_types[guid]()
else:
obj = UnknownObject(guid)
data = fileobj.read(size - 24)
obj.parse(self, data, fileobj, size)
self.objects.append(obj)
@staticmethod
def score(filename, fileobj, header):
return header.startswith(HeaderObject.GUID) * 2
Open = ASF

534
lib/mutagen/easyid3.py Normal file
View File

@@ -0,0 +1,534 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""Easier access to ID3 tags.
EasyID3 is a wrapper around mutagen.id3.ID3 to make ID3 tags appear
more like Vorbis or APEv2 tags.
"""
import mutagen.id3
from ._compat import iteritems, text_type, PY2
from mutagen import Metadata
from mutagen._util import DictMixin, dict_match
from mutagen.id3 import ID3, error, delete, ID3FileType
__all__ = ['EasyID3', 'Open', 'delete']
class EasyID3KeyError(KeyError, ValueError, error):
"""Raised when trying to get/set an invalid key.
Subclasses both KeyError and ValueError for API compatibility,
catching KeyError is preferred.
"""
class EasyID3(DictMixin, Metadata):
"""A file with an ID3 tag.
Like Vorbis comments, EasyID3 keys are case-insensitive ASCII
strings. Only a subset of ID3 frames are supported by default. Use
EasyID3.RegisterKey and its wrappers to support more.
You can also set the GetFallback, SetFallback, and DeleteFallback
to generic key getter/setter/deleter functions, which are called
if no specific handler is registered for a key. Additionally,
ListFallback can be used to supply an arbitrary list of extra
keys. These can be set on EasyID3 or on individual instances after
creation.
To use an EasyID3 class with mutagen.mp3.MP3::
from mutagen.mp3 import EasyMP3 as MP3
MP3(filename)
Because many of the attributes are constructed on the fly, things
like the following will not work::
ezid3["performer"].append("Joe")
Instead, you must do::
values = ezid3["performer"]
values.append("Joe")
ezid3["performer"] = values
"""
Set = {}
Get = {}
Delete = {}
List = {}
# For compatibility.
valid_keys = Get
GetFallback = None
SetFallback = None
DeleteFallback = None
ListFallback = None
@classmethod
def RegisterKey(cls, key,
getter=None, setter=None, deleter=None, lister=None):
"""Register a new key mapping.
A key mapping is four functions, a getter, setter, deleter,
and lister. The key may be either a string or a glob pattern.
The getter, deleted, and lister receive an ID3 instance and
the requested key name. The setter also receives the desired
value, which will be a list of strings.
The getter, setter, and deleter are used to implement __getitem__,
__setitem__, and __delitem__.
The lister is used to implement keys(). It should return a
list of keys that are actually in the ID3 instance, provided
by its associated getter.
"""
key = key.lower()
if getter is not None:
cls.Get[key] = getter
if setter is not None:
cls.Set[key] = setter
if deleter is not None:
cls.Delete[key] = deleter
if lister is not None:
cls.List[key] = lister
@classmethod
def RegisterTextKey(cls, key, frameid):
"""Register a text key.
If the key you need to register is a simple one-to-one mapping
of ID3 frame name to EasyID3 key, then you can use this
function::
EasyID3.RegisterTextKey("title", "TIT2")
"""
def getter(id3, key):
return list(id3[frameid])
def setter(id3, key, value):
try:
frame = id3[frameid]
except KeyError:
id3.add(mutagen.id3.Frames[frameid](encoding=3, text=value))
else:
frame.encoding = 3
frame.text = value
def deleter(id3, key):
del(id3[frameid])
cls.RegisterKey(key, getter, setter, deleter)
@classmethod
def RegisterTXXXKey(cls, key, desc):
"""Register a user-defined text frame key.
Some ID3 tags are stored in TXXX frames, which allow a
freeform 'description' which acts as a subkey,
e.g. TXXX:BARCODE.::
EasyID3.RegisterTXXXKey('barcode', 'BARCODE').
"""
frameid = "TXXX:" + desc
def getter(id3, key):
return list(id3[frameid])
def setter(id3, key, value):
try:
frame = id3[frameid]
except KeyError:
enc = 0
# Store 8859-1 if we can, per MusicBrainz spec.
for v in value:
if v and max(v) > u'\x7f':
enc = 3
break
id3.add(mutagen.id3.TXXX(encoding=enc, text=value, desc=desc))
else:
frame.text = value
def deleter(id3, key):
del(id3[frameid])
cls.RegisterKey(key, getter, setter, deleter)
def __init__(self, filename=None):
self.__id3 = ID3()
if filename is not None:
self.load(filename)
load = property(lambda s: s.__id3.load,
lambda s, v: setattr(s.__id3, 'load', v))
def save(self, *args, **kwargs):
# ignore v2_version until we support 2.3 here
kwargs.pop("v2_version", None)
self.__id3.save(*args, **kwargs)
delete = property(lambda s: s.__id3.delete,
lambda s, v: setattr(s.__id3, 'delete', v))
filename = property(lambda s: s.__id3.filename,
lambda s, fn: setattr(s.__id3, 'filename', fn))
size = property(lambda s: s.__id3.size,
lambda s, fn: setattr(s.__id3, 'size', s))
def __getitem__(self, key):
key = key.lower()
func = dict_match(self.Get, key, self.GetFallback)
if func is not None:
return func(self.__id3, key)
else:
raise EasyID3KeyError("%r is not a valid key" % key)
def __setitem__(self, key, value):
key = key.lower()
if PY2:
if isinstance(value, basestring):
value = [value]
else:
if isinstance(value, text_type):
value = [value]
func = dict_match(self.Set, key, self.SetFallback)
if func is not None:
return func(self.__id3, key, value)
else:
raise EasyID3KeyError("%r is not a valid key" % key)
def __delitem__(self, key):
key = key.lower()
func = dict_match(self.Delete, key, self.DeleteFallback)
if func is not None:
return func(self.__id3, key)
else:
raise EasyID3KeyError("%r is not a valid key" % key)
def keys(self):
keys = []
for key in self.Get.keys():
if key in self.List:
keys.extend(self.List[key](self.__id3, key))
elif key in self:
keys.append(key)
if self.ListFallback is not None:
keys.extend(self.ListFallback(self.__id3, ""))
return keys
def pprint(self):
"""Print tag key=value pairs."""
strings = []
for key in sorted(self.keys()):
values = self[key]
for value in values:
strings.append("%s=%s" % (key, value))
return "\n".join(strings)
Open = EasyID3
def genre_get(id3, key):
return id3["TCON"].genres
def genre_set(id3, key, value):
try:
frame = id3["TCON"]
except KeyError:
id3.add(mutagen.id3.TCON(encoding=3, text=value))
else:
frame.encoding = 3
frame.genres = value
def genre_delete(id3, key):
del(id3["TCON"])
def date_get(id3, key):
return [stamp.text for stamp in id3["TDRC"].text]
def date_set(id3, key, value):
id3.add(mutagen.id3.TDRC(encoding=3, text=value))
def date_delete(id3, key):
del(id3["TDRC"])
def original_date_get(id3, key):
return [stamp.text for stamp in id3["TDOR"].text]
def original_date_set(id3, key, value):
id3.add(mutagen.id3.TDOR(encoding=3, text=value))
def original_date_delete(id3, key):
del(id3["TDOR"])
def performer_get(id3, key):
people = []
wanted_role = key.split(":", 1)[1]
try:
mcl = id3["TMCL"]
except KeyError:
raise KeyError(key)
for role, person in mcl.people:
if role == wanted_role:
people.append(person)
if people:
return people
else:
raise KeyError(key)
def performer_set(id3, key, value):
wanted_role = key.split(":", 1)[1]
try:
mcl = id3["TMCL"]
except KeyError:
mcl = mutagen.id3.TMCL(encoding=3, people=[])
id3.add(mcl)
mcl.encoding = 3
people = [p for p in mcl.people if p[0] != wanted_role]
for v in value:
people.append((wanted_role, v))
mcl.people = people
def performer_delete(id3, key):
wanted_role = key.split(":", 1)[1]
try:
mcl = id3["TMCL"]
except KeyError:
raise KeyError(key)
people = [p for p in mcl.people if p[0] != wanted_role]
if people == mcl.people:
raise KeyError(key)
elif people:
mcl.people = people
else:
del(id3["TMCL"])
def performer_list(id3, key):
try:
mcl = id3["TMCL"]
except KeyError:
return []
else:
return list(set("performer:" + p[0] for p in mcl.people))
def musicbrainz_trackid_get(id3, key):
return [id3["UFID:http://musicbrainz.org"].data.decode('ascii')]
def musicbrainz_trackid_set(id3, key, value):
if len(value) != 1:
raise ValueError("only one track ID may be set per song")
value = value[0].encode('ascii')
try:
frame = id3["UFID:http://musicbrainz.org"]
except KeyError:
frame = mutagen.id3.UFID(owner="http://musicbrainz.org", data=value)
id3.add(frame)
else:
frame.data = value
def musicbrainz_trackid_delete(id3, key):
del(id3["UFID:http://musicbrainz.org"])
def website_get(id3, key):
urls = [frame.url for frame in id3.getall("WOAR")]
if urls:
return urls
else:
raise EasyID3KeyError(key)
def website_set(id3, key, value):
id3.delall("WOAR")
for v in value:
id3.add(mutagen.id3.WOAR(url=v))
def website_delete(id3, key):
id3.delall("WOAR")
def gain_get(id3, key):
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
raise EasyID3KeyError(key)
else:
return [u"%+f dB" % frame.gain]
def gain_set(id3, key, value):
if len(value) != 1:
raise ValueError(
"there must be exactly one gain value, not %r.", value)
gain = float(value[0].split()[0])
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
frame = mutagen.id3.RVA2(desc=key[11:-5], gain=0, peak=0, channel=1)
id3.add(frame)
frame.gain = gain
def gain_delete(id3, key):
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
pass
else:
if frame.peak:
frame.gain = 0.0
else:
del(id3["RVA2:" + key[11:-5]])
def peak_get(id3, key):
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
raise EasyID3KeyError(key)
else:
return [u"%f" % frame.peak]
def peak_set(id3, key, value):
if len(value) != 1:
raise ValueError(
"there must be exactly one peak value, not %r.", value)
peak = float(value[0])
if peak >= 2 or peak < 0:
raise ValueError("peak must be => 0 and < 2.")
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
frame = mutagen.id3.RVA2(desc=key[11:-5], gain=0, peak=0, channel=1)
id3.add(frame)
frame.peak = peak
def peak_delete(id3, key):
try:
frame = id3["RVA2:" + key[11:-5]]
except KeyError:
pass
else:
if frame.gain:
frame.peak = 0.0
else:
del(id3["RVA2:" + key[11:-5]])
def peakgain_list(id3, key):
keys = []
for frame in id3.getall("RVA2"):
keys.append("replaygain_%s_gain" % frame.desc)
keys.append("replaygain_%s_peak" % frame.desc)
return keys
for frameid, key in iteritems({
"TALB": "album",
"TBPM": "bpm",
"TCMP": "compilation", # iTunes extension
"TCOM": "composer",
"TCOP": "copyright",
"TENC": "encodedby",
"TEXT": "lyricist",
"TLEN": "length",
"TMED": "media",
"TMOO": "mood",
"TIT2": "title",
"TIT3": "version",
"TPE1": "artist",
"TPE2": "performer",
"TPE3": "conductor",
"TPE4": "arranger",
"TPOS": "discnumber",
"TPUB": "organization",
"TRCK": "tracknumber",
"TOLY": "author",
"TSO2": "albumartistsort", # iTunes extension
"TSOA": "albumsort",
"TSOC": "composersort", # iTunes extension
"TSOP": "artistsort",
"TSOT": "titlesort",
"TSRC": "isrc",
"TSST": "discsubtitle",
"TLAN": "language",
}):
EasyID3.RegisterTextKey(key, frameid)
EasyID3.RegisterKey("genre", genre_get, genre_set, genre_delete)
EasyID3.RegisterKey("date", date_get, date_set, date_delete)
EasyID3.RegisterKey("originaldate", original_date_get, original_date_set,
original_date_delete)
EasyID3.RegisterKey(
"performer:*", performer_get, performer_set, performer_delete,
performer_list)
EasyID3.RegisterKey("musicbrainz_trackid", musicbrainz_trackid_get,
musicbrainz_trackid_set, musicbrainz_trackid_delete)
EasyID3.RegisterKey("website", website_get, website_set, website_delete)
EasyID3.RegisterKey(
"replaygain_*_gain", gain_get, gain_set, gain_delete, peakgain_list)
EasyID3.RegisterKey("replaygain_*_peak", peak_get, peak_set, peak_delete)
# At various times, information for this came from
# http://musicbrainz.org/docs/specs/metadata_tags.html
# http://bugs.musicbrainz.org/ticket/1383
# http://musicbrainz.org/doc/MusicBrainzTag
for desc, key in iteritems({
u"MusicBrainz Artist Id": "musicbrainz_artistid",
u"MusicBrainz Album Id": "musicbrainz_albumid",
u"MusicBrainz Album Artist Id": "musicbrainz_albumartistid",
u"MusicBrainz TRM Id": "musicbrainz_trmid",
u"MusicIP PUID": "musicip_puid",
u"MusicMagic Fingerprint": "musicip_fingerprint",
u"MusicBrainz Album Status": "musicbrainz_albumstatus",
u"MusicBrainz Album Type": "musicbrainz_albumtype",
u"MusicBrainz Album Release Country": "releasecountry",
u"MusicBrainz Disc Id": "musicbrainz_discid",
u"ASIN": "asin",
u"ALBUMARTISTSORT": "albumartistsort",
u"BARCODE": "barcode",
u"CATALOGNUMBER": "catalognumber",
u"MusicBrainz Release Track Id": "musicbrainz_releasetrackid",
u"MusicBrainz Release Group Id": "musicbrainz_releasegroupid",
u"MusicBrainz Work Id": "musicbrainz_workid",
u"Acoustid Fingerprint": "acoustid_fingerprint",
u"Acoustid Id": "acoustid_id",
}):
EasyID3.RegisterTXXXKey(key, desc)
class EasyID3FileType(ID3FileType):
"""Like ID3FileType, but uses EasyID3 for tags."""
ID3 = EasyID3

284
lib/mutagen/easymp4.py Normal file
View File

@@ -0,0 +1,284 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2009 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
from mutagen import Metadata
from mutagen._util import DictMixin, dict_match
from mutagen.mp4 import MP4, MP4Tags, error, delete
from ._compat import PY2, text_type, PY3
__all__ = ["EasyMP4Tags", "EasyMP4", "delete", "error"]
class EasyMP4KeyError(error, KeyError, ValueError):
pass
class EasyMP4Tags(DictMixin, Metadata):
"""A file with MPEG-4 iTunes metadata.
Like Vorbis comments, EasyMP4Tags keys are case-insensitive ASCII
strings, and values are a list of Unicode strings (and these lists
are always of length 0 or 1).
If you need access to the full MP4 metadata feature set, you should use
MP4, not EasyMP4.
"""
Set = {}
Get = {}
Delete = {}
List = {}
def __init__(self, *args, **kwargs):
self.__mp4 = MP4Tags(*args, **kwargs)
self.load = self.__mp4.load
self.save = self.__mp4.save
self.delete = self.__mp4.delete
filename = property(lambda s: s.__mp4.filename,
lambda s, fn: setattr(s.__mp4, 'filename', fn))
@classmethod
def RegisterKey(cls, key,
getter=None, setter=None, deleter=None, lister=None):
"""Register a new key mapping.
A key mapping is four functions, a getter, setter, deleter,
and lister. The key may be either a string or a glob pattern.
The getter, deleted, and lister receive an MP4Tags instance
and the requested key name. The setter also receives the
desired value, which will be a list of strings.
The getter, setter, and deleter are used to implement __getitem__,
__setitem__, and __delitem__.
The lister is used to implement keys(). It should return a
list of keys that are actually in the MP4 instance, provided
by its associated getter.
"""
key = key.lower()
if getter is not None:
cls.Get[key] = getter
if setter is not None:
cls.Set[key] = setter
if deleter is not None:
cls.Delete[key] = deleter
if lister is not None:
cls.List[key] = lister
@classmethod
def RegisterTextKey(cls, key, atomid):
"""Register a text key.
If the key you need to register is a simple one-to-one mapping
of MP4 atom name to EasyMP4Tags key, then you can use this
function::
EasyMP4Tags.RegisterTextKey("artist", "\xa9ART")
"""
def getter(tags, key):
return tags[atomid]
def setter(tags, key, value):
tags[atomid] = value
def deleter(tags, key):
del(tags[atomid])
cls.RegisterKey(key, getter, setter, deleter)
@classmethod
def RegisterIntKey(cls, key, atomid, min_value=0, max_value=(2 ** 16) - 1):
"""Register a scalar integer key.
"""
def getter(tags, key):
return list(map(text_type, tags[atomid]))
def setter(tags, key, value):
clamp = lambda x: int(min(max(min_value, x), max_value))
tags[atomid] = [clamp(v) for v in map(int, value)]
def deleter(tags, key):
del(tags[atomid])
cls.RegisterKey(key, getter, setter, deleter)
@classmethod
def RegisterIntPairKey(cls, key, atomid, min_value=0,
max_value=(2 ** 16) - 1):
def getter(tags, key):
ret = []
for (track, total) in tags[atomid]:
if total:
ret.append(u"%d/%d" % (track, total))
else:
ret.append(text_type(track))
return ret
def setter(tags, key, value):
clamp = lambda x: int(min(max(min_value, x), max_value))
data = []
for v in value:
try:
tracks, total = v.split("/")
tracks = clamp(int(tracks))
total = clamp(int(total))
except (ValueError, TypeError):
tracks = clamp(int(v))
total = min_value
data.append((tracks, total))
tags[atomid] = data
def deleter(tags, key):
del(tags[atomid])
cls.RegisterKey(key, getter, setter, deleter)
@classmethod
def RegisterFreeformKey(cls, key, name, mean="com.apple.iTunes"):
"""Register a text key.
If the key you need to register is a simple one-to-one mapping
of MP4 freeform atom (----) and name to EasyMP4Tags key, then
you can use this function::
EasyMP4Tags.RegisterFreeformKey(
"musicbrainz_artistid", "MusicBrainz Artist Id")
"""
atomid = "----:" + mean + ":" + name
def getter(tags, key):
return [s.decode("utf-8", "replace") for s in tags[atomid]]
def setter(tags, key, value):
encoded = []
for v in value:
if not isinstance(v, text_type):
if PY3:
raise TypeError("%r not str" % v)
v = v.decode("utf-8")
encoded.append(v.encode("utf-8"))
tags[atomid] = encoded
def deleter(tags, key):
del(tags[atomid])
cls.RegisterKey(key, getter, setter, deleter)
def __getitem__(self, key):
key = key.lower()
func = dict_match(self.Get, key)
if func is not None:
return func(self.__mp4, key)
else:
raise EasyMP4KeyError("%r is not a valid key" % key)
def __setitem__(self, key, value):
key = key.lower()
if PY2:
if isinstance(value, basestring):
value = [value]
else:
if isinstance(value, text_type):
value = [value]
func = dict_match(self.Set, key)
if func is not None:
return func(self.__mp4, key, value)
else:
raise EasyMP4KeyError("%r is not a valid key" % key)
def __delitem__(self, key):
key = key.lower()
func = dict_match(self.Delete, key)
if func is not None:
return func(self.__mp4, key)
else:
raise EasyMP4KeyError("%r is not a valid key" % key)
def keys(self):
keys = []
for key in self.Get.keys():
if key in self.List:
keys.extend(self.List[key](self.__mp4, key))
elif key in self:
keys.append(key)
return keys
def pprint(self):
"""Print tag key=value pairs."""
strings = []
for key in sorted(self.keys()):
values = self[key]
for value in values:
strings.append("%s=%s" % (key, value))
return "\n".join(strings)
for atomid, key in {
'\xa9nam': 'title',
'\xa9alb': 'album',
'\xa9ART': 'artist',
'aART': 'albumartist',
'\xa9day': 'date',
'\xa9cmt': 'comment',
'desc': 'description',
'\xa9grp': 'grouping',
'\xa9gen': 'genre',
'cprt': 'copyright',
'soal': 'albumsort',
'soaa': 'albumartistsort',
'soar': 'artistsort',
'sonm': 'titlesort',
'soco': 'composersort',
}.items():
EasyMP4Tags.RegisterTextKey(key, atomid)
for name, key in {
'MusicBrainz Artist Id': 'musicbrainz_artistid',
'MusicBrainz Track Id': 'musicbrainz_trackid',
'MusicBrainz Album Id': 'musicbrainz_albumid',
'MusicBrainz Album Artist Id': 'musicbrainz_albumartistid',
'MusicIP PUID': 'musicip_puid',
'MusicBrainz Album Status': 'musicbrainz_albumstatus',
'MusicBrainz Album Type': 'musicbrainz_albumtype',
'MusicBrainz Release Country': 'releasecountry',
}.items():
EasyMP4Tags.RegisterFreeformKey(key, name)
for name, key in {
"tmpo": "bpm",
}.items():
EasyMP4Tags.RegisterIntKey(key, name)
for name, key in {
"trkn": "tracknumber",
"disk": "discnumber",
}.items():
EasyMP4Tags.RegisterIntPairKey(key, name)
class EasyMP4(MP4):
"""Like :class:`MP4 <mutagen.mp4.MP4>`,
but uses :class:`EasyMP4Tags` for tags.
:ivar info: :class:`MP4Info <mutagen.mp4.MP4Info>`
:ivar tags: :class:`EasyMP4Tags`
"""
MP4Tags = EasyMP4Tags
Get = EasyMP4Tags.Get
Set = EasyMP4Tags.Set
Delete = EasyMP4Tags.Delete
List = EasyMP4Tags.List
RegisterTextKey = EasyMP4Tags.RegisterTextKey
RegisterKey = EasyMP4Tags.RegisterKey

838
lib/mutagen/flac.py Normal file
View File

@@ -0,0 +1,838 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""Read and write FLAC Vorbis comments and stream information.
Read more about FLAC at http://flac.sourceforge.net.
FLAC supports arbitrary metadata blocks. The two most interesting ones
are the FLAC stream information block, and the Vorbis comment block;
these are also the only ones Mutagen can currently read.
This module does not handle Ogg FLAC files.
Based off documentation available at
http://flac.sourceforge.net/format.html
"""
__all__ = ["FLAC", "Open", "delete"]
import struct
from ._vorbis import VCommentDict
import mutagen
from ._compat import cBytesIO, endswith, chr_
from mutagen._util import insert_bytes, MutagenError
from mutagen.id3 import BitPaddedInt
from functools import reduce
class error(IOError, MutagenError):
pass
class FLACNoHeaderError(error):
pass
class FLACVorbisError(ValueError, error):
pass
def to_int_be(data):
"""Convert an arbitrarily-long string to a long using big-endian
byte order."""
return reduce(lambda a, b: (a << 8) + b, bytearray(data), 0)
class StrictFileObject(object):
"""Wraps a file-like object and raises an exception if the requested
amount of data to read isn't returned."""
def __init__(self, fileobj):
self._fileobj = fileobj
for m in ["close", "tell", "seek", "write", "name"]:
if hasattr(fileobj, m):
setattr(self, m, getattr(fileobj, m))
def read(self, size=-1):
data = self._fileobj.read(size)
if size >= 0 and len(data) != size:
raise error("file said %d bytes, read %d bytes" % (
size, len(data)))
return data
def tryread(self, *args):
return self._fileobj.read(*args)
class MetadataBlock(object):
"""A generic block of FLAC metadata.
This class is extended by specific used as an ancestor for more specific
blocks, and also as a container for data blobs of unknown blocks.
Attributes:
* data -- raw binary data for this block
"""
_distrust_size = False
def __init__(self, data):
"""Parse the given data string or file-like as a metadata block.
The metadata header should not be included."""
if data is not None:
if not isinstance(data, StrictFileObject):
if isinstance(data, bytes):
data = cBytesIO(data)
elif not hasattr(data, 'read'):
raise TypeError(
"StreamInfo requires string data or a file-like")
data = StrictFileObject(data)
self.load(data)
def load(self, data):
self.data = data.read()
def write(self):
return self.data
@staticmethod
def writeblocks(blocks):
"""Render metadata block as a byte string."""
data = []
codes = [[block.code, block.write()] for block in blocks]
codes[-1][0] |= 128
for code, datum in codes:
byte = chr_(code)
if len(datum) > 2 ** 24:
raise error("block is too long to write")
length = struct.pack(">I", len(datum))[-3:]
data.append(byte + length + datum)
return b"".join(data)
@staticmethod
def group_padding(blocks):
"""Consolidate FLAC padding metadata blocks.
The overall size of the rendered blocks does not change, so
this adds several bytes of padding for each merged block.
"""
paddings = [b for b in blocks if isinstance(b, Padding)]
for p in paddings:
blocks.remove(p)
# total padding size is the sum of padding sizes plus 4 bytes
# per removed header.
size = sum(padding.length for padding in paddings)
padding = Padding()
padding.length = size + 4 * (len(paddings) - 1)
blocks.append(padding)
class StreamInfo(MetadataBlock, mutagen.StreamInfo):
"""FLAC stream information.
This contains information about the audio data in the FLAC file.
Unlike most stream information objects in Mutagen, changes to this
one will rewritten to the file when it is saved. Unless you are
actually changing the audio stream itself, don't change any
attributes of this block.
Attributes:
* min_blocksize -- minimum audio block size
* max_blocksize -- maximum audio block size
* sample_rate -- audio sample rate in Hz
* channels -- audio channels (1 for mono, 2 for stereo)
* bits_per_sample -- bits per sample
* total_samples -- total samples in file
* length -- audio length in seconds
"""
code = 0
def __eq__(self, other):
try:
return (self.min_blocksize == other.min_blocksize and
self.max_blocksize == other.max_blocksize and
self.sample_rate == other.sample_rate and
self.channels == other.channels and
self.bits_per_sample == other.bits_per_sample and
self.total_samples == other.total_samples)
except:
return False
__hash__ = MetadataBlock.__hash__
def load(self, data):
self.min_blocksize = int(to_int_be(data.read(2)))
self.max_blocksize = int(to_int_be(data.read(2)))
self.min_framesize = int(to_int_be(data.read(3)))
self.max_framesize = int(to_int_be(data.read(3)))
# first 16 bits of sample rate
sample_first = to_int_be(data.read(2))
# last 4 bits of sample rate, 3 of channels, first 1 of bits/sample
sample_channels_bps = to_int_be(data.read(1))
# last 4 of bits/sample, 36 of total samples
bps_total = to_int_be(data.read(5))
sample_tail = sample_channels_bps >> 4
self.sample_rate = int((sample_first << 4) + sample_tail)
if not self.sample_rate:
raise error("A sample rate value of 0 is invalid")
self.channels = int(((sample_channels_bps >> 1) & 7) + 1)
bps_tail = bps_total >> 36
bps_head = (sample_channels_bps & 1) << 4
self.bits_per_sample = int(bps_head + bps_tail + 1)
self.total_samples = bps_total & 0xFFFFFFFFF
self.length = self.total_samples / float(self.sample_rate)
self.md5_signature = to_int_be(data.read(16))
def write(self):
f = cBytesIO()
f.write(struct.pack(">I", self.min_blocksize)[-2:])
f.write(struct.pack(">I", self.max_blocksize)[-2:])
f.write(struct.pack(">I", self.min_framesize)[-3:])
f.write(struct.pack(">I", self.max_framesize)[-3:])
# first 16 bits of sample rate
f.write(struct.pack(">I", self.sample_rate >> 4)[-2:])
# 4 bits sample, 3 channel, 1 bps
byte = (self.sample_rate & 0xF) << 4
byte += ((self.channels - 1) & 7) << 1
byte += ((self.bits_per_sample - 1) >> 4) & 1
f.write(chr_(byte))
# 4 bits of bps, 4 of sample count
byte = ((self.bits_per_sample - 1) & 0xF) << 4
byte += (self.total_samples >> 32) & 0xF
f.write(chr_(byte))
# last 32 of sample count
f.write(struct.pack(">I", self.total_samples & 0xFFFFFFFF))
# MD5 signature
sig = self.md5_signature
f.write(struct.pack(
">4I", (sig >> 96) & 0xFFFFFFFF, (sig >> 64) & 0xFFFFFFFF,
(sig >> 32) & 0xFFFFFFFF, sig & 0xFFFFFFFF))
return f.getvalue()
def pprint(self):
return "FLAC, %.2f seconds, %d Hz" % (self.length, self.sample_rate)
class SeekPoint(tuple):
"""A single seek point in a FLAC file.
Placeholder seek points have first_sample of 0xFFFFFFFFFFFFFFFFL,
and byte_offset and num_samples undefined. Seek points must be
sorted in ascending order by first_sample number. Seek points must
be unique by first_sample number, except for placeholder
points. Placeholder points must occur last in the table and there
may be any number of them.
Attributes:
* first_sample -- sample number of first sample in the target frame
* byte_offset -- offset from first frame to target frame
* num_samples -- number of samples in target frame
"""
def __new__(cls, first_sample, byte_offset, num_samples):
return super(cls, SeekPoint).__new__(
cls, (first_sample, byte_offset, num_samples))
first_sample = property(lambda self: self[0])
byte_offset = property(lambda self: self[1])
num_samples = property(lambda self: self[2])
class SeekTable(MetadataBlock):
"""Read and write FLAC seek tables.
Attributes:
* seekpoints -- list of SeekPoint objects
"""
__SEEKPOINT_FORMAT = '>QQH'
__SEEKPOINT_SIZE = struct.calcsize(__SEEKPOINT_FORMAT)
code = 3
def __init__(self, data):
self.seekpoints = []
super(SeekTable, self).__init__(data)
def __eq__(self, other):
try:
return (self.seekpoints == other.seekpoints)
except (AttributeError, TypeError):
return False
__hash__ = MetadataBlock.__hash__
def load(self, data):
self.seekpoints = []
sp = data.tryread(self.__SEEKPOINT_SIZE)
while len(sp) == self.__SEEKPOINT_SIZE:
self.seekpoints.append(SeekPoint(
*struct.unpack(self.__SEEKPOINT_FORMAT, sp)))
sp = data.tryread(self.__SEEKPOINT_SIZE)
def write(self):
f = cBytesIO()
for seekpoint in self.seekpoints:
packed = struct.pack(
self.__SEEKPOINT_FORMAT,
seekpoint.first_sample, seekpoint.byte_offset,
seekpoint.num_samples)
f.write(packed)
return f.getvalue()
def __repr__(self):
return "<%s seekpoints=%r>" % (type(self).__name__, self.seekpoints)
class VCFLACDict(VCommentDict):
"""Read and write FLAC Vorbis comments.
FLACs don't use the framing bit at the end of the comment block.
So this extends VCommentDict to not use the framing bit.
"""
code = 4
_distrust_size = True
def load(self, data, errors='replace', framing=False):
super(VCFLACDict, self).load(data, errors=errors, framing=framing)
def write(self, framing=False):
return super(VCFLACDict, self).write(framing=framing)
class CueSheetTrackIndex(tuple):
"""Index for a track in a cuesheet.
For CD-DA, an index_number of 0 corresponds to the track
pre-gap. The first index in a track must have a number of 0 or 1,
and subsequently, index_numbers must increase by 1. Index_numbers
must be unique within a track. And index_offset must be evenly
divisible by 588 samples.
Attributes:
* index_number -- index point number
* index_offset -- offset in samples from track start
"""
def __new__(cls, index_number, index_offset):
return super(cls, CueSheetTrackIndex).__new__(
cls, (index_number, index_offset))
index_number = property(lambda self: self[0])
index_offset = property(lambda self: self[1])
class CueSheetTrack(object):
"""A track in a cuesheet.
For CD-DA, track_numbers must be 1-99, or 170 for the
lead-out. Track_numbers must be unique within a cue sheet. There
must be atleast one index in every track except the lead-out track
which must have none.
Attributes:
* track_number -- track number
* start_offset -- track offset in samples from start of FLAC stream
* isrc -- ISRC code
* type -- 0 for audio, 1 for digital data
* pre_emphasis -- true if the track is recorded with pre-emphasis
* indexes -- list of CueSheetTrackIndex objects
"""
def __init__(self, track_number, start_offset, isrc='', type_=0,
pre_emphasis=False):
self.track_number = track_number
self.start_offset = start_offset
self.isrc = isrc
self.type = type_
self.pre_emphasis = pre_emphasis
self.indexes = []
def __eq__(self, other):
try:
return (self.track_number == other.track_number and
self.start_offset == other.start_offset and
self.isrc == other.isrc and
self.type == other.type and
self.pre_emphasis == other.pre_emphasis and
self.indexes == other.indexes)
except (AttributeError, TypeError):
return False
__hash__ = object.__hash__
def __repr__(self):
return (("<%s number=%r, offset=%d, isrc=%r, type=%r, "
"pre_emphasis=%r, indexes=%r)>") %
(type(self).__name__, self.track_number, self.start_offset,
self.isrc, self.type, self.pre_emphasis, self.indexes))
class CueSheet(MetadataBlock):
"""Read and write FLAC embedded cue sheets.
Number of tracks should be from 1 to 100. There should always be
exactly one lead-out track and that track must be the last track
in the cue sheet.
Attributes:
* media_catalog_number -- media catalog number in ASCII
* lead_in_samples -- number of lead-in samples
* compact_disc -- true if the cuesheet corresponds to a compact disc
* tracks -- list of CueSheetTrack objects
* lead_out -- lead-out as CueSheetTrack or None if lead-out was not found
"""
__CUESHEET_FORMAT = '>128sQB258xB'
__CUESHEET_SIZE = struct.calcsize(__CUESHEET_FORMAT)
__CUESHEET_TRACK_FORMAT = '>QB12sB13xB'
__CUESHEET_TRACK_SIZE = struct.calcsize(__CUESHEET_TRACK_FORMAT)
__CUESHEET_TRACKINDEX_FORMAT = '>QB3x'
__CUESHEET_TRACKINDEX_SIZE = struct.calcsize(__CUESHEET_TRACKINDEX_FORMAT)
code = 5
media_catalog_number = b''
lead_in_samples = 88200
compact_disc = True
def __init__(self, data):
self.tracks = []
super(CueSheet, self).__init__(data)
def __eq__(self, other):
try:
return (self.media_catalog_number == other.media_catalog_number and
self.lead_in_samples == other.lead_in_samples and
self.compact_disc == other.compact_disc and
self.tracks == other.tracks)
except (AttributeError, TypeError):
return False
__hash__ = MetadataBlock.__hash__
def load(self, data):
header = data.read(self.__CUESHEET_SIZE)
media_catalog_number, lead_in_samples, flags, num_tracks = \
struct.unpack(self.__CUESHEET_FORMAT, header)
self.media_catalog_number = media_catalog_number.rstrip(b'\0')
self.lead_in_samples = lead_in_samples
self.compact_disc = bool(flags & 0x80)
self.tracks = []
for i in range(num_tracks):
track = data.read(self.__CUESHEET_TRACK_SIZE)
start_offset, track_number, isrc_padded, flags, num_indexes = \
struct.unpack(self.__CUESHEET_TRACK_FORMAT, track)
isrc = isrc_padded.rstrip(b'\0')
type_ = (flags & 0x80) >> 7
pre_emphasis = bool(flags & 0x40)
val = CueSheetTrack(
track_number, start_offset, isrc, type_, pre_emphasis)
for j in range(num_indexes):
index = data.read(self.__CUESHEET_TRACKINDEX_SIZE)
index_offset, index_number = struct.unpack(
self.__CUESHEET_TRACKINDEX_FORMAT, index)
val.indexes.append(
CueSheetTrackIndex(index_number, index_offset))
self.tracks.append(val)
def write(self):
f = cBytesIO()
flags = 0
if self.compact_disc:
flags |= 0x80
packed = struct.pack(
self.__CUESHEET_FORMAT, self.media_catalog_number,
self.lead_in_samples, flags, len(self.tracks))
f.write(packed)
for track in self.tracks:
track_flags = 0
track_flags |= (track.type & 1) << 7
if track.pre_emphasis:
track_flags |= 0x40
track_packed = struct.pack(
self.__CUESHEET_TRACK_FORMAT, track.start_offset,
track.track_number, track.isrc, track_flags,
len(track.indexes))
f.write(track_packed)
for index in track.indexes:
index_packed = struct.pack(
self.__CUESHEET_TRACKINDEX_FORMAT,
index.index_offset, index.index_number)
f.write(index_packed)
return f.getvalue()
def __repr__(self):
return (("<%s media_catalog_number=%r, lead_in=%r, compact_disc=%r, "
"tracks=%r>") %
(type(self).__name__, self.media_catalog_number,
self.lead_in_samples, self.compact_disc, self.tracks))
class Picture(MetadataBlock):
"""Read and write FLAC embed pictures.
Attributes:
* type -- picture type (same as types for ID3 APIC frames)
* mime -- MIME type of the picture
* desc -- picture's description
* width -- width in pixels
* height -- height in pixels
* depth -- color depth in bits-per-pixel
* colors -- number of colors for indexed palettes (like GIF),
0 for non-indexed
* data -- picture data
"""
code = 6
_distrust_size = True
def __init__(self, data=None):
self.type = 0
self.mime = u''
self.desc = u''
self.width = 0
self.height = 0
self.depth = 0
self.colors = 0
self.data = b''
super(Picture, self).__init__(data)
def __eq__(self, other):
try:
return (self.type == other.type and
self.mime == other.mime and
self.desc == other.desc and
self.width == other.width and
self.height == other.height and
self.depth == other.depth and
self.colors == other.colors and
self.data == other.data)
except (AttributeError, TypeError):
return False
__hash__ = MetadataBlock.__hash__
def load(self, data):
self.type, length = struct.unpack('>2I', data.read(8))
self.mime = data.read(length).decode('UTF-8', 'replace')
length, = struct.unpack('>I', data.read(4))
self.desc = data.read(length).decode('UTF-8', 'replace')
(self.width, self.height, self.depth,
self.colors, length) = struct.unpack('>5I', data.read(20))
self.data = data.read(length)
def write(self):
f = cBytesIO()
mime = self.mime.encode('UTF-8')
f.write(struct.pack('>2I', self.type, len(mime)))
f.write(mime)
desc = self.desc.encode('UTF-8')
f.write(struct.pack('>I', len(desc)))
f.write(desc)
f.write(struct.pack('>5I', self.width, self.height, self.depth,
self.colors, len(self.data)))
f.write(self.data)
return f.getvalue()
def __repr__(self):
return "<%s '%s' (%d bytes)>" % (type(self).__name__, self.mime,
len(self.data))
class Padding(MetadataBlock):
"""Empty padding space for metadata blocks.
To avoid rewriting the entire FLAC file when editing comments,
metadata is often padded. Padding should occur at the end, and no
more than one padding block should be in any FLAC file. Mutagen
handles this with MetadataBlock.group_padding.
"""
code = 1
def __init__(self, data=b""):
super(Padding, self).__init__(data)
def load(self, data):
self.length = len(data.read())
def write(self):
try:
return b"\x00" * self.length
# On some 64 bit platforms this won't generate a MemoryError
# or OverflowError since you might have enough RAM, but it
# still generates a ValueError. On other 64 bit platforms,
# this will still succeed for extremely large values.
# Those should never happen in the real world, and if they
# do, writeblocks will catch it.
except (OverflowError, ValueError, MemoryError):
raise error("cannot write %d bytes" % self.length)
def __eq__(self, other):
return isinstance(other, Padding) and self.length == other.length
__hash__ = MetadataBlock.__hash__
def __repr__(self):
return "<%s (%d bytes)>" % (type(self).__name__, self.length)
class FLAC(mutagen.FileType):
"""A FLAC audio file.
Attributes:
* info -- stream information (length, bitrate, sample rate)
* tags -- metadata tags, if any
* cuesheet -- CueSheet object, if any
* seektable -- SeekTable object, if any
* pictures -- list of embedded pictures
"""
_mimes = ["audio/x-flac", "application/x-flac"]
METADATA_BLOCKS = [StreamInfo, Padding, None, SeekTable, VCFLACDict,
CueSheet, Picture]
"""Known metadata block types, indexed by ID."""
@staticmethod
def score(filename, fileobj, header_data):
return (header_data.startswith(b"fLaC") +
endswith(filename.lower(), ".flac") * 3)
def __read_metadata_block(self, fileobj):
byte = ord(fileobj.read(1))
size = to_int_be(fileobj.read(3))
code = byte & 0x7F
last_block = bool(byte & 0x80)
try:
block_type = self.METADATA_BLOCKS[code] or MetadataBlock
except IndexError:
block_type = MetadataBlock
if block_type._distrust_size:
# Some jackass is writing broken Metadata block length
# for Vorbis comment blocks, and the FLAC reference
# implementaton can parse them (mostly by accident),
# so we have to too. Instead of parsing the size
# given, parse an actual Vorbis comment, leaving
# fileobj in the right position.
# http://code.google.com/p/mutagen/issues/detail?id=52
# ..same for the Picture block:
# http://code.google.com/p/mutagen/issues/detail?id=106
block = block_type(fileobj)
else:
data = fileobj.read(size)
block = block_type(data)
block.code = code
if block.code == VCFLACDict.code:
if self.tags is None:
self.tags = block
else:
raise FLACVorbisError("> 1 Vorbis comment block found")
elif block.code == CueSheet.code:
if self.cuesheet is None:
self.cuesheet = block
else:
raise error("> 1 CueSheet block found")
elif block.code == SeekTable.code:
if self.seektable is None:
self.seektable = block
else:
raise error("> 1 SeekTable block found")
self.metadata_blocks.append(block)
return not last_block
def add_tags(self):
"""Add a Vorbis comment block to the file."""
if self.tags is None:
self.tags = VCFLACDict()
self.metadata_blocks.append(self.tags)
else:
raise FLACVorbisError("a Vorbis comment already exists")
add_vorbiscomment = add_tags
def delete(self, filename=None):
"""Remove Vorbis comments from a file.
If no filename is given, the one most recently loaded is used.
"""
if filename is None:
filename = self.filename
for s in list(self.metadata_blocks):
if isinstance(s, VCFLACDict):
self.metadata_blocks.remove(s)
self.tags = None
self.save()
break
vc = property(lambda s: s.tags, doc="Alias for tags; don't use this.")
def load(self, filename):
"""Load file information from a filename."""
self.metadata_blocks = []
self.tags = None
self.cuesheet = None
self.seektable = None
self.filename = filename
fileobj = StrictFileObject(open(filename, "rb"))
try:
self.__check_header(fileobj)
while self.__read_metadata_block(fileobj):
pass
finally:
fileobj.close()
try:
self.metadata_blocks[0].length
except (AttributeError, IndexError):
raise FLACNoHeaderError("Stream info block not found")
@property
def info(self):
return self.metadata_blocks[0]
def add_picture(self, picture):
"""Add a new picture to the file."""
self.metadata_blocks.append(picture)
def clear_pictures(self):
"""Delete all pictures from the file."""
blocks = [b for b in self.metadata_blocks if b.code != Picture.code]
self.metadata_blocks = blocks
@property
def pictures(self):
"""List of embedded pictures"""
return [b for b in self.metadata_blocks if b.code == Picture.code]
def save(self, filename=None, deleteid3=False):
"""Save metadata blocks to a file.
If no filename is given, the one most recently loaded is used.
"""
if filename is None:
filename = self.filename
f = open(filename, 'rb+')
try:
# Ensure we've got padding at the end, and only at the end.
# If adding makes it too large, we'll scale it down later.
self.metadata_blocks.append(Padding(b'\x00' * 1020))
MetadataBlock.group_padding(self.metadata_blocks)
header = self.__check_header(f)
# "fLaC" and maybe ID3
available = self.__find_audio_offset(f) - header
data = MetadataBlock.writeblocks(self.metadata_blocks)
# Delete ID3v2
if deleteid3 and header > 4:
available += header - 4
header = 4
if len(data) > available:
# If we have too much data, see if we can reduce padding.
padding = self.metadata_blocks[-1]
newlength = padding.length - (len(data) - available)
if newlength > 0:
padding.length = newlength
data = MetadataBlock.writeblocks(self.metadata_blocks)
assert len(data) == available
elif len(data) < available:
# If we have too little data, increase padding.
self.metadata_blocks[-1].length += (available - len(data))
data = MetadataBlock.writeblocks(self.metadata_blocks)
assert len(data) == available
if len(data) != available:
# We couldn't reduce the padding enough.
diff = (len(data) - available)
insert_bytes(f, diff, header)
f.seek(header - 4)
f.write(b"fLaC" + data)
# Delete ID3v1
if deleteid3:
try:
f.seek(-128, 2)
except IOError:
pass
else:
if f.read(3) == b"TAG":
f.seek(-128, 2)
f.truncate()
finally:
f.close()
def __find_audio_offset(self, fileobj):
byte = 0x00
while not (byte & 0x80):
byte = ord(fileobj.read(1))
size = to_int_be(fileobj.read(3))
try:
block_type = self.METADATA_BLOCKS[byte & 0x7F]
except IndexError:
block_type = None
if block_type and block_type._distrust_size:
# See comments in read_metadata_block; the size can't
# be trusted for Vorbis comment blocks and Picture block
block_type(fileobj)
else:
fileobj.read(size)
return fileobj.tell()
def __check_header(self, fileobj):
size = 4
header = fileobj.read(4)
if header != b"fLaC":
size = None
if header[:3] == b"ID3":
size = 14 + BitPaddedInt(fileobj.read(6)[2:])
fileobj.seek(size - 4)
if fileobj.read(4) != b"fLaC":
size = None
if size is None:
raise FLACNoHeaderError(
"%r is not a valid FLAC file" % fileobj.name)
return size
Open = FLAC
def delete(filename):
"""Remove tags from a file."""
FLAC(filename).delete()

1034
lib/mutagen/id3/__init__.py Normal file
View File

File diff suppressed because it is too large Load Diff

1916
lib/mutagen/id3/_frames.py Normal file
View File

File diff suppressed because it is too large Load Diff

543
lib/mutagen/id3/_specs.py Normal file
View File

@@ -0,0 +1,543 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005 Michael Urman
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
import struct
from struct import unpack, pack
from warnings import warn
from .._compat import text_type, chr_, PY3, swap_to_string, string_types
from .._util import total_ordering, decode_terminated, enum
from ._util import ID3JunkFrameError, ID3Warning, BitPaddedInt
class Spec(object):
def __init__(self, name):
self.name = name
def __hash__(self):
raise TypeError("Spec objects are unhashable")
def _validate23(self, frame, value, **kwargs):
"""Return a possibly modified value which, if written,
results in valid id3v2.3 data.
"""
return value
def read(self, frame, value):
raise NotImplementedError
def write(self, frame, value):
raise NotImplementedError
def validate(self, frame, value):
"""Returns the validated data or raises ValueError/TypeError"""
raise NotImplementedError
class ByteSpec(Spec):
def read(self, frame, data):
return bytearray(data)[0], data[1:]
def write(self, frame, value):
return chr_(value)
def validate(self, frame, value):
if value is not None:
chr_(value)
return value
class IntegerSpec(Spec):
def read(self, frame, data):
return int(BitPaddedInt(data, bits=8)), b''
def write(self, frame, value):
return BitPaddedInt.to_str(value, bits=8, width=-1)
def validate(self, frame, value):
return value
class SizedIntegerSpec(Spec):
def __init__(self, name, size):
self.name, self.__sz = name, size
def read(self, frame, data):
return int(BitPaddedInt(data[:self.__sz], bits=8)), data[self.__sz:]
def write(self, frame, value):
return BitPaddedInt.to_str(value, bits=8, width=self.__sz)
def validate(self, frame, value):
return value
@enum
class Encoding(object):
LATIN1 = 0
UTF16 = 1
UTF16BE = 2
UTF8 = 3
class EncodingSpec(ByteSpec):
def read(self, frame, data):
enc, data = super(EncodingSpec, self).read(frame, data)
if enc < 16:
return enc, data
else:
return 0, chr_(enc) + data
def validate(self, frame, value):
if value is None:
return None
if 0 <= value <= 3:
return value
raise ValueError('Invalid Encoding: %r' % value)
def _validate23(self, frame, value, **kwargs):
# only 0, 1 are valid in v2.3, default to utf-16
return min(1, value)
class StringSpec(Spec):
"""A fixed size ASCII only payload."""
def __init__(self, name, length):
super(StringSpec, self).__init__(name)
self.len = length
def read(s, frame, data):
chunk = data[:s.len]
try:
ascii = chunk.decode("ascii")
except UnicodeDecodeError:
raise ID3JunkFrameError("not ascii")
else:
if PY3:
chunk = ascii
return chunk, data[s.len:]
def write(s, frame, value):
if value is None:
return b'\x00' * s.len
else:
if PY3:
value = value.encode("ascii")
return (bytes(value) + b'\x00' * s.len)[:s.len]
def validate(s, frame, value):
if value is None:
return None
if PY3:
if not isinstance(value, str):
raise TypeError("%s has to be str" % s.name)
value.encode("ascii")
else:
if not isinstance(value, bytes):
value = value.encode("ascii")
if len(value) == s.len:
return value
raise ValueError('Invalid StringSpec[%d] data: %r' % (s.len, value))
class BinaryDataSpec(Spec):
def read(self, frame, data):
return data, b''
def write(self, frame, value):
if value is None:
return b""
if isinstance(value, bytes):
return value
value = text_type(value).encode("ascii")
return value
def validate(self, frame, value):
if value is None:
return None
if isinstance(value, bytes):
return value
elif PY3:
raise TypeError("%s has to be bytes" % self.name)
value = text_type(value).encode("ascii")
return value
class EncodedTextSpec(Spec):
# Okay, seriously. This is private and defined explicitly and
# completely by the ID3 specification. You can't just add
# encodings here however you want.
_encodings = (
('latin1', b'\x00'),
('utf16', b'\x00\x00'),
('utf_16_be', b'\x00\x00'),
('utf8', b'\x00')
)
def read(self, frame, data):
enc, term = self._encodings[frame.encoding]
try:
# allow missing termination
return decode_terminated(data, enc, strict=False)
except ValueError:
# utf-16 termination with missing BOM, or single NULL
if not data[:len(term)].strip(b"\x00"):
return u"", data[len(term):]
# utf-16 data with single NULL, see issue 169
try:
return decode_terminated(data + b"\x00", enc)
except ValueError:
raise ID3JunkFrameError
def write(self, frame, value):
enc, term = self._encodings[frame.encoding]
return value.encode(enc) + term
def validate(self, frame, value):
return text_type(value)
class MultiSpec(Spec):
def __init__(self, name, *specs, **kw):
super(MultiSpec, self).__init__(name)
self.specs = specs
self.sep = kw.get('sep')
def read(self, frame, data):
values = []
while data:
record = []
for spec in self.specs:
value, data = spec.read(frame, data)
record.append(value)
if len(self.specs) != 1:
values.append(record)
else:
values.append(record[0])
return values, data
def write(self, frame, value):
data = []
if len(self.specs) == 1:
for v in value:
data.append(self.specs[0].write(frame, v))
else:
for record in value:
for v, s in zip(record, self.specs):
data.append(s.write(frame, v))
return b''.join(data)
def validate(self, frame, value):
if value is None:
return []
if self.sep and isinstance(value, string_types):
value = value.split(self.sep)
if isinstance(value, list):
if len(self.specs) == 1:
return [self.specs[0].validate(frame, v) for v in value]
else:
return [
[s.validate(frame, v) for (v, s) in zip(val, self.specs)]
for val in value]
raise ValueError('Invalid MultiSpec data: %r' % value)
def _validate23(self, frame, value, **kwargs):
if len(self.specs) != 1:
return [[s._validate23(frame, v, **kwargs)
for (v, s) in zip(val, self.specs)]
for val in value]
spec = self.specs[0]
# Merge single text spec multispecs only.
# (TimeStampSpec beeing the exception, but it's not a valid v2.3 frame)
if not isinstance(spec, EncodedTextSpec) or \
isinstance(spec, TimeStampSpec):
return value
value = [spec._validate23(frame, v, **kwargs) for v in value]
if kwargs.get("sep") is not None:
return [spec.validate(frame, kwargs["sep"].join(value))]
return value
class EncodedNumericTextSpec(EncodedTextSpec):
pass
class EncodedNumericPartTextSpec(EncodedTextSpec):
pass
class Latin1TextSpec(EncodedTextSpec):
def read(self, frame, data):
if b'\x00' in data:
data, ret = data.split(b'\x00', 1)
else:
ret = b''
return data.decode('latin1'), ret
def write(self, data, value):
return value.encode('latin1') + b'\x00'
def validate(self, frame, value):
return text_type(value)
@swap_to_string
@total_ordering
class ID3TimeStamp(object):
"""A time stamp in ID3v2 format.
This is a restricted form of the ISO 8601 standard; time stamps
take the form of:
YYYY-MM-DD HH:MM:SS
Or some partial form (YYYY-MM-DD HH, YYYY, etc.).
The 'text' attribute contains the raw text data of the time stamp.
"""
import re
def __init__(self, text):
if isinstance(text, ID3TimeStamp):
text = text.text
elif not isinstance(text, text_type):
if PY3:
raise TypeError("not a str")
text = text.decode("utf-8")
self.text = text
__formats = ['%04d'] + ['%02d'] * 5
__seps = ['-', '-', ' ', ':', ':', 'x']
def get_text(self):
parts = [self.year, self.month, self.day,
self.hour, self.minute, self.second]
pieces = []
for i, part in enumerate(parts):
if part is None:
break
pieces.append(self.__formats[i] % part + self.__seps[i])
return u''.join(pieces)[:-1]
def set_text(self, text, splitre=re.compile('[-T:/.]|\s+')):
year, month, day, hour, minute, second = \
splitre.split(text + ':::::')[:6]
for a in 'year month day hour minute second'.split():
try:
v = int(locals()[a])
except ValueError:
v = None
setattr(self, a, v)
text = property(get_text, set_text, doc="ID3v2.4 date and time.")
def __str__(self):
return self.text
def __bytes__(self):
return self.text.encode("utf-8")
def __repr__(self):
return repr(self.text)
def __eq__(self, other):
return self.text == other.text
def __lt__(self, other):
return self.text < other.text
__hash__ = object.__hash__
def encode(self, *args):
return self.text.encode(*args)
class TimeStampSpec(EncodedTextSpec):
def read(self, frame, data):
value, data = super(TimeStampSpec, self).read(frame, data)
return self.validate(frame, value), data
def write(self, frame, data):
return super(TimeStampSpec, self).write(frame,
data.text.replace(' ', 'T'))
def validate(self, frame, value):
try:
return ID3TimeStamp(value)
except TypeError:
raise ValueError("Invalid ID3TimeStamp: %r" % value)
class ChannelSpec(ByteSpec):
(OTHER, MASTER, FRONTRIGHT, FRONTLEFT, BACKRIGHT, BACKLEFT, FRONTCENTRE,
BACKCENTRE, SUBWOOFER) = range(9)
class VolumeAdjustmentSpec(Spec):
def read(self, frame, data):
value, = unpack('>h', data[0:2])
return value / 512.0, data[2:]
def write(self, frame, value):
number = int(round(value * 512))
# pack only fails in 2.7, do it manually in 2.6
if not -32768 <= number <= 32767:
raise struct.error
return pack('>h', number)
def validate(self, frame, value):
if value is not None:
try:
self.write(frame, value)
except struct.error:
raise ValueError("out of range")
return value
class VolumePeakSpec(Spec):
def read(self, frame, data):
# http://bugs.xmms.org/attachment.cgi?id=113&action=view
peak = 0
data_array = bytearray(data)
bits = data_array[0]
vol_bytes = min(4, (bits + 7) >> 3)
# not enough frame data
if vol_bytes + 1 > len(data):
raise ID3JunkFrameError
shift = ((8 - (bits & 7)) & 7) + (4 - vol_bytes) * 8
for i in range(1, vol_bytes + 1):
peak *= 256
peak += data_array[i]
peak *= 2 ** shift
return (float(peak) / (2 ** 31 - 1)), data[1 + vol_bytes:]
def write(self, frame, value):
number = int(round(value * 32768))
# pack only fails in 2.7, do it manually in 2.6
if not 0 <= number <= 65535:
raise struct.error
# always write as 16 bits for sanity.
return b"\x10" + pack('>H', number)
def validate(self, frame, value):
if value is not None:
try:
self.write(frame, value)
except struct.error:
raise ValueError("out of range")
return value
class SynchronizedTextSpec(EncodedTextSpec):
def read(self, frame, data):
texts = []
encoding, term = self._encodings[frame.encoding]
while data:
try:
value, data = decode_terminated(data, encoding)
except ValueError:
raise ID3JunkFrameError
if len(data) < 4:
raise ID3JunkFrameError
time, = struct.unpack(">I", data[:4])
texts.append((value, time))
data = data[4:]
return texts, b""
def write(self, frame, value):
data = []
encoding, term = self._encodings[frame.encoding]
for text, time in value:
text = text.encode(encoding) + term
data.append(text + struct.pack(">I", time))
return b"".join(data)
def validate(self, frame, value):
return value
class KeyEventSpec(Spec):
def read(self, frame, data):
events = []
while len(data) >= 5:
events.append(struct.unpack(">bI", data[:5]))
data = data[5:]
return events, data
def write(self, frame, value):
return b"".join(struct.pack(">bI", *event) for event in value)
def validate(self, frame, value):
return value
class VolumeAdjustmentsSpec(Spec):
# Not to be confused with VolumeAdjustmentSpec.
def read(self, frame, data):
adjustments = {}
while len(data) >= 4:
freq, adj = struct.unpack(">Hh", data[:4])
data = data[4:]
freq /= 2.0
adj /= 512.0
adjustments[freq] = adj
adjustments = sorted(adjustments.items())
return adjustments, data
def write(self, frame, value):
value.sort()
return b"".join(struct.pack(">Hh", int(freq * 2), int(adj * 512))
for (freq, adj) in value)
def validate(self, frame, value):
return value
class ASPIIndexSpec(Spec):
def read(self, frame, data):
if frame.b == 16:
format = "H"
size = 2
elif frame.b == 8:
format = "B"
size = 1
else:
warn("invalid bit count in ASPI (%d)" % frame.b, ID3Warning)
return [], data
indexes = data[:frame.N * size]
data = data[frame.N * size:]
return list(struct.unpack(">" + format * frame.N, indexes)), data
def write(self, frame, values):
if frame.b == 16:
format = "H"
elif frame.b == 8:
format = "B"
else:
raise ValueError("frame.b must be 8 or 16")
return struct.pack(">" + format * frame.N, *values)
def validate(self, frame, values):
return values

165
lib/mutagen/id3/_util.py Normal file
View File

@@ -0,0 +1,165 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2005 Michael Urman
# 2013 Christoph Reiter
# 2014 Ben Ockmore
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
from .._compat import long_, integer_types
from .._util import MutagenError
class error(MutagenError):
pass
class ID3NoHeaderError(error, ValueError):
pass
class ID3BadUnsynchData(error, ValueError):
pass
class ID3BadCompressedData(error, ValueError):
pass
class ID3TagError(error, ValueError):
pass
class ID3UnsupportedVersionError(error, NotImplementedError):
pass
class ID3EncryptionUnsupportedError(error, NotImplementedError):
pass
class ID3JunkFrameError(error, ValueError):
pass
class ID3Warning(error, UserWarning):
pass
class unsynch(object):
@staticmethod
def decode(value):
fragments = bytearray(value).split(b'\xff')
if len(fragments) > 1 and not fragments[-1]:
raise ValueError('string ended unsafe')
for f in fragments[1:]:
if (not f) or (f[0] >= 0xE0):
raise ValueError('invalid sync-safe string')
if f[0] == 0x00:
del f[0]
return bytes(bytearray(b'\xff').join(fragments))
@staticmethod
def encode(value):
fragments = bytearray(value).split(b'\xff')
for f in fragments[1:]:
if (not f) or (f[0] >= 0xE0) or (f[0] == 0x00):
f.insert(0, 0x00)
return bytes(bytearray(b'\xff').join(fragments))
class _BitPaddedMixin(object):
def as_str(self, width=4, minwidth=4):
return self.to_str(self, self.bits, self.bigendian, width, minwidth)
@staticmethod
def to_str(value, bits=7, bigendian=True, width=4, minwidth=4):
mask = (1 << bits) - 1
if width != -1:
index = 0
bytes_ = bytearray(width)
try:
while value:
bytes_[index] = value & mask
value >>= bits
index += 1
except IndexError:
raise ValueError('Value too wide (>%d bytes)' % width)
else:
# PCNT and POPM use growing integers
# of at least 4 bytes (=minwidth) as counters.
bytes_ = bytearray()
append = bytes_.append
while value:
append(value & mask)
value >>= bits
bytes_ = bytes_.ljust(minwidth, b"\x00")
if bigendian:
bytes_.reverse()
return bytes(bytes_)
@staticmethod
def has_valid_padding(value, bits=7):
"""Whether the padding bits are all zero"""
assert bits <= 8
mask = (((1 << (8 - bits)) - 1) << bits)
if isinstance(value, integer_types):
while value:
if value & mask:
return False
value >>= 8
elif isinstance(value, bytes):
for byte in bytearray(value):
if byte & mask:
return False
else:
raise TypeError
return True
class BitPaddedInt(int, _BitPaddedMixin):
def __new__(cls, value, bits=7, bigendian=True):
mask = (1 << (bits)) - 1
numeric_value = 0
shift = 0
if isinstance(value, integer_types):
while value:
numeric_value += (value & mask) << shift
value >>= 8
shift += bits
elif isinstance(value, bytes):
if bigendian:
value = reversed(value)
for byte in bytearray(value):
numeric_value += (byte & mask) << shift
shift += bits
else:
raise TypeError
if isinstance(numeric_value, int):
self = int.__new__(BitPaddedInt, numeric_value)
else:
self = long_.__new__(BitPaddedLong, numeric_value)
self.bits = bits
self.bigendian = bigendian
return self
class BitPaddedLong(long_, _BitPaddedMixin):
pass

545
lib/mutagen/m4a.py Normal file
View File

@@ -0,0 +1,545 @@
# Copyright 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
import sys
if sys.version_info[0] != 2:
raise ImportError("No longer available with Python 3, use mutagen.mp4")
"""Read and write MPEG-4 audio files with iTunes metadata.
This module will read MPEG-4 audio information and metadata,
as found in Apple's M4A (aka MP4, M4B, M4P) files.
There is no official specification for this format. The source code
for TagLib, FAAD, and various MPEG specifications at
http://developer.apple.com/documentation/QuickTime/QTFF/,
http://www.geocities.com/xhelmboyx/quicktime/formats/mp4-layout.txt,
and http://wiki.multimedia.cx/index.php?title=Apple_QuickTime were all
consulted.
This module does not support 64 bit atom sizes, and so will not
work on metadata over 4GB.
"""
import struct
import sys
from cStringIO import StringIO
from ._compat import reraise
from mutagen import FileType, Metadata, StreamInfo
from mutagen._constants import GENRES
from mutagen._util import cdata, insert_bytes, delete_bytes, DictProxy, \
MutagenError
class error(IOError, MutagenError):
pass
class M4AMetadataError(error):
pass
class M4AStreamInfoError(error):
pass
class M4AMetadataValueError(ValueError, M4AMetadataError):
pass
import warnings
warnings.warn(
"mutagen.m4a is deprecated; use mutagen.mp4 instead.", DeprecationWarning)
# This is not an exhaustive list of container atoms, but just the
# ones this module needs to peek inside.
_CONTAINERS = ["moov", "udta", "trak", "mdia", "meta", "ilst",
"stbl", "minf", "stsd"]
_SKIP_SIZE = {"meta": 4}
__all__ = ['M4A', 'Open', 'delete', 'M4ACover']
class M4ACover(str):
"""A cover artwork.
Attributes:
imageformat -- format of the image (either FORMAT_JPEG or FORMAT_PNG)
"""
FORMAT_JPEG = 0x0D
FORMAT_PNG = 0x0E
def __new__(cls, data, imageformat=None):
self = str.__new__(cls, data)
if imageformat is None:
imageformat = M4ACover.FORMAT_JPEG
self.imageformat = imageformat
try:
self.format
except AttributeError:
self.format = imageformat
return self
class Atom(object):
"""An individual atom.
Attributes:
children -- list child atoms (or None for non-container atoms)
length -- length of this atom, including length and name
name -- four byte name of the atom, as a str
offset -- location in the constructor-given fileobj of this atom
This structure should only be used internally by Mutagen.
"""
children = None
def __init__(self, fileobj):
self.offset = fileobj.tell()
self.length, self.name = struct.unpack(">I4s", fileobj.read(8))
if self.length == 1:
raise error("64 bit atom sizes are not supported")
elif self.length < 8:
return
if self.name in _CONTAINERS:
self.children = []
fileobj.seek(_SKIP_SIZE.get(self.name, 0), 1)
while fileobj.tell() < self.offset + self.length:
self.children.append(Atom(fileobj))
else:
fileobj.seek(self.offset + self.length, 0)
@staticmethod
def render(name, data):
"""Render raw atom data."""
# this raises OverflowError if Py_ssize_t can't handle the atom data
size = len(data) + 8
if size <= 0xFFFFFFFF:
return struct.pack(">I4s", size, name) + data
else:
return struct.pack(">I4sQ", 1, name, size + 8) + data
def __getitem__(self, remaining):
"""Look up a child atom, potentially recursively.
e.g. atom['udta', 'meta'] => <Atom name='meta' ...>
"""
if not remaining:
return self
elif self.children is None:
raise KeyError("%r is not a container" % self.name)
for child in self.children:
if child.name == remaining[0]:
return child[remaining[1:]]
else:
raise KeyError("%r not found" % remaining[0])
def __repr__(self):
klass = self.__class__.__name__
if self.children is None:
return "<%s name=%r length=%r offset=%r>" % (
klass, self.name, self.length, self.offset)
else:
children = "\n".join([" " + line for child in self.children
for line in repr(child).splitlines()])
return "<%s name=%r length=%r offset=%r\n%s>" % (
klass, self.name, self.length, self.offset, children)
class Atoms(object):
"""Root atoms in a given file.
Attributes:
atoms -- a list of top-level atoms as Atom objects
This structure should only be used internally by Mutagen.
"""
def __init__(self, fileobj):
self.atoms = []
fileobj.seek(0, 2)
end = fileobj.tell()
fileobj.seek(0)
while fileobj.tell() < end:
self.atoms.append(Atom(fileobj))
def path(self, *names):
"""Look up and return the complete path of an atom.
For example, atoms.path('moov', 'udta', 'meta') will return a
list of three atoms, corresponding to the moov, udta, and meta
atoms.
"""
path = [self]
for name in names:
path.append(path[-1][name, ])
return path[1:]
def __getitem__(self, names):
"""Look up a child atom.
'names' may be a list of atoms (['moov', 'udta']) or a string
specifying the complete path ('moov.udta').
"""
if isinstance(names, basestring):
names = names.split(".")
for child in self.atoms:
if child.name == names[0]:
return child[names[1:]]
else:
raise KeyError("%s not found" % names[0])
def __repr__(self):
return "\n".join([repr(child) for child in self.atoms])
class M4ATags(DictProxy, Metadata):
"""Dictionary containing Apple iTunes metadata list key/values.
Keys are four byte identifiers, except for freeform ('----')
keys. Values are usually unicode strings, but some atoms have a
special structure:
cpil -- boolean
trkn, disk -- tuple of 16 bit ints (current, total)
tmpo -- 16 bit int
covr -- list of M4ACover objects (which are tagged strs)
gnre -- not supported. Use '\\xa9gen' instead.
The freeform '----' frames use a key in the format '----:mean:name'
where 'mean' is usually 'com.apple.iTunes' and 'name' is a unique
identifier for this frame. The value is a str, but is probably
text that can be decoded as UTF-8.
M4A tag data cannot exist outside of the structure of an M4A file,
so this class should not be manually instantiated.
Unknown non-text tags are removed.
"""
def load(self, atoms, fileobj):
try:
ilst = atoms["moov.udta.meta.ilst"]
except KeyError as key:
raise M4AMetadataError(key)
for atom in ilst.children:
fileobj.seek(atom.offset + 8)
data = fileobj.read(atom.length - 8)
parse = self.__atoms.get(atom.name, (M4ATags.__parse_text,))[0]
parse(self, atom, data)
@staticmethod
def __key_sort(item1, item2):
(key1, v1) = item1
(key2, v2) = item2
# iTunes always writes the tags in order of "relevance", try
# to copy it as closely as possible.
order = ["\xa9nam", "\xa9ART", "\xa9wrt", "\xa9alb",
"\xa9gen", "gnre", "trkn", "disk",
"\xa9day", "cpil", "tmpo", "\xa9too",
"----", "covr", "\xa9lyr"]
order = dict(zip(order, range(len(order))))
last = len(order)
# If there's no key-based way to distinguish, order by length.
# If there's still no way, go by string comparison on the
# values, so we at least have something determinstic.
return (cmp(order.get(key1[:4], last), order.get(key2[:4], last)) or
cmp(len(v1), len(v2)) or cmp(v1, v2))
def save(self, filename):
"""Save the metadata to the given filename."""
values = []
items = self.items()
items.sort(self.__key_sort)
for key, value in items:
render = self.__atoms.get(
key[:4], (None, M4ATags.__render_text))[1]
values.append(render(self, key, value))
data = Atom.render("ilst", "".join(values))
# Find the old atoms.
fileobj = open(filename, "rb+")
try:
atoms = Atoms(fileobj)
moov = atoms["moov"]
if moov != atoms.atoms[-1]:
# "Free" the old moov block. Something in the mdat
# block is not happy when its offset changes and it
# won't play back. So, rather than try to figure that
# out, just move the moov atom to the end of the file.
offset = self.__move_moov(fileobj, moov)
else:
offset = 0
try:
path = atoms.path("moov", "udta", "meta", "ilst")
except KeyError:
self.__save_new(fileobj, atoms, data, offset)
else:
self.__save_existing(fileobj, atoms, path, data, offset)
finally:
fileobj.close()
def __move_moov(self, fileobj, moov):
fileobj.seek(moov.offset)
data = fileobj.read(moov.length)
fileobj.seek(moov.offset)
free = Atom.render("free", "\x00" * (moov.length - 8))
fileobj.write(free)
fileobj.seek(0, 2)
# Figure out how far we have to shift all our successive
# seek calls, relative to what the atoms say.
old_end = fileobj.tell()
fileobj.write(data)
return old_end - moov.offset
def __save_new(self, fileobj, atoms, ilst, offset):
hdlr = Atom.render("hdlr", "\x00" * 8 + "mdirappl" + "\x00" * 9)
meta = Atom.render("meta", "\x00\x00\x00\x00" + hdlr + ilst)
moov, udta = atoms.path("moov", "udta")
insert_bytes(fileobj, len(meta), udta.offset + offset + 8)
fileobj.seek(udta.offset + offset + 8)
fileobj.write(meta)
self.__update_parents(fileobj, [moov, udta], len(meta), offset)
def __save_existing(self, fileobj, atoms, path, data, offset):
# Replace the old ilst atom.
ilst = path.pop()
delta = len(data) - ilst.length
fileobj.seek(ilst.offset + offset)
if delta > 0:
insert_bytes(fileobj, delta, ilst.offset + offset)
elif delta < 0:
delete_bytes(fileobj, -delta, ilst.offset + offset)
fileobj.seek(ilst.offset + offset)
fileobj.write(data)
self.__update_parents(fileobj, path, delta, offset)
def __update_parents(self, fileobj, path, delta, offset):
# Update all parent atoms with the new size.
for atom in path:
fileobj.seek(atom.offset + offset)
size = cdata.uint_be(fileobj.read(4)) + delta
fileobj.seek(atom.offset + offset)
fileobj.write(cdata.to_uint_be(size))
def __render_data(self, key, flags, data):
data = struct.pack(">2I", flags, 0) + data
return Atom.render(key, Atom.render("data", data))
def __parse_freeform(self, atom, data):
try:
fileobj = StringIO(data)
mean_length = cdata.uint_be(fileobj.read(4))
# skip over 8 bytes of atom name, flags
mean = fileobj.read(mean_length - 4)[8:]
name_length = cdata.uint_be(fileobj.read(4))
name = fileobj.read(name_length - 4)[8:]
value_length = cdata.uint_be(fileobj.read(4))
# Name, flags, and reserved bytes
value = fileobj.read(value_length - 4)[12:]
except struct.error:
# Some ---- atoms have no data atom, I have no clue why
# they actually end up in the file.
pass
else:
self["%s:%s:%s" % (atom.name, mean, name)] = value
def __render_freeform(self, key, value):
dummy, mean, name = key.split(":", 2)
mean = struct.pack(">I4sI", len(mean) + 12, "mean", 0) + mean
name = struct.pack(">I4sI", len(name) + 12, "name", 0) + name
value = struct.pack(">I4s2I", len(value) + 16, "data", 0x1, 0) + value
final = mean + name + value
return Atom.render("----", final)
def __parse_pair(self, atom, data):
self[atom.name] = struct.unpack(">2H", data[18:22])
def __render_pair(self, key, value):
track, total = value
if 0 <= track < 1 << 16 and 0 <= total < 1 << 16:
data = struct.pack(">4H", 0, track, total, 0)
return self.__render_data(key, 0, data)
else:
raise M4AMetadataValueError("invalid numeric pair %r" % (value,))
def __render_pair_no_trailing(self, key, value):
track, total = value
if 0 <= track < 1 << 16 and 0 <= total < 1 << 16:
data = struct.pack(">3H", 0, track, total)
return self.__render_data(key, 0, data)
else:
raise M4AMetadataValueError("invalid numeric pair %r" % (value,))
def __parse_genre(self, atom, data):
# Translate to a freeform genre.
genre = cdata.short_be(data[16:18])
if "\xa9gen" not in self:
try:
self["\xa9gen"] = GENRES[genre - 1]
except IndexError:
pass
def __parse_tempo(self, atom, data):
self[atom.name] = cdata.short_be(data[16:18])
def __render_tempo(self, key, value):
if 0 <= value < 1 << 16:
return self.__render_data(key, 0x15, cdata.to_ushort_be(value))
else:
raise M4AMetadataValueError("invalid short integer %r" % value)
def __parse_compilation(self, atom, data):
try:
self[atom.name] = bool(ord(data[16:17]))
except TypeError:
self[atom.name] = False
def __render_compilation(self, key, value):
return self.__render_data(key, 0x15, chr(bool(value)))
def __parse_cover(self, atom, data):
length, name, imageformat = struct.unpack(">I4sI", data[:12])
if name != "data":
raise M4AMetadataError(
"unexpected atom %r inside 'covr'" % name)
if imageformat not in (M4ACover.FORMAT_JPEG, M4ACover.FORMAT_PNG):
imageformat = M4ACover.FORMAT_JPEG
self[atom.name] = M4ACover(data[16:length], imageformat)
def __render_cover(self, key, value):
try:
imageformat = value.imageformat
except AttributeError:
imageformat = M4ACover.FORMAT_JPEG
data = Atom.render("data", struct.pack(">2I", imageformat, 0) + value)
return Atom.render(key, data)
def __parse_text(self, atom, data):
flags = cdata.uint_be(data[8:12])
if flags == 1:
self[atom.name] = data[16:].decode('utf-8', 'replace')
def __render_text(self, key, value):
return self.__render_data(key, 0x1, value.encode('utf-8'))
def delete(self, filename):
self.clear()
self.save(filename)
__atoms = {
"----": (__parse_freeform, __render_freeform),
"trkn": (__parse_pair, __render_pair),
"disk": (__parse_pair, __render_pair_no_trailing),
"gnre": (__parse_genre, None),
"tmpo": (__parse_tempo, __render_tempo),
"cpil": (__parse_compilation, __render_compilation),
"covr": (__parse_cover, __render_cover),
}
def pprint(self):
values = []
for key, value in self.iteritems():
key = key.decode('latin1')
try:
values.append("%s=%s" % (key, value))
except UnicodeDecodeError:
values.append("%s=[%d bytes of data]" % (key, len(value)))
return "\n".join(values)
class M4AInfo(StreamInfo):
"""MPEG-4 stream information.
Attributes:
bitrate -- bitrate in bits per second, as an int
length -- file length in seconds, as a float
"""
bitrate = 0
def __init__(self, atoms, fileobj):
hdlr = atoms["moov.trak.mdia.hdlr"]
fileobj.seek(hdlr.offset)
if "soun" not in fileobj.read(hdlr.length):
raise M4AStreamInfoError("track has no audio data")
mdhd = atoms["moov.trak.mdia.mdhd"]
fileobj.seek(mdhd.offset)
data = fileobj.read(mdhd.length)
if ord(data[8]) == 0:
offset = 20
fmt = ">2I"
else:
offset = 28
fmt = ">IQ"
end = offset + struct.calcsize(fmt)
unit, length = struct.unpack(fmt, data[offset:end])
self.length = float(length) / unit
try:
atom = atoms["moov.trak.mdia.minf.stbl.stsd"]
fileobj.seek(atom.offset)
data = fileobj.read(atom.length)
self.bitrate = cdata.uint_be(data[-17:-13])
except (ValueError, KeyError):
# Bitrate values are optional.
pass
def pprint(self):
return "MPEG-4 audio, %.2f seconds, %d bps" % (
self.length, self.bitrate)
class M4A(FileType):
"""An MPEG-4 audio file, probably containing AAC.
If more than one track is present in the file, the first is used.
Only audio ('soun') tracks will be read.
"""
_mimes = ["audio/mp4", "audio/x-m4a", "audio/mpeg4", "audio/aac"]
def load(self, filename):
self.filename = filename
fileobj = open(filename, "rb")
try:
atoms = Atoms(fileobj)
try:
self.info = M4AInfo(atoms, fileobj)
except StandardError as err:
reraise(M4AStreamInfoError, err, sys.exc_info()[2])
try:
self.tags = M4ATags(atoms, fileobj)
except M4AMetadataError:
self.tags = None
except StandardError as err:
reraise(M4AMetadataError, err, sys.exc_info()[2])
finally:
fileobj.close()
def add_tags(self):
self.tags = M4ATags()
@staticmethod
def score(filename, fileobj, header):
return ("ftyp" in header) + ("mp4" in header)
Open = M4A
def delete(filename):
"""Remove tags from a file."""
M4A(filename).delete()

86
lib/mutagen/monkeysaudio.py Normal file
View File

@@ -0,0 +1,86 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Lukas Lalinsky
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Monkey's Audio streams with APEv2 tags.
Monkey's Audio is a very efficient lossless audio compressor developed
by Matt Ashland.
For more information, see http://www.monkeysaudio.com/.
"""
__all__ = ["MonkeysAudio", "Open", "delete"]
import struct
from ._compat import endswith
from mutagen import StreamInfo
from mutagen.apev2 import APEv2File, error, delete
from mutagen._util import cdata
class MonkeysAudioHeaderError(error):
pass
class MonkeysAudioInfo(StreamInfo):
"""Monkey's Audio stream information.
Attributes:
* channels -- number of audio channels
* length -- file length in seconds, as a float
* sample_rate -- audio sampling rate in Hz
* bits_per_sample -- bits per sample
* version -- Monkey's Audio stream version, as a float (eg: 3.99)
"""
def __init__(self, fileobj):
header = fileobj.read(76)
if len(header) != 76 or not header.startswith(b"MAC "):
raise MonkeysAudioHeaderError("not a Monkey's Audio file")
self.version = cdata.ushort_le(header[4:6])
if self.version >= 3980:
(blocks_per_frame, final_frame_blocks, total_frames,
self.bits_per_sample, self.channels,
self.sample_rate) = struct.unpack("<IIIHHI", header[56:76])
else:
compression_level = cdata.ushort_le(header[6:8])
self.channels, self.sample_rate = struct.unpack(
"<HI", header[10:16])
total_frames, final_frame_blocks = struct.unpack(
"<II", header[24:32])
if self.version >= 3950:
blocks_per_frame = 73728 * 4
elif self.version >= 3900 or (self.version >= 3800 and
compression_level == 4):
blocks_per_frame = 73728
else:
blocks_per_frame = 9216
self.version /= 1000.0
self.length = 0.0
if (self.sample_rate != 0) and (total_frames > 0):
total_blocks = ((total_frames - 1) * blocks_per_frame +
final_frame_blocks)
self.length = float(total_blocks) / self.sample_rate
def pprint(self):
return "Monkey's Audio %.2f, %.2f seconds, %d Hz" % (
self.version, self.length, self.sample_rate)
class MonkeysAudio(APEv2File):
_Info = MonkeysAudioInfo
_mimes = ["audio/ape", "audio/x-ape"]
@staticmethod
def score(filename, fileobj, header):
return header.startswith(b"MAC ") + endswith(filename.lower(), ".ape")
Open = MonkeysAudio

289
lib/mutagen/mp3.py Normal file
View File

@@ -0,0 +1,289 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""MPEG audio stream information and tags."""
import os
import struct
from ._compat import endswith
from mutagen import StreamInfo
from mutagen._util import MutagenError
from mutagen.id3 import ID3FileType, BitPaddedInt, delete
__all__ = ["MP3", "Open", "delete", "MP3"]
class error(RuntimeError, MutagenError):
pass
class HeaderNotFoundError(error, IOError):
pass
class InvalidMPEGHeader(error, IOError):
pass
# Mode values.
STEREO, JOINTSTEREO, DUALCHANNEL, MONO = range(4)
class MPEGInfo(StreamInfo):
"""MPEG audio stream information
Parse information about an MPEG audio file. This also reads the
Xing VBR header format.
This code was implemented based on the format documentation at
http://mpgedit.org/mpgedit/mpeg_format/mpeghdr.htm.
Useful attributes:
* length -- audio length, in seconds
* bitrate -- audio bitrate, in bits per second
* sketchy -- if true, the file may not be valid MPEG audio
Useless attributes:
* version -- MPEG version (1, 2, 2.5)
* layer -- 1, 2, or 3
* mode -- One of STEREO, JOINTSTEREO, DUALCHANNEL, or MONO (0-3)
* protected -- whether or not the file is "protected"
* padding -- whether or not audio frames are padded
* sample_rate -- audio sample rate, in Hz
"""
# Map (version, layer) tuples to bitrates.
__BITRATE = {
(1, 1): [0, 32, 64, 96, 128, 160, 192, 224,
256, 288, 320, 352, 384, 416, 448],
(1, 2): [0, 32, 48, 56, 64, 80, 96, 112, 128,
160, 192, 224, 256, 320, 384],
(1, 3): [0, 32, 40, 48, 56, 64, 80, 96, 112,
128, 160, 192, 224, 256, 320],
(2, 1): [0, 32, 48, 56, 64, 80, 96, 112, 128,
144, 160, 176, 192, 224, 256],
(2, 2): [0, 8, 16, 24, 32, 40, 48, 56, 64,
80, 96, 112, 128, 144, 160],
}
__BITRATE[(2, 3)] = __BITRATE[(2, 2)]
for i in range(1, 4):
__BITRATE[(2.5, i)] = __BITRATE[(2, i)]
# Map version to sample rates.
__RATES = {
1: [44100, 48000, 32000],
2: [22050, 24000, 16000],
2.5: [11025, 12000, 8000]
}
sketchy = False
def __init__(self, fileobj, offset=None):
"""Parse MPEG stream information from a file-like object.
If an offset argument is given, it is used to start looking
for stream information and Xing headers; otherwise, ID3v2 tags
will be skipped automatically. A correct offset can make
loading files significantly faster.
"""
try:
size = os.path.getsize(fileobj.name)
except (IOError, OSError, AttributeError):
fileobj.seek(0, 2)
size = fileobj.tell()
# If we don't get an offset, try to skip an ID3v2 tag.
if offset is None:
fileobj.seek(0, 0)
idata = fileobj.read(10)
try:
id3, insize = struct.unpack('>3sxxx4s', idata)
except struct.error:
id3, insize = '', 0
insize = BitPaddedInt(insize)
if id3 == b'ID3' and insize > 0:
offset = insize + 10
else:
offset = 0
# Try to find two valid headers (meaning, very likely MPEG data)
# at the given offset, 30% through the file, 60% through the file,
# and 90% through the file.
for i in [offset, 0.3 * size, 0.6 * size, 0.9 * size]:
try:
self.__try(fileobj, int(i), size - offset)
except error:
pass
else:
break
# If we can't find any two consecutive frames, try to find just
# one frame back at the original offset given.
else:
self.__try(fileobj, offset, size - offset, False)
self.sketchy = True
def __try(self, fileobj, offset, real_size, check_second=True):
# This is going to be one really long function; bear with it,
# because there's not really a sane point to cut it up.
fileobj.seek(offset, 0)
# We "know" we have an MPEG file if we find two frames that look like
# valid MPEG data. If we can't find them in 32k of reads, something
# is horribly wrong (the longest frame can only be about 4k). This
# is assuming the offset didn't lie.
data = fileobj.read(32768)
frame_1 = data.find(b"\xff")
while 0 <= frame_1 <= (len(data) - 4):
frame_data = struct.unpack(">I", data[frame_1:frame_1 + 4])[0]
if ((frame_data >> 16) & 0xE0) != 0xE0:
frame_1 = data.find(b"\xff", frame_1 + 2)
else:
version = (frame_data >> 19) & 0x3
layer = (frame_data >> 17) & 0x3
protection = (frame_data >> 16) & 0x1
bitrate = (frame_data >> 12) & 0xF
sample_rate = (frame_data >> 10) & 0x3
padding = (frame_data >> 9) & 0x1
# private = (frame_data >> 8) & 0x1
self.mode = (frame_data >> 6) & 0x3
# mode_extension = (frame_data >> 4) & 0x3
# copyright = (frame_data >> 3) & 0x1
# original = (frame_data >> 2) & 0x1
# emphasis = (frame_data >> 0) & 0x3
if (version == 1 or layer == 0 or sample_rate == 0x3 or
bitrate == 0 or bitrate == 0xF):
frame_1 = data.find(b"\xff", frame_1 + 2)
else:
break
else:
raise HeaderNotFoundError("can't sync to an MPEG frame")
# There is a serious problem here, which is that many flags
# in an MPEG header are backwards.
self.version = [2.5, None, 2, 1][version]
self.layer = 4 - layer
self.protected = not protection
self.padding = bool(padding)
self.bitrate = self.__BITRATE[(self.version, self.layer)][bitrate]
self.bitrate *= 1000
self.sample_rate = self.__RATES[self.version][sample_rate]
if self.layer == 1:
frame_length = (
(12 * self.bitrate // self.sample_rate) + padding) * 4
frame_size = 384
elif self.version >= 2 and self.layer == 3:
frame_length = (72 * self.bitrate // self.sample_rate) + padding
frame_size = 576
else:
frame_length = (144 * self.bitrate // self.sample_rate) + padding
frame_size = 1152
if check_second:
possible = int(frame_1 + frame_length)
if possible > len(data) + 4:
raise HeaderNotFoundError("can't sync to second MPEG frame")
try:
frame_data = struct.unpack(
">H", data[possible:possible + 2])[0]
except struct.error:
raise HeaderNotFoundError("can't sync to second MPEG frame")
if (frame_data & 0xFFE0) != 0xFFE0:
raise HeaderNotFoundError("can't sync to second MPEG frame")
self.length = 8 * real_size / float(self.bitrate)
# Try to find/parse the Xing header, which trumps the above length
# and bitrate calculation.
fileobj.seek(offset, 0)
data = fileobj.read(32768)
try:
xing = data[:-4].index(b"Xing")
except ValueError:
# Try to find/parse the VBRI header, which trumps the above length
# calculation.
try:
vbri = data[:-24].index(b"VBRI")
except ValueError:
pass
else:
# If a VBRI header was found, this is definitely MPEG audio.
self.sketchy = False
vbri_version = struct.unpack('>H', data[vbri + 4:vbri + 6])[0]
if vbri_version == 1:
frame_count = struct.unpack(
'>I', data[vbri + 14:vbri + 18])[0]
samples = float(frame_size * frame_count)
self.length = (samples / self.sample_rate) or self.length
else:
# If a Xing header was found, this is definitely MPEG audio.
self.sketchy = False
flags = struct.unpack('>I', data[xing + 4:xing + 8])[0]
if flags & 0x1:
frame_count = struct.unpack('>I', data[xing + 8:xing + 12])[0]
samples = float(frame_size * frame_count)
self.length = (samples / self.sample_rate) or self.length
if flags & 0x2:
bitrate_data = struct.unpack(
'>I', data[xing + 12:xing + 16])[0]
self.bitrate = int((bitrate_data * 8) // self.length)
def pprint(self):
s = "MPEG %s layer %d, %d bps, %s Hz, %.2f seconds" % (
self.version, self.layer, self.bitrate, self.sample_rate,
self.length)
if self.sketchy:
s += " (sketchy)"
return s
class MP3(ID3FileType):
"""An MPEG audio (usually MPEG-1 Layer 3) file.
:ivar info: :class:`MPEGInfo`
:ivar tags: :class:`ID3 <mutagen.id3.ID3>`
"""
_Info = MPEGInfo
_mimes = ["audio/mpeg", "audio/mpg", "audio/x-mpeg"]
@property
def mime(self):
l = self.info.layer
return ["audio/mp%d" % l, "audio/x-mp%d" % l] + super(MP3, self).mime
@staticmethod
def score(filename, fileobj, header_data):
filename = filename.lower()
return (header_data.startswith(b"ID3") * 2 +
endswith(filename, b".mp3") +
endswith(filename, b".mp2") + endswith(filename, b".mpg") +
endswith(filename, b".mpeg"))
Open = MP3
class EasyMP3(MP3):
"""Like MP3, but uses EasyID3 for tags.
:ivar info: :class:`MPEGInfo`
:ivar tags: :class:`EasyID3 <mutagen.easyid3.EasyID3>`
"""
from mutagen.easyid3 import EasyID3 as ID3
ID3 = ID3

965
lib/mutagen/mp4/__init__.py Normal file
View File

@@ -0,0 +1,965 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write MPEG-4 audio files with iTunes metadata.
This module will read MPEG-4 audio information and metadata,
as found in Apple's MP4 (aka M4A, M4B, M4P) files.
There is no official specification for this format. The source code
for TagLib, FAAD, and various MPEG specifications at
* http://developer.apple.com/documentation/QuickTime/QTFF/
* http://www.geocities.com/xhelmboyx/quicktime/formats/mp4-layout.txt
* http://standards.iso.org/ittf/PubliclyAvailableStandards/\
c041828_ISO_IEC_14496-12_2005(E).zip
* http://wiki.multimedia.cx/index.php?title=Apple_QuickTime
were all consulted.
"""
import struct
import sys
from mutagen import FileType, Metadata, StreamInfo
from mutagen._constants import GENRES
from mutagen._util import (cdata, insert_bytes, DictProxy, MutagenError,
hashable, enum)
from mutagen._compat import (reraise, PY2, string_types, text_type, chr_,
iteritems, PY3, cBytesIO)
from ._atom import Atoms, Atom, AtomError
from ._util import parse_full_atom
from ._as_entry import AudioSampleEntry, ASEntryError
class error(IOError, MutagenError):
pass
class MP4MetadataError(error):
pass
class MP4StreamInfoError(error):
pass
class MP4MetadataValueError(ValueError, MP4MetadataError):
pass
__all__ = ['MP4', 'Open', 'delete', 'MP4Cover', 'MP4FreeForm', 'AtomDataType']
@enum
class AtomDataType(object):
"""Enum for `dataformat` attribute of MP4FreeForm.
.. versionadded:: 1.25
"""
IMPLICIT = 0
"""for use with tags for which no type needs to be indicated because
only one type is allowed"""
UTF8 = 1
"""without any count or null terminator"""
UTF16 = 2
"""also known as UTF-16BE"""
SJIS = 3
"""deprecated unless it is needed for special Japanese characters"""
HTML = 6
"""the HTML file header specifies which HTML version"""
XML = 7
"""the XML header must identify the DTD or schemas"""
UUID = 8
"""also known as GUID; stored as 16 bytes in binary (valid as an ID)"""
ISRC = 9
"""stored as UTF-8 text (valid as an ID)"""
MI3P = 10
"""stored as UTF-8 text (valid as an ID)"""
GIF = 12
"""(deprecated) a GIF image"""
JPEG = 13
"""a JPEG image"""
PNG = 14
"""PNG image"""
URL = 15
"""absolute, in UTF-8 characters"""
DURATION = 16
"""in milliseconds, 32-bit integer"""
DATETIME = 17
"""in UTC, counting seconds since midnight, January 1, 1904;
32 or 64-bits"""
GENRES = 18
"""a list of enumerated values"""
INTEGER = 21
"""a signed big-endian integer with length one of { 1,2,3,4,8 } bytes"""
RIAA_PA = 24
"""RIAA parental advisory; { -1=no, 1=yes, 0=unspecified },
8-bit ingteger"""
UPC = 25
"""Universal Product Code, in text UTF-8 format (valid as an ID)"""
BMP = 27
"""Windows bitmap image"""
@hashable
class MP4Cover(bytes):
"""A cover artwork.
Attributes:
* imageformat -- format of the image (either FORMAT_JPEG or FORMAT_PNG)
"""
FORMAT_JPEG = AtomDataType.JPEG
FORMAT_PNG = AtomDataType.PNG
def __new__(cls, data, *args, **kwargs):
return bytes.__new__(cls, data)
def __init__(self, data, imageformat=FORMAT_JPEG):
self.imageformat = imageformat
__hash__ = bytes.__hash__
def __eq__(self, other):
if not isinstance(other, MP4Cover):
return NotImplemented
if not bytes.__eq__(self, other):
return False
if self.imageformat != other.imageformat:
return False
return True
def __ne__(self, other):
return not self.__eq__(other)
def __repr__(self):
return "%s(%r, %r)" % (
type(self).__name__, bytes(self),
AtomDataType(self.imageformat))
@hashable
class MP4FreeForm(bytes):
"""A freeform value.
Attributes:
* dataformat -- format of the data (see AtomDataType)
"""
FORMAT_DATA = AtomDataType.IMPLICIT # deprecated
FORMAT_TEXT = AtomDataType.UTF8 # deprecated
def __new__(cls, data, *args, **kwargs):
return bytes.__new__(cls, data)
def __init__(self, data, dataformat=AtomDataType.UTF8, version=0):
self.dataformat = dataformat
self.version = version
__hash__ = bytes.__hash__
def __eq__(self, other):
if not isinstance(other, MP4FreeForm):
return NotImplemented
if not bytes.__eq__(self, other):
return False
if self.dataformat != other.dataformat:
return False
if self.version != other.version:
return False
return True
def __ne__(self, other):
return not self.__eq__(other)
def __repr__(self):
return "%s(%r, %r)" % (
type(self).__name__, bytes(self),
AtomDataType(self.dataformat))
def _name2key(name):
if PY2:
return name
return name.decode("latin-1")
def _key2name(key):
if PY2:
return key
return key.encode("latin-1")
class MP4Tags(DictProxy, Metadata):
r"""Dictionary containing Apple iTunes metadata list key/values.
Keys are four byte identifiers, except for freeform ('----')
keys. Values are usually unicode strings, but some atoms have a
special structure:
Text values (multiple values per key are supported):
* '\\xa9nam' -- track title
* '\\xa9alb' -- album
* '\\xa9ART' -- artist
* 'aART' -- album artist
* '\\xa9wrt' -- composer
* '\\xa9day' -- year
* '\\xa9cmt' -- comment
* 'desc' -- description (usually used in podcasts)
* 'purd' -- purchase date
* '\\xa9grp' -- grouping
* '\\xa9gen' -- genre
* '\\xa9lyr' -- lyrics
* 'purl' -- podcast URL
* 'egid' -- podcast episode GUID
* 'catg' -- podcast category
* 'keyw' -- podcast keywords
* '\\xa9too' -- encoded by
* 'cprt' -- copyright
* 'soal' -- album sort order
* 'soaa' -- album artist sort order
* 'soar' -- artist sort order
* 'sonm' -- title sort order
* 'soco' -- composer sort order
* 'sosn' -- show sort order
* 'tvsh' -- show name
Boolean values:
* 'cpil' -- part of a compilation
* 'pgap' -- part of a gapless album
* 'pcst' -- podcast (iTunes reads this only on import)
Tuples of ints (multiple values per key are supported):
* 'trkn' -- track number, total tracks
* 'disk' -- disc number, total discs
Others:
* 'tmpo' -- tempo/BPM, 16 bit int
* 'covr' -- cover artwork, list of MP4Cover objects (which are
tagged strs)
* 'gnre' -- ID3v1 genre. Not supported, use '\\xa9gen' instead.
The freeform '----' frames use a key in the format '----:mean:name'
where 'mean' is usually 'com.apple.iTunes' and 'name' is a unique
identifier for this frame. The value is a str, but is probably
text that can be decoded as UTF-8. Multiple values per key are
supported.
MP4 tag data cannot exist outside of the structure of an MP4 file,
so this class should not be manually instantiated.
Unknown non-text tags and tags that failed to parse will be written
back as is.
"""
def __init__(self, *args, **kwargs):
self._failed_atoms = {}
super(MP4Tags, self).__init__(*args, **kwargs)
def load(self, atoms, fileobj):
try:
ilst = atoms[b"moov.udta.meta.ilst"]
except KeyError as key:
raise MP4MetadataError(key)
for atom in ilst.children:
ok, data = atom.read(fileobj)
if not ok:
raise MP4MetadataError("Not enough data")
try:
if atom.name in self.__atoms:
info = self.__atoms[atom.name]
info[0](self, atom, data)
else:
# unknown atom, try as text
self.__parse_text(atom, data, implicit=False)
except MP4MetadataError:
# parsing failed, save them so we can write them back
key = _name2key(atom.name)
self._failed_atoms.setdefault(key, []).append(data)
def __setitem__(self, key, value):
if not isinstance(key, str):
raise TypeError("key has to be str")
super(MP4Tags, self).__setitem__(key, value)
@classmethod
def _can_load(cls, atoms):
return b"moov.udta.meta.ilst" in atoms
@staticmethod
def __key_sort(item):
(key, v) = item
# iTunes always writes the tags in order of "relevance", try
# to copy it as closely as possible.
order = [b"\xa9nam", b"\xa9ART", b"\xa9wrt", b"\xa9alb",
b"\xa9gen", b"gnre", b"trkn", b"disk",
b"\xa9day", b"cpil", b"pgap", b"pcst", b"tmpo",
b"\xa9too", b"----", b"covr", b"\xa9lyr"]
order = dict(zip(order, range(len(order))))
last = len(order)
# If there's no key-based way to distinguish, order by length.
# If there's still no way, go by string comparison on the
# values, so we at least have something determinstic.
return (order.get(key[:4], last), len(repr(v)), repr(v))
def save(self, filename):
"""Save the metadata to the given filename."""
values = []
items = sorted(self.items(), key=self.__key_sort)
for key, value in items:
atom_name = _key2name(key)[:4]
if atom_name in self.__atoms:
render_func = self.__atoms[atom_name][1]
else:
render_func = type(self).__render_text
try:
values.append(render_func(self, key, value))
except (TypeError, ValueError) as s:
reraise(MP4MetadataValueError, s, sys.exc_info()[2])
for atom_name, failed in iteritems(self._failed_atoms):
# don't write atoms back if we have added a new one with
# the same name, this excludes freeform which can have
# multiple atoms with the same key (most parsers seem to be able
# to handle that)
if atom_name in self:
assert atom_name != b"----"
continue
for data in failed:
values.append(Atom.render(_key2name(atom_name), data))
data = Atom.render(b"ilst", b"".join(values))
# Find the old atoms.
with open(filename, "rb+") as fileobj:
try:
atoms = Atoms(fileobj)
except AtomError as err:
reraise(error, err, sys.exc_info()[2])
try:
path = atoms.path(b"moov", b"udta", b"meta", b"ilst")
except KeyError:
self.__save_new(fileobj, atoms, data)
else:
self.__save_existing(fileobj, atoms, path, data)
def __pad_ilst(self, data, length=None):
if length is None:
length = ((len(data) + 1023) & ~1023) - len(data)
return Atom.render(b"free", b"\x00" * length)
def __save_new(self, fileobj, atoms, ilst):
hdlr = Atom.render(b"hdlr", b"\x00" * 8 + b"mdirappl" + b"\x00" * 9)
meta = Atom.render(
b"meta", b"\x00\x00\x00\x00" + hdlr + ilst + self.__pad_ilst(ilst))
try:
path = atoms.path(b"moov", b"udta")
except KeyError:
# moov.udta not found -- create one
path = atoms.path(b"moov")
meta = Atom.render(b"udta", meta)
offset = path[-1].offset + 8
insert_bytes(fileobj, len(meta), offset)
fileobj.seek(offset)
fileobj.write(meta)
self.__update_parents(fileobj, path, len(meta))
self.__update_offsets(fileobj, atoms, len(meta), offset)
def __save_existing(self, fileobj, atoms, path, data):
# Replace the old ilst atom.
ilst = path.pop()
offset = ilst.offset
length = ilst.length
# Check for padding "free" atoms
meta = path[-1]
index = meta.children.index(ilst)
try:
prev = meta.children[index - 1]
if prev.name == b"free":
offset = prev.offset
length += prev.length
except IndexError:
pass
try:
next = meta.children[index + 1]
if next.name == b"free":
length += next.length
except IndexError:
pass
delta = len(data) - length
if delta > 0 or (delta < 0 and delta > -8):
data += self.__pad_ilst(data)
delta = len(data) - length
insert_bytes(fileobj, delta, offset)
elif delta < 0:
data += self.__pad_ilst(data, -delta - 8)
delta = 0
fileobj.seek(offset)
fileobj.write(data)
self.__update_parents(fileobj, path, delta)
self.__update_offsets(fileobj, atoms, delta, offset)
def __update_parents(self, fileobj, path, delta):
"""Update all parent atoms with the new size."""
for atom in path:
fileobj.seek(atom.offset)
size = cdata.uint_be(fileobj.read(4))
if size == 1: # 64bit
# skip name (4B) and read size (8B)
size = cdata.ulonglong_be(fileobj.read(12)[4:])
fileobj.seek(atom.offset + 8)
fileobj.write(cdata.to_ulonglong_be(size + delta))
else: # 32bit
fileobj.seek(atom.offset)
fileobj.write(cdata.to_uint_be(size + delta))
def __update_offset_table(self, fileobj, fmt, atom, delta, offset):
"""Update offset table in the specified atom."""
if atom.offset > offset:
atom.offset += delta
fileobj.seek(atom.offset + 12)
data = fileobj.read(atom.length - 12)
fmt = fmt % cdata.uint_be(data[:4])
offsets = struct.unpack(fmt, data[4:])
offsets = [o + (0, delta)[offset < o] for o in offsets]
fileobj.seek(atom.offset + 16)
fileobj.write(struct.pack(fmt, *offsets))
def __update_tfhd(self, fileobj, atom, delta, offset):
if atom.offset > offset:
atom.offset += delta
fileobj.seek(atom.offset + 9)
data = fileobj.read(atom.length - 9)
flags = cdata.uint_be(b"\x00" + data[:3])
if flags & 1:
o = cdata.ulonglong_be(data[7:15])
if o > offset:
o += delta
fileobj.seek(atom.offset + 16)
fileobj.write(cdata.to_ulonglong_be(o))
def __update_offsets(self, fileobj, atoms, delta, offset):
"""Update offset tables in all 'stco' and 'co64' atoms."""
if delta == 0:
return
moov = atoms[b"moov"]
for atom in moov.findall(b'stco', True):
self.__update_offset_table(fileobj, ">%dI", atom, delta, offset)
for atom in moov.findall(b'co64', True):
self.__update_offset_table(fileobj, ">%dQ", atom, delta, offset)
try:
for atom in atoms[b"moof"].findall(b'tfhd', True):
self.__update_tfhd(fileobj, atom, delta, offset)
except KeyError:
pass
def __parse_data(self, atom, data):
pos = 0
while pos < atom.length - 8:
head = data[pos:pos + 12]
if len(head) != 12:
raise MP4MetadataError("truncated atom % r" % atom.name)
length, name = struct.unpack(">I4s", head[:8])
version = ord(head[8:9])
flags = struct.unpack(">I", b"\x00" + head[9:12])[0]
if name != b"data":
raise MP4MetadataError(
"unexpected atom %r inside %r" % (name, atom.name))
chunk = data[pos + 16:pos + length]
if len(chunk) != length - 16:
raise MP4MetadataError("truncated atom % r" % atom.name)
yield version, flags, chunk
pos += length
def __add(self, key, value, single=False):
assert isinstance(key, str)
if single:
self[key] = value
else:
self.setdefault(key, []).extend(value)
def __render_data(self, key, version, flags, value):
return Atom.render(_key2name(key), b"".join([
Atom.render(
b"data", struct.pack(">2I", version << 24 | flags, 0) + data)
for data in value]))
def __parse_freeform(self, atom, data):
length = cdata.uint_be(data[:4])
mean = data[12:length]
pos = length
length = cdata.uint_be(data[pos:pos + 4])
name = data[pos + 12:pos + length]
pos += length
value = []
while pos < atom.length - 8:
length, atom_name = struct.unpack(">I4s", data[pos:pos + 8])
if atom_name != b"data":
raise MP4MetadataError(
"unexpected atom %r inside %r" % (atom_name, atom.name))
version = ord(data[pos + 8:pos + 8 + 1])
flags = struct.unpack(">I", b"\x00" + data[pos + 9:pos + 12])[0]
value.append(MP4FreeForm(data[pos + 16:pos + length],
dataformat=flags, version=version))
pos += length
key = _name2key(atom.name + b":" + mean + b":" + name)
self.__add(key, value)
def __render_freeform(self, key, value):
if isinstance(value, bytes):
value = [value]
dummy, mean, name = _key2name(key).split(b":", 2)
mean = struct.pack(">I4sI", len(mean) + 12, b"mean", 0) + mean
name = struct.pack(">I4sI", len(name) + 12, b"name", 0) + name
data = b""
for v in value:
flags = AtomDataType.UTF8
version = 0
if isinstance(v, MP4FreeForm):
flags = v.dataformat
version = v.version
data += struct.pack(
">I4s2I", len(v) + 16, b"data", version << 24 | flags, 0)
data += v
return Atom.render(b"----", mean + name + data)
def __parse_pair(self, atom, data):
key = _name2key(atom.name)
values = [struct.unpack(">2H", d[2:6]) for
version, flags, d in self.__parse_data(atom, data)]
self.__add(key, values)
def __render_pair(self, key, value):
data = []
for (track, total) in value:
if 0 <= track < 1 << 16 and 0 <= total < 1 << 16:
data.append(struct.pack(">4H", 0, track, total, 0))
else:
raise MP4MetadataValueError(
"invalid numeric pair %r" % ((track, total),))
return self.__render_data(key, 0, AtomDataType.IMPLICIT, data)
def __render_pair_no_trailing(self, key, value):
data = []
for (track, total) in value:
if 0 <= track < 1 << 16 and 0 <= total < 1 << 16:
data.append(struct.pack(">3H", 0, track, total))
else:
raise MP4MetadataValueError(
"invalid numeric pair %r" % ((track, total),))
return self.__render_data(key, 0, AtomDataType.IMPLICIT, data)
def __parse_genre(self, atom, data):
values = []
for version, flags, data in self.__parse_data(atom, data):
# version = 0, flags = 0
if len(data) != 2:
raise MP4MetadataValueError("invalid genre")
genre = cdata.short_be(data)
# Translate to a freeform genre.
try:
genre = GENRES[genre - 1]
except IndexError:
# this will make us write it back at least
raise MP4MetadataValueError("unknown genre")
values.append(genre)
key = _name2key(b"\xa9gen")
self.__add(key, values)
def __parse_tempo(self, atom, data):
values = []
for version, flags, data in self.__parse_data(atom, data):
# version = 0, flags = 0 or 21
if len(data) != 2:
raise MP4MetadataValueError("invalid tempo")
values.append(cdata.ushort_be(data))
key = _name2key(atom.name)
self.__add(key, values)
def __render_tempo(self, key, value):
try:
if len(value) == 0:
return self.__render_data(key, 0, AtomDataType.INTEGER, b"")
if (min(value) < 0) or (max(value) >= 2 ** 16):
raise MP4MetadataValueError(
"invalid 16 bit integers: %r" % value)
except TypeError:
raise MP4MetadataValueError(
"tmpo must be a list of 16 bit integers")
values = [cdata.to_ushort_be(v) for v in value]
return self.__render_data(key, 0, AtomDataType.INTEGER, values)
def __parse_bool(self, atom, data):
for version, flags, data in self.__parse_data(atom, data):
if len(data) != 1:
raise MP4MetadataValueError("invalid bool")
value = bool(ord(data))
key = _name2key(atom.name)
self.__add(key, value, single=True)
def __render_bool(self, key, value):
return self.__render_data(
key, 0, AtomDataType.INTEGER, [chr_(bool(value))])
def __parse_cover(self, atom, data):
values = []
pos = 0
while pos < atom.length - 8:
length, name, imageformat = struct.unpack(">I4sI",
data[pos:pos + 12])
if name != b"data":
if name == b"name":
pos += length
continue
raise MP4MetadataError(
"unexpected atom %r inside 'covr'" % name)
if imageformat not in (MP4Cover.FORMAT_JPEG, MP4Cover.FORMAT_PNG):
# Sometimes AtomDataType.IMPLICIT or simply wrong.
# In all cases it was jpeg, so default to it
imageformat = MP4Cover.FORMAT_JPEG
cover = MP4Cover(data[pos + 16:pos + length], imageformat)
values.append(cover)
pos += length
key = _name2key(atom.name)
self.__add(key, values)
def __render_cover(self, key, value):
atom_data = []
for cover in value:
try:
imageformat = cover.imageformat
except AttributeError:
imageformat = MP4Cover.FORMAT_JPEG
atom_data.append(Atom.render(
b"data", struct.pack(">2I", imageformat, 0) + cover))
return Atom.render(_key2name(key), b"".join(atom_data))
def __parse_text(self, atom, data, implicit=True):
# implicit = False, for parsing unknown atoms only take utf8 ones.
# For known ones we can assume the implicit are utf8 too.
values = []
for version, flags, atom_data in self.__parse_data(atom, data):
if implicit:
if flags not in (AtomDataType.IMPLICIT, AtomDataType.UTF8):
raise MP4MetadataError(
"Unknown atom type %r for %r" % (flags, atom.name))
else:
if flags != AtomDataType.UTF8:
raise MP4MetadataError(
"%r is not text, ignore" % atom.name)
try:
text = atom_data.decode("utf-8")
except UnicodeDecodeError as e:
raise MP4MetadataError("%s: %s" % (atom.name, e))
values.append(text)
key = _name2key(atom.name)
self.__add(key, values)
def __render_text(self, key, value, flags=AtomDataType.UTF8):
if isinstance(value, string_types):
value = [value]
encoded = []
for v in value:
if not isinstance(v, text_type):
if PY3:
raise TypeError("%r not str" % v)
v = v.decode("utf-8")
encoded.append(v.encode("utf-8"))
return self.__render_data(key, 0, flags, encoded)
def delete(self, filename):
"""Remove the metadata from the given filename."""
self._failed_atoms.clear()
self.clear()
self.save(filename)
__atoms = {
b"----": (__parse_freeform, __render_freeform),
b"trkn": (__parse_pair, __render_pair),
b"disk": (__parse_pair, __render_pair_no_trailing),
b"gnre": (__parse_genre, None),
b"tmpo": (__parse_tempo, __render_tempo),
b"cpil": (__parse_bool, __render_bool),
b"pgap": (__parse_bool, __render_bool),
b"pcst": (__parse_bool, __render_bool),
b"covr": (__parse_cover, __render_cover),
b"purl": (__parse_text, __render_text),
b"egid": (__parse_text, __render_text),
}
# these allow implicit flags and parse as text
for name in [b"\xa9nam", b"\xa9alb", b"\xa9ART", b"aART", b"\xa9wrt",
b"\xa9day", b"\xa9cmt", b"desc", b"purd", b"\xa9grp",
b"\xa9gen", b"\xa9lyr", b"catg", b"keyw", b"\xa9too",
b"cprt", b"soal", b"soaa", b"soar", b"sonm", b"soco",
b"sosn", b"tvsh"]:
__atoms[name] = (__parse_text, __render_text)
def pprint(self):
values = []
for key, value in iteritems(self):
if not isinstance(key, text_type):
key = key.decode("latin-1")
if key == "covr":
values.append("%s=%s" % (key, ", ".join(
["[%d bytes of data]" % len(data) for data in value])))
elif isinstance(value, list):
values.append("%s=%s" %
(key, " / ".join(map(text_type, value))))
else:
values.append("%s=%s" % (key, value))
return "\n".join(values)
class MP4Info(StreamInfo):
"""MPEG-4 stream information.
Attributes:
* bitrate -- bitrate in bits per second, as an int
* length -- file length in seconds, as a float
* channels -- number of audio channels
* sample_rate -- audio sampling rate in Hz
* bits_per_sample -- bits per sample
* codec (string):
* if starting with ``"mp4a"`` uses an mp4a audio codec
(see the codec parameter in rfc6381 for details e.g. ``"mp4a.40.2"``)
* for everything else see a list of possible values at
http://www.mp4ra.org/codecs.html
e.g. ``"mp4a"``, ``"alac"``, ``"mp4a.40.2"``, ``"ac-3"`` etc.
* codec_description (string):
Name of the codec used (ALAC, AAC LC, AC-3...). Values might change in
the future, use for display purposes only.
"""
bitrate = 0
channels = 0
sample_rate = 0
bits_per_sample = 0
codec = u""
codec_name = u""
def __init__(self, atoms, fileobj):
try:
moov = atoms[b"moov"]
except KeyError:
raise MP4StreamInfoError("not a MP4 file")
for trak in moov.findall(b"trak"):
hdlr = trak[b"mdia", b"hdlr"]
ok, data = hdlr.read(fileobj)
if not ok:
raise MP4StreamInfoError("Not enough data")
if data[8:12] == b"soun":
break
else:
raise MP4StreamInfoError("track has no audio data")
mdhd = trak[b"mdia", b"mdhd"]
ok, data = mdhd.read(fileobj)
if not ok:
raise MP4StreamInfoError("Not enough data")
try:
version, flags, data = parse_full_atom(data)
except ValueError as e:
raise MP4StreamInfoError(e)
if version == 0:
offset = 8
fmt = ">2I"
elif version == 1:
offset = 16
fmt = ">IQ"
else:
raise MP4StreamInfoError("Unknown mdhd version %d" % version)
end = offset + struct.calcsize(fmt)
unit, length = struct.unpack(fmt, data[offset:end])
try:
self.length = float(length) / unit
except ZeroDivisionError:
self.length = 0
try:
atom = trak[b"mdia", b"minf", b"stbl", b"stsd"]
except KeyError:
pass
else:
self._parse_stsd(atom, fileobj)
def _parse_stsd(self, atom, fileobj):
"""Sets channels, bits_per_sample, sample_rate and optionally bitrate.
Can raise MP4StreamInfoError.
"""
assert atom.name == b"stsd"
ok, data = atom.read(fileobj)
if not ok:
raise MP4StreamInfoError("Invalid stsd")
try:
version, flags, data = parse_full_atom(data)
except ValueError as e:
raise MP4StreamInfoError(e)
if version != 0:
raise MP4StreamInfoError("Unsupported stsd version")
try:
num_entries, offset = cdata.uint32_be_from(data, 0)
except cdata.error as e:
raise MP4StreamInfoError(e)
if num_entries == 0:
return
# look at the first entry if there is one
entry_fileobj = cBytesIO(data[offset:])
try:
entry_atom = Atom(entry_fileobj)
except AtomError as e:
raise MP4StreamInfoError(e)
try:
entry = AudioSampleEntry(entry_atom, entry_fileobj)
except ASEntryError as e:
raise MP4StreamInfoError(e)
else:
self.channels = entry.channels
self.bits_per_sample = entry.sample_size
self.sample_rate = entry.sample_rate
self.bitrate = entry.bitrate
self.codec = entry.codec
self.codec_description = entry.codec_description
def pprint(self):
return "MPEG-4 audio (%s), %.2f seconds, %d bps" % (
self.codec_description, self.length, self.bitrate)
class MP4(FileType):
"""An MPEG-4 audio file, probably containing AAC.
If more than one track is present in the file, the first is used.
Only audio ('soun') tracks will be read.
:ivar info: :class:`MP4Info`
:ivar tags: :class:`MP4Tags`
"""
MP4Tags = MP4Tags
_mimes = ["audio/mp4", "audio/x-m4a", "audio/mpeg4", "audio/aac"]
def load(self, filename):
self.filename = filename
with open(filename, "rb") as fileobj:
try:
atoms = Atoms(fileobj)
except AtomError as err:
reraise(error, err, sys.exc_info()[2])
try:
self.info = MP4Info(atoms, fileobj)
except error:
raise
except Exception as err:
reraise(MP4StreamInfoError, err, sys.exc_info()[2])
if not MP4Tags._can_load(atoms):
self.tags = None
else:
try:
self.tags = self.MP4Tags(atoms, fileobj)
except error:
raise
except Exception as err:
reraise(MP4MetadataError, err, sys.exc_info()[2])
def add_tags(self):
if self.tags is None:
self.tags = self.MP4Tags()
else:
raise error("an MP4 tag already exists")
@staticmethod
def score(filename, fileobj, header_data):
return (b"ftyp" in header_data) + (b"mp4" in header_data)
Open = MP4
def delete(filename):
"""Remove tags from a file."""
MP4(filename).delete()

541
lib/mutagen/mp4/_as_entry.py Normal file
View File

@@ -0,0 +1,541 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2014 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
from mutagen._compat import cBytesIO, xrange
from mutagen.aac import ProgramConfigElement
from mutagen._util import BitReader, BitReaderError, cdata, text_type
from ._util import parse_full_atom
from ._atom import Atom, AtomError
class ASEntryError(Exception):
pass
class AudioSampleEntry(object):
"""Parses an AudioSampleEntry atom.
Private API.
Attrs:
channels (int): number of channels
sample_size (int): sample size in bits
sample_rate (int): sample rate in Hz
bitrate (int): bits per second (0 means unknown)
codec (string):
audio codec, either 'mp4a[.*][.*]' (rfc6381) or 'alac'
codec_description (string): descriptive codec name e.g. "AAC LC+SBR"
Can raise ASEntryError.
"""
channels = 0
sample_size = 0
sample_rate = 0
bitrate = 0
codec = None
codec_description = None
def __init__(self, atom, fileobj):
ok, data = atom.read(fileobj)
if not ok:
raise ASEntryError("too short %r atom" % atom.name)
fileobj = cBytesIO(data)
r = BitReader(fileobj)
try:
# SampleEntry
r.skip(6 * 8) # reserved
r.skip(2 * 8) # data_ref_index
# AudioSampleEntry
r.skip(8 * 8) # reserved
self.channels = r.bits(16)
self.sample_size = r.bits(16)
r.skip(2 * 8) # pre_defined
r.skip(2 * 8) # reserved
self.sample_rate = r.bits(32) >> 16
except BitReaderError as e:
raise ASEntryError(e)
assert r.is_aligned()
try:
extra = Atom(fileobj)
except AtomError as e:
raise ASEntryError(e)
self.codec = atom.name.decode("latin-1")
self.codec_description = None
if atom.name == b"mp4a" and extra.name == b"esds":
self._parse_esds(extra, fileobj)
elif atom.name == b"alac" and extra.name == b"alac":
self._parse_alac(extra, fileobj)
elif atom.name == b"ac-3" and extra.name == b"dac3":
self._parse_dac3(extra, fileobj)
if self.codec_description is None:
self.codec_description = self.codec.upper()
def _parse_dac3(self, atom, fileobj):
# ETSI TS 102 366
assert atom.name == b"dac3"
ok, data = atom.read(fileobj)
if not ok:
raise ASEntryError("truncated %s atom" % atom.name)
fileobj = cBytesIO(data)
r = BitReader(fileobj)
# sample_rate in AudioSampleEntry covers values in
# fscod2 and not just fscod, so ignore fscod here.
try:
r.skip(2 + 5 + 3) # fscod, bsid, bsmod
acmod = r.bits(3)
lfeon = r.bits(1)
bit_rate_code = r.bits(5)
r.skip(5) # reserved
except BitReaderError as e:
raise ASEntryError(e)
self.channels = [2, 1, 2, 3, 3, 4, 4, 5][acmod] + lfeon
try:
self.bitrate = [
32, 40, 48, 56, 64, 80, 96, 112, 128, 160, 192,
224, 256, 320, 384, 448, 512, 576, 640][bit_rate_code] * 1000
except IndexError:
pass
def _parse_alac(self, atom, fileobj):
# https://alac.macosforge.org/trac/browser/trunk/
# ALACMagicCookieDescription.txt
assert atom.name == b"alac"
ok, data = atom.read(fileobj)
if not ok:
raise ASEntryError("truncated %s atom" % atom.name)
try:
version, flags, data = parse_full_atom(data)
except ValueError as e:
raise ASEntryError(e)
if version != 0:
raise ASEntryError("Unsupported version %d" % version)
fileobj = cBytesIO(data)
r = BitReader(fileobj)
try:
# for some files the AudioSampleEntry values default to 44100/2chan
# and the real info is in the alac cookie, so prefer it
r.skip(32) # frameLength
compatibleVersion = r.bits(8)
if compatibleVersion != 0:
return
self.sample_size = r.bits(8)
r.skip(8 + 8 + 8)
self.channels = r.bits(8)
r.skip(16 + 32)
self.bitrate = r.bits(32)
self.sample_rate = r.bits(32)
except BitReaderError as e:
raise ASEntryError(e)
def _parse_esds(self, esds, fileobj):
assert esds.name == b"esds"
ok, data = esds.read(fileobj)
if not ok:
raise ASEntryError("truncated %s atom" % esds.name)
try:
version, flags, data = parse_full_atom(data)
except ValueError as e:
raise ASEntryError(e)
if version != 0:
raise ASEntryError("Unsupported version %d" % version)
fileobj = cBytesIO(data)
r = BitReader(fileobj)
try:
tag = r.bits(8)
if tag != ES_Descriptor.TAG:
raise ASEntryError("unexpected descriptor: %d" % tag)
assert r.is_aligned()
except BitReaderError as e:
raise ASEntryError(e)
try:
decSpecificInfo = ES_Descriptor.parse(fileobj)
except DescriptorError as e:
raise ASEntryError(e)
dec_conf_desc = decSpecificInfo.decConfigDescr
self.bitrate = dec_conf_desc.avgBitrate
self.codec += dec_conf_desc.codec_param
self.codec_description = dec_conf_desc.codec_desc
decSpecificInfo = dec_conf_desc.decSpecificInfo
if decSpecificInfo is not None:
if decSpecificInfo.channels != 0:
self.channels = decSpecificInfo.channels
if decSpecificInfo.sample_rate != 0:
self.sample_rate = decSpecificInfo.sample_rate
class DescriptorError(Exception):
pass
class BaseDescriptor(object):
TAG = None
@classmethod
def _parse_desc_length_file(cls, fileobj):
"""May raise ValueError"""
value = 0
for i in xrange(4):
try:
b = cdata.uint8(fileobj.read(1))
except cdata.error as e:
raise ValueError(e)
value = (value << 7) | (b & 0x7f)
if not b >> 7:
break
else:
raise ValueError("invalid descriptor length")
return value
@classmethod
def parse(cls, fileobj):
"""Returns a parsed instance of the called type.
The file position is right after the descriptor after this returns.
Raises DescriptorError
"""
try:
length = cls._parse_desc_length_file(fileobj)
except ValueError as e:
raise DescriptorError(e)
pos = fileobj.tell()
instance = cls(fileobj, length)
left = length - (fileobj.tell() - pos)
if left < 0:
raise DescriptorError("descriptor parsing read too much data")
fileobj.seek(left, 1)
return instance
class ES_Descriptor(BaseDescriptor):
TAG = 0x3
def __init__(self, fileobj, length):
"""Raises DescriptorError"""
r = BitReader(fileobj)
try:
self.ES_ID = r.bits(16)
self.streamDependenceFlag = r.bits(1)
self.URL_Flag = r.bits(1)
self.OCRstreamFlag = r.bits(1)
self.streamPriority = r.bits(5)
if self.streamDependenceFlag:
self.dependsOn_ES_ID = r.bits(16)
if self.URL_Flag:
URLlength = r.bits(8)
self.URLstring = r.bytes(URLlength)
if self.OCRstreamFlag:
self.OCR_ES_Id = r.bits(16)
tag = r.bits(8)
except BitReaderError as e:
raise DescriptorError(e)
if tag != DecoderConfigDescriptor.TAG:
raise DescriptorError("unexpected DecoderConfigDescrTag %d" % tag)
assert r.is_aligned()
self.decConfigDescr = DecoderConfigDescriptor.parse(fileobj)
class DecoderConfigDescriptor(BaseDescriptor):
TAG = 0x4
decSpecificInfo = None
"""A DecoderSpecificInfo, optional"""
def __init__(self, fileobj, length):
"""Raises DescriptorError"""
r = BitReader(fileobj)
try:
self.objectTypeIndication = r.bits(8)
self.streamType = r.bits(6)
self.upStream = r.bits(1)
self.reserved = r.bits(1)
self.bufferSizeDB = r.bits(24)
self.maxBitrate = r.bits(32)
self.avgBitrate = r.bits(32)
if (self.objectTypeIndication, self.streamType) != (0x40, 0x5):
return
# all from here is optional
if length * 8 == r.get_position():
return
tag = r.bits(8)
except BitReaderError as e:
raise DescriptorError(e)
if tag == DecoderSpecificInfo.TAG:
assert r.is_aligned()
self.decSpecificInfo = DecoderSpecificInfo.parse(fileobj)
@property
def codec_param(self):
"""string"""
param = u".%X" % self.objectTypeIndication
info = self.decSpecificInfo
if info is not None:
param += u".%d" % info.audioObjectType
return param
@property
def codec_desc(self):
"""string or None"""
info = self.decSpecificInfo
desc = None
if info is not None:
desc = info.description
return desc
class DecoderSpecificInfo(BaseDescriptor):
TAG = 0x5
_TYPE_NAMES = [
None, "AAC MAIN", "AAC LC", "AAC SSR", "AAC LTP", "SBR",
"AAC scalable", "TwinVQ", "CELP", "HVXC", None, None, "TTSI",
"Main synthetic", "Wavetable synthesis", "General MIDI",
"Algorithmic Synthesis and Audio FX", "ER AAC LC", None, "ER AAC LTP",
"ER AAC scalable", "ER Twin VQ", "ER BSAC", "ER AAC LD", "ER CELP",
"ER HVXC", "ER HILN", "ER Parametric", "SSC", "PS", "MPEG Surround",
None, "Layer-1", "Layer-2", "Layer-3", "DST", "ALS", "SLS",
"SLS non-core", "ER AAC ELD", "SMR Simple", "SMR Main", "USAC",
"SAOC", "LD MPEG Surround", "USAC"
]
_FREQS = [
96000, 88200, 64000, 48000, 44100, 32000, 24000, 22050, 16000,
12000, 11025, 8000, 7350,
]
@property
def description(self):
"""string or None if unknown"""
name = None
try:
name = self._TYPE_NAMES[self.audioObjectType]
except IndexError:
pass
if name is None:
return
if self.sbrPresentFlag == 1:
name += "+SBR"
if self.psPresentFlag == 1:
name += "+PS"
return text_type(name)
@property
def sample_rate(self):
"""0 means unknown"""
if self.sbrPresentFlag == 1:
return self.extensionSamplingFrequency
elif self.sbrPresentFlag == 0:
return self.samplingFrequency
else:
# these are all types that support SBR
aot_can_sbr = (1, 2, 3, 4, 6, 17, 19, 20, 22)
if self.audioObjectType not in aot_can_sbr:
return self.samplingFrequency
# there shouldn't be SBR for > 48KHz
if self.samplingFrequency > 24000:
return self.samplingFrequency
# either samplingFrequency or samplingFrequency * 2
return 0
@property
def channels(self):
"""channel count or 0 for unknown"""
# from ProgramConfigElement()
if hasattr(self, "pce_channels"):
return self.pce_channels
conf = getattr(
self, "extensionChannelConfiguration", self.channelConfiguration)
if conf == 1:
if self.psPresentFlag == -1:
return 0
elif self.psPresentFlag == 1:
return 2
else:
return 1
elif conf == 7:
return 8
elif conf > 7:
return 0
else:
return conf
def _get_audio_object_type(self, r):
"""Raises BitReaderError"""
audioObjectType = r.bits(5)
if audioObjectType == 31:
audioObjectTypeExt = r.bits(6)
audioObjectType = 32 + audioObjectTypeExt
return audioObjectType
def _get_sampling_freq(self, r):
"""Raises BitReaderError"""
samplingFrequencyIndex = r.bits(4)
if samplingFrequencyIndex == 0xf:
samplingFrequency = r.bits(24)
else:
try:
samplingFrequency = self._FREQS[samplingFrequencyIndex]
except IndexError:
samplingFrequency = 0
return samplingFrequency
def __init__(self, fileobj, length):
"""Raises DescriptorError"""
r = BitReader(fileobj)
try:
self._parse(r, length)
except BitReaderError as e:
raise DescriptorError(e)
def _parse(self, r, length):
"""Raises BitReaderError"""
def bits_left():
return length * 8 - r.get_position()
self.audioObjectType = self._get_audio_object_type(r)
self.samplingFrequency = self._get_sampling_freq(r)
self.channelConfiguration = r.bits(4)
self.sbrPresentFlag = -1
self.psPresentFlag = -1
if self.audioObjectType in (5, 29):
self.extensionAudioObjectType = 5
self.sbrPresentFlag = 1
if self.audioObjectType == 29:
self.psPresentFlag = 1
self.extensionSamplingFrequency = self._get_sampling_freq(r)
self.audioObjectType = self._get_audio_object_type(r)
if self.audioObjectType == 22:
self.extensionChannelConfiguration = r.bits(4)
else:
self.extensionAudioObjectType = 0
if self.audioObjectType in (1, 2, 3, 4, 6, 7, 17, 19, 20, 21, 22, 23):
try:
GASpecificConfig(r, self)
except NotImplementedError:
# unsupported, (warn?)
return
else:
# unsupported
return
if self.audioObjectType in (
17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 39):
epConfig = r.bits(2)
if epConfig in (2, 3):
# unsupported
return
if self.extensionAudioObjectType != 5 and bits_left() >= 16:
syncExtensionType = r.bits(11)
if syncExtensionType == 0x2b7:
self.extensionAudioObjectType = self._get_audio_object_type(r)
if self.extensionAudioObjectType == 5:
self.sbrPresentFlag = r.bits(1)
if self.sbrPresentFlag == 1:
self.extensionSamplingFrequency = \
self._get_sampling_freq(r)
if bits_left() >= 12:
syncExtensionType = r.bits(11)
if syncExtensionType == 0x548:
self.psPresentFlag = r.bits(1)
if self.extensionAudioObjectType == 22:
self.sbrPresentFlag = r.bits(1)
if self.sbrPresentFlag == 1:
self.extensionSamplingFrequency = \
self._get_sampling_freq(r)
self.extensionChannelConfiguration = r.bits(4)
def GASpecificConfig(r, info):
"""Reads GASpecificConfig which is needed to get the data after that
(there is no length defined to skip it) and to read program_config_element
which can contain channel counts.
May raise BitReaderError on error or
NotImplementedError if some reserved data was set.
"""
assert isinstance(info, DecoderSpecificInfo)
r.skip(1) # frameLengthFlag
dependsOnCoreCoder = r.bits(1)
if dependsOnCoreCoder:
r.skip(14)
extensionFlag = r.bits(1)
if not info.channelConfiguration:
pce = ProgramConfigElement(r)
info.pce_channels = pce.channels
if info.audioObjectType == 6 or info.audioObjectType == 20:
r.skip(3)
if extensionFlag:
if info.audioObjectType == 22:
r.skip(5 + 11)
if info.audioObjectType in (17, 19, 20, 23):
r.skip(1 + 1 + 1)
extensionFlag3 = r.bits(1)
if extensionFlag3 != 0:
raise NotImplementedError("extensionFlag3 set")

190
lib/mutagen/mp4/_atom.py Normal file
View File

@@ -0,0 +1,190 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
import struct
from mutagen._compat import PY2
# This is not an exhaustive list of container atoms, but just the
# ones this module needs to peek inside.
_CONTAINERS = [b"moov", b"udta", b"trak", b"mdia", b"meta", b"ilst",
b"stbl", b"minf", b"moof", b"traf"]
_SKIP_SIZE = {b"meta": 4}
class AtomError(Exception):
pass
class Atom(object):
"""An individual atom.
Attributes:
children -- list child atoms (or None for non-container atoms)
length -- length of this atom, including length and name
name -- four byte name of the atom, as a str
offset -- location in the constructor-given fileobj of this atom
This structure should only be used internally by Mutagen.
"""
children = None
def __init__(self, fileobj, level=0):
"""May raise AtomError"""
self.offset = fileobj.tell()
try:
self.length, self.name = struct.unpack(">I4s", fileobj.read(8))
except struct.error:
raise AtomError("truncated data")
self._dataoffset = self.offset + 8
if self.length == 1:
try:
self.length, = struct.unpack(">Q", fileobj.read(8))
except struct.error:
raise AtomError("truncated data")
self._dataoffset += 8
if self.length < 16:
raise AtomError(
"64 bit atom length can only be 16 and higher")
elif self.length == 0:
if level != 0:
raise AtomError(
"only a top-level atom can have zero length")
# Only the last atom is supposed to have a zero-length, meaning it
# extends to the end of file.
fileobj.seek(0, 2)
self.length = fileobj.tell() - self.offset
fileobj.seek(self.offset + 8, 0)
elif self.length < 8:
raise AtomError(
"atom length can only be 0, 1 or 8 and higher")
if self.name in _CONTAINERS:
self.children = []
fileobj.seek(_SKIP_SIZE.get(self.name, 0), 1)
while fileobj.tell() < self.offset + self.length:
self.children.append(Atom(fileobj, level + 1))
else:
fileobj.seek(self.offset + self.length, 0)
def read(self, fileobj):
"""Return if all data could be read and the atom payload"""
fileobj.seek(self._dataoffset, 0)
length = self.length - (self._dataoffset - self.offset)
data = fileobj.read(length)
return len(data) == length, data
@staticmethod
def render(name, data):
"""Render raw atom data."""
# this raises OverflowError if Py_ssize_t can't handle the atom data
size = len(data) + 8
if size <= 0xFFFFFFFF:
return struct.pack(">I4s", size, name) + data
else:
return struct.pack(">I4sQ", 1, name, size + 8) + data
def findall(self, name, recursive=False):
"""Recursively find all child atoms by specified name."""
if self.children is not None:
for child in self.children:
if child.name == name:
yield child
if recursive:
for atom in child.findall(name, True):
yield atom
def __getitem__(self, remaining):
"""Look up a child atom, potentially recursively.
e.g. atom['udta', 'meta'] => <Atom name='meta' ...>
"""
if not remaining:
return self
elif self.children is None:
raise KeyError("%r is not a container" % self.name)
for child in self.children:
if child.name == remaining[0]:
return child[remaining[1:]]
else:
raise KeyError("%r not found" % remaining[0])
def __repr__(self):
cls = self.__class__.__name__
if self.children is None:
return "<%s name=%r length=%r offset=%r>" % (
cls, self.name, self.length, self.offset)
else:
children = "\n".join([" " + line for child in self.children
for line in repr(child).splitlines()])
return "<%s name=%r length=%r offset=%r\n%s>" % (
cls, self.name, self.length, self.offset, children)
class Atoms(object):
"""Root atoms in a given file.
Attributes:
atoms -- a list of top-level atoms as Atom objects
This structure should only be used internally by Mutagen.
"""
def __init__(self, fileobj):
self.atoms = []
fileobj.seek(0, 2)
end = fileobj.tell()
fileobj.seek(0)
while fileobj.tell() + 8 <= end:
self.atoms.append(Atom(fileobj))
def path(self, *names):
"""Look up and return the complete path of an atom.
For example, atoms.path('moov', 'udta', 'meta') will return a
list of three atoms, corresponding to the moov, udta, and meta
atoms.
"""
path = [self]
for name in names:
path.append(path[-1][name, ])
return path[1:]
def __contains__(self, names):
try:
self[names]
except KeyError:
return False
return True
def __getitem__(self, names):
"""Look up a child atom.
'names' may be a list of atoms (['moov', 'udta']) or a string
specifying the complete path ('moov.udta').
"""
if PY2:
if isinstance(names, basestring):
names = names.split(b".")
else:
if isinstance(names, bytes):
names = names.split(b".")
for child in self.atoms:
if child.name == names[0]:
return child[names[1:]]
else:
raise KeyError("%s not found" % names[0])
def __repr__(self):
return "\n".join([repr(child) for child in self.atoms])

21
lib/mutagen/mp4/_util.py Normal file
View File

@@ -0,0 +1,21 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2014 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
from mutagen._util import cdata
def parse_full_atom(data):
"""Some atoms are versioned. Split them up in (version, flags, payload).
Can raise ValueError.
"""
if len(data) < 4:
raise ValueError("not enough data")
version = ord(data[0:1])
flags = cdata.uint_be(b"\x00" + data[1:4])
return version, flags, data[4:]

270
lib/mutagen/musepack.py Normal file
View File

@@ -0,0 +1,270 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Lukas Lalinsky
# Copyright (C) 2012 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Musepack audio streams with APEv2 tags.
Musepack is an audio format originally based on the MPEG-1 Layer-2
algorithms. Stream versions 4 through 7 are supported.
For more information, see http://www.musepack.net/.
"""
__all__ = ["Musepack", "Open", "delete"]
import struct
from ._compat import endswith, xrange
from mutagen import StreamInfo
from mutagen.apev2 import APEv2File, error, delete
from mutagen.id3 import BitPaddedInt
from mutagen._util import cdata
class MusepackHeaderError(error):
pass
RATES = [44100, 48000, 37800, 32000]
def _parse_sv8_int(fileobj, limit=9):
"""Reads (max limit) bytes from fileobj until the MSB is zero.
All 7 LSB will be merged to a big endian uint.
Raises ValueError in case not MSB is zero, or EOFError in
case the file ended before limit is reached.
Returns (parsed number, number of bytes read)
"""
num = 0
for i in xrange(limit):
c = fileobj.read(1)
if len(c) != 1:
raise EOFError
c = bytearray(c)
num = (num << 7) | (c[0] & 0x7F)
if not c[0] & 0x80:
return num, i + 1
if limit > 0:
raise ValueError
return 0, 0
def _calc_sv8_gain(gain):
# 64.82 taken from mpcdec
return 64.82 - gain / 256.0
def _calc_sv8_peak(peak):
return (10 ** (peak / (256.0 * 20.0)) / 65535.0)
class MusepackInfo(StreamInfo):
"""Musepack stream information.
Attributes:
* channels -- number of audio channels
* length -- file length in seconds, as a float
* sample_rate -- audio sampling rate in Hz
* bitrate -- audio bitrate, in bits per second
* version -- Musepack stream version
Optional Attributes:
* title_gain, title_peak -- Replay Gain and peak data for this song
* album_gain, album_peak -- Replay Gain and peak data for this album
These attributes are only available in stream version 7/8. The
gains are a float, +/- some dB. The peaks are a percentage [0..1] of
the maximum amplitude. This means to get a number comparable to
VorbisGain, you must multiply the peak by 2.
"""
def __init__(self, fileobj):
header = fileobj.read(4)
if len(header) != 4:
raise MusepackHeaderError("not a Musepack file")
# Skip ID3v2 tags
if header[:3] == b"ID3":
header = fileobj.read(6)
if len(header) != 6:
raise MusepackHeaderError("not a Musepack file")
size = 10 + BitPaddedInt(header[2:6])
fileobj.seek(size)
header = fileobj.read(4)
if len(header) != 4:
raise MusepackHeaderError("not a Musepack file")
if header.startswith(b"MPCK"):
self.__parse_sv8(fileobj)
else:
self.__parse_sv467(fileobj)
if not self.bitrate and self.length != 0:
fileobj.seek(0, 2)
self.bitrate = int(round(fileobj.tell() * 8 / self.length))
def __parse_sv8(self, fileobj):
# SV8 http://trac.musepack.net/trac/wiki/SV8Specification
key_size = 2
mandatory_packets = [b"SH", b"RG"]
def check_frame_key(key):
if ((len(frame_type) != key_size) or
(not b'AA' <= frame_type <= b'ZZ')):
raise MusepackHeaderError("Invalid frame key.")
frame_type = fileobj.read(key_size)
check_frame_key(frame_type)
while frame_type not in (b"AP", b"SE") and mandatory_packets:
try:
frame_size, slen = _parse_sv8_int(fileobj)
except (EOFError, ValueError):
raise MusepackHeaderError("Invalid packet size.")
data_size = frame_size - key_size - slen
# packets can be at maximum data_size big and are padded with zeros
if frame_type == b"SH":
mandatory_packets.remove(frame_type)
self.__parse_stream_header(fileobj, data_size)
elif frame_type == b"RG":
mandatory_packets.remove(frame_type)
self.__parse_replaygain_packet(fileobj, data_size)
else:
fileobj.seek(data_size, 1)
frame_type = fileobj.read(key_size)
check_frame_key(frame_type)
if mandatory_packets:
raise MusepackHeaderError("Missing mandatory packets: %s." %
", ".join(map(repr, mandatory_packets)))
self.length = float(self.samples) / self.sample_rate
self.bitrate = 0
def __parse_stream_header(self, fileobj, data_size):
# skip CRC
fileobj.seek(4, 1)
remaining_size = data_size - 4
try:
self.version = bytearray(fileobj.read(1))[0]
except TypeError:
raise MusepackHeaderError("SH packet ended unexpectedly.")
remaining_size -= 1
try:
samples, l1 = _parse_sv8_int(fileobj)
samples_skip, l2 = _parse_sv8_int(fileobj)
except (EOFError, ValueError):
raise MusepackHeaderError(
"SH packet: Invalid sample counts.")
self.samples = samples - samples_skip
remaining_size -= l1 + l2
data = fileobj.read(remaining_size)
if len(data) != remaining_size:
raise MusepackHeaderError("SH packet ended unexpectedly.")
self.sample_rate = RATES[bytearray(data)[0] >> 5]
self.channels = (bytearray(data)[1] >> 4) + 1
def __parse_replaygain_packet(self, fileobj, data_size):
data = fileobj.read(data_size)
if data_size < 9:
raise MusepackHeaderError("Invalid RG packet size.")
if len(data) != data_size:
raise MusepackHeaderError("RG packet ended unexpectedly.")
title_gain = cdata.short_be(data[1:3])
title_peak = cdata.short_be(data[3:5])
album_gain = cdata.short_be(data[5:7])
album_peak = cdata.short_be(data[7:9])
if title_gain:
self.title_gain = _calc_sv8_gain(title_gain)
if title_peak:
self.title_peak = _calc_sv8_peak(title_peak)
if album_gain:
self.album_gain = _calc_sv8_gain(album_gain)
if album_peak:
self.album_peak = _calc_sv8_peak(album_peak)
def __parse_sv467(self, fileobj):
fileobj.seek(-4, 1)
header = fileobj.read(32)
if len(header) != 32:
raise MusepackHeaderError("not a Musepack file")
# SV7
if header.startswith(b"MP+"):
self.version = bytearray(header)[3] & 0xF
if self.version < 7:
raise MusepackHeaderError("not a Musepack file")
frames = cdata.uint_le(header[4:8])
flags = cdata.uint_le(header[8:12])
self.title_peak, self.title_gain = struct.unpack(
"<Hh", header[12:16])
self.album_peak, self.album_gain = struct.unpack(
"<Hh", header[16:20])
self.title_gain /= 100.0
self.album_gain /= 100.0
self.title_peak /= 65535.0
self.album_peak /= 65535.0
self.sample_rate = RATES[(flags >> 16) & 0x0003]
self.bitrate = 0
# SV4-SV6
else:
header_dword = cdata.uint_le(header[0:4])
self.version = (header_dword >> 11) & 0x03FF
if self.version < 4 or self.version > 6:
raise MusepackHeaderError("not a Musepack file")
self.bitrate = (header_dword >> 23) & 0x01FF
self.sample_rate = 44100
if self.version >= 5:
frames = cdata.uint_le(header[4:8])
else:
frames = cdata.ushort_le(header[6:8])
if self.version < 6:
frames -= 1
self.channels = 2
self.length = float(frames * 1152 - 576) / self.sample_rate
def pprint(self):
rg_data = []
if hasattr(self, "title_gain"):
rg_data.append("%+0.2f (title)" % self.title_gain)
if hasattr(self, "album_gain"):
rg_data.append("%+0.2f (album)" % self.album_gain)
rg_data = (rg_data and ", Gain: " + ", ".join(rg_data)) or ""
return "Musepack SV%d, %.2f seconds, %d Hz, %d bps%s" % (
self.version, self.length, self.sample_rate, self.bitrate, rg_data)
class Musepack(APEv2File):
_Info = MusepackInfo
_mimes = ["audio/x-musepack", "audio/x-mpc"]
@staticmethod
def score(filename, fileobj, header):
filename = filename.lower()
return (header.startswith(b"MP+") + header.startswith(b"MPCK") +
endswith(filename, b".mpc"))
Open = Musepack

508
lib/mutagen/ogg.py Normal file
View File

@@ -0,0 +1,508 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg bitstreams and pages.
This module reads and writes a subset of the Ogg bitstream format
version 0. It does *not* read or write Ogg Vorbis files! For that,
you should use mutagen.oggvorbis.
This implementation is based on the RFC 3533 standard found at
http://www.xiph.org/ogg/doc/rfc3533.txt.
"""
import struct
import sys
import zlib
from mutagen import FileType
from mutagen._util import cdata, insert_bytes, delete_bytes, MutagenError
from ._compat import cBytesIO, reraise, chr_
class error(IOError, MutagenError):
"""Ogg stream parsing errors."""
pass
class OggPage(object):
"""A single Ogg page (not necessarily a single encoded packet).
A page is a header of 26 bytes, followed by the length of the
data, followed by the data.
The constructor is givin a file-like object pointing to the start
of an Ogg page. After the constructor is finished it is pointing
to the start of the next page.
Attributes:
* version -- stream structure version (currently always 0)
* position -- absolute stream position (default -1)
* serial -- logical stream serial number (default 0)
* sequence -- page sequence number within logical stream (default 0)
* offset -- offset this page was read from (default None)
* complete -- if the last packet on this page is complete (default True)
* packets -- list of raw packet data (default [])
Note that if 'complete' is false, the next page's 'continued'
property must be true (so set both when constructing pages).
If a file-like object is supplied to the constructor, the above
attributes will be filled in based on it.
"""
version = 0
__type_flags = 0
position = 0
serial = 0
sequence = 0
offset = None
complete = True
def __init__(self, fileobj=None):
self.packets = []
if fileobj is None:
return
self.offset = fileobj.tell()
header = fileobj.read(27)
if len(header) == 0:
raise EOFError
try:
(oggs, self.version, self.__type_flags,
self.position, self.serial, self.sequence,
crc, segments) = struct.unpack("<4sBBqIIiB", header)
except struct.error:
raise error("unable to read full header; got %r" % header)
if oggs != b"OggS":
raise error("read %r, expected %r, at 0x%x" % (
oggs, b"OggS", fileobj.tell() - 27))
if self.version != 0:
raise error("version %r unsupported" % self.version)
total = 0
lacings = []
lacing_bytes = fileobj.read(segments)
if len(lacing_bytes) != segments:
raise error("unable to read %r lacing bytes" % segments)
for c in bytearray(lacing_bytes):
total += c
if c < 255:
lacings.append(total)
total = 0
if total:
lacings.append(total)
self.complete = False
self.packets = [fileobj.read(l) for l in lacings]
if [len(p) for p in self.packets] != lacings:
raise error("unable to read full data")
def __eq__(self, other):
"""Two Ogg pages are the same if they write the same data."""
try:
return (self.write() == other.write())
except AttributeError:
return False
__hash__ = object.__hash__
def __repr__(self):
attrs = ['version', 'position', 'serial', 'sequence', 'offset',
'complete', 'continued', 'first', 'last']
values = ["%s=%r" % (attr, getattr(self, attr)) for attr in attrs]
return "<%s %s, %d bytes in %d packets>" % (
type(self).__name__, " ".join(values), sum(map(len, self.packets)),
len(self.packets))
def write(self):
"""Return a string encoding of the page header and data.
A ValueError is raised if the data is too big to fit in a
single page.
"""
data = [
struct.pack("<4sBBqIIi", b"OggS", self.version, self.__type_flags,
self.position, self.serial, self.sequence, 0)
]
lacing_data = []
for datum in self.packets:
quot, rem = divmod(len(datum), 255)
lacing_data.append(b"\xff" * quot + chr_(rem))
lacing_data = b"".join(lacing_data)
if not self.complete and lacing_data.endswith(b"\x00"):
lacing_data = lacing_data[:-1]
data.append(chr_(len(lacing_data)))
data.append(lacing_data)
data.extend(self.packets)
data = b"".join(data)
# Python's CRC is swapped relative to Ogg's needs.
# crc32 returns uint prior to py2.6 on some platforms, so force uint
crc = (~zlib.crc32(data.translate(cdata.bitswap), -1)) & 0xffffffff
# Although we're using to_uint_be, this actually makes the CRC
# a proper le integer, since Python's CRC is byteswapped.
crc = cdata.to_uint_be(crc).translate(cdata.bitswap)
data = data[:22] + crc + data[26:]
return data
@property
def size(self):
"""Total frame size."""
size = 27 # Initial header size
for datum in self.packets:
quot, rem = divmod(len(datum), 255)
size += quot + 1
if not self.complete and rem == 0:
# Packet contains a multiple of 255 bytes and is not
# terminated, so we don't have a \x00 at the end.
size -= 1
size += sum(map(len, self.packets))
return size
def __set_flag(self, bit, val):
mask = 1 << bit
if val:
self.__type_flags |= mask
else:
self.__type_flags &= ~mask
continued = property(
lambda self: cdata.test_bit(self.__type_flags, 0),
lambda self, v: self.__set_flag(0, v),
doc="The first packet is continued from the previous page.")
first = property(
lambda self: cdata.test_bit(self.__type_flags, 1),
lambda self, v: self.__set_flag(1, v),
doc="This is the first page of a logical bitstream.")
last = property(
lambda self: cdata.test_bit(self.__type_flags, 2),
lambda self, v: self.__set_flag(2, v),
doc="This is the last page of a logical bitstream.")
@staticmethod
def renumber(fileobj, serial, start):
"""Renumber pages belonging to a specified logical stream.
fileobj must be opened with mode r+b or w+b.
Starting at page number 'start', renumber all pages belonging
to logical stream 'serial'. Other pages will be ignored.
fileobj must point to the start of a valid Ogg page; any
occuring after it and part of the specified logical stream
will be numbered. No adjustment will be made to the data in
the pages nor the granule position; only the page number, and
so also the CRC.
If an error occurs (e.g. non-Ogg data is found), fileobj will
be left pointing to the place in the stream the error occured,
but the invalid data will be left intact (since this function
does not change the total file size).
"""
number = start
while True:
try:
page = OggPage(fileobj)
except EOFError:
break
else:
if page.serial != serial:
# Wrong stream, skip this page.
continue
# Changing the number can't change the page size,
# so seeking back based on the current size is safe.
fileobj.seek(-page.size, 1)
page.sequence = number
fileobj.write(page.write())
fileobj.seek(page.offset + page.size, 0)
number += 1
@staticmethod
def to_packets(pages, strict=False):
"""Construct a list of packet data from a list of Ogg pages.
If strict is true, the first page must start a new packet,
and the last page must end the last packet.
"""
serial = pages[0].serial
sequence = pages[0].sequence
packets = []
if strict:
if pages[0].continued:
raise ValueError("first packet is continued")
if not pages[-1].complete:
raise ValueError("last packet does not complete")
elif pages and pages[0].continued:
packets.append([b""])
for page in pages:
if serial != page.serial:
raise ValueError("invalid serial number in %r" % page)
elif sequence != page.sequence:
raise ValueError("bad sequence number in %r" % page)
else:
sequence += 1
if page.continued:
packets[-1].append(page.packets[0])
else:
packets.append([page.packets[0]])
packets.extend([p] for p in page.packets[1:])
return [b"".join(p) for p in packets]
@staticmethod
def from_packets(packets, sequence=0, default_size=4096,
wiggle_room=2048):
"""Construct a list of Ogg pages from a list of packet data.
The algorithm will generate pages of approximately
default_size in size (rounded down to the nearest multiple of
255). However, it will also allow pages to increase to
approximately default_size + wiggle_room if allowing the
wiggle room would finish a packet (only one packet will be
finished in this way per page; if the next packet would fit
into the wiggle room, it still starts on a new page).
This method reduces packet fragmentation when packet sizes are
slightly larger than the default page size, while still
ensuring most pages are of the average size.
Pages are numbered started at 'sequence'; other information is
uninitialized.
"""
chunk_size = (default_size // 255) * 255
pages = []
page = OggPage()
page.sequence = sequence
for packet in packets:
page.packets.append(b"")
while packet:
data, packet = packet[:chunk_size], packet[chunk_size:]
if page.size < default_size and len(page.packets) < 255:
page.packets[-1] += data
else:
# If we've put any packet data into this page yet,
# we need to mark it incomplete. However, we can
# also have just started this packet on an already
# full page, in which case, just start the new
# page with this packet.
if page.packets[-1]:
page.complete = False
if len(page.packets) == 1:
page.position = -1
else:
page.packets.pop(-1)
pages.append(page)
page = OggPage()
page.continued = not pages[-1].complete
page.sequence = pages[-1].sequence + 1
page.packets.append(data)
if len(packet) < wiggle_room:
page.packets[-1] += packet
packet = b""
if page.packets:
pages.append(page)
return pages
@classmethod
def replace(cls, fileobj, old_pages, new_pages):
"""Replace old_pages with new_pages within fileobj.
old_pages must have come from reading fileobj originally.
new_pages are assumed to have the 'same' data as old_pages,
and so the serial and sequence numbers will be copied, as will
the flags for the first and last pages.
fileobj will be resized and pages renumbered as necessary. As
such, it must be opened r+b or w+b.
"""
# Number the new pages starting from the first old page.
first = old_pages[0].sequence
for page, seq in zip(new_pages, range(first, first + len(new_pages))):
page.sequence = seq
page.serial = old_pages[0].serial
new_pages[0].first = old_pages[0].first
new_pages[0].last = old_pages[0].last
new_pages[0].continued = old_pages[0].continued
new_pages[-1].first = old_pages[-1].first
new_pages[-1].last = old_pages[-1].last
new_pages[-1].complete = old_pages[-1].complete
if not new_pages[-1].complete and len(new_pages[-1].packets) == 1:
new_pages[-1].position = -1
new_data = b"".join(cls.write(p) for p in new_pages)
# Make room in the file for the new data.
delta = len(new_data)
fileobj.seek(old_pages[0].offset, 0)
insert_bytes(fileobj, delta, old_pages[0].offset)
fileobj.seek(old_pages[0].offset, 0)
fileobj.write(new_data)
new_data_end = old_pages[0].offset + delta
# Go through the old pages and delete them. Since we shifted
# the data down the file, we need to adjust their offsets. We
# also need to go backwards, so we don't adjust the deltas of
# the other pages.
old_pages.reverse()
for old_page in old_pages:
adj_offset = old_page.offset + delta
delete_bytes(fileobj, old_page.size, adj_offset)
# Finally, if there's any discrepency in length, we need to
# renumber the pages for the logical stream.
if len(old_pages) != len(new_pages):
fileobj.seek(new_data_end, 0)
serial = new_pages[-1].serial
sequence = new_pages[-1].sequence + 1
cls.renumber(fileobj, serial, sequence)
@staticmethod
def find_last(fileobj, serial):
"""Find the last page of the stream 'serial'.
If the file is not multiplexed this function is fast. If it is,
it must read the whole the stream.
This finds the last page in the actual file object, or the last
page in the stream (with eos set), whichever comes first.
"""
# For non-muxed streams, look at the last page.
try:
fileobj.seek(-256 * 256, 2)
except IOError:
# The file is less than 64k in length.
fileobj.seek(0)
data = fileobj.read()
try:
index = data.rindex(b"OggS")
except ValueError:
raise error("unable to find final Ogg header")
bytesobj = cBytesIO(data[index:])
best_page = None
try:
page = OggPage(bytesobj)
except error:
pass
else:
if page.serial == serial:
if page.last:
return page
else:
best_page = page
else:
best_page = None
# The stream is muxed, so use the slow way.
fileobj.seek(0)
try:
page = OggPage(fileobj)
while not page.last:
page = OggPage(fileobj)
while page.serial != serial:
page = OggPage(fileobj)
best_page = page
return page
except error:
return best_page
except EOFError:
return best_page
class OggFileType(FileType):
"""An generic Ogg file."""
_Info = None
_Tags = None
_Error = None
_mimes = ["application/ogg", "application/x-ogg"]
def load(self, filename):
"""Load file information from a filename."""
self.filename = filename
fileobj = open(filename, "rb")
try:
try:
self.info = self._Info(fileobj)
self.tags = self._Tags(fileobj, self.info)
self.info._post_tags(fileobj)
except error as e:
reraise(self._Error, e, sys.exc_info()[2])
except EOFError:
raise self._Error("no appropriate stream found")
finally:
fileobj.close()
def delete(self, filename=None):
"""Remove tags from a file.
If no filename is given, the one most recently loaded is used.
"""
if filename is None:
filename = self.filename
self.tags.clear()
fileobj = open(filename, "rb+")
try:
try:
self.tags._inject(fileobj)
except error as e:
reraise(self._Error, e, sys.exc_info()[2])
except EOFError:
raise self._Error("no appropriate stream found")
finally:
fileobj.close()
def save(self, filename=None):
"""Save a tag to a file.
If no filename is given, the one most recently loaded is used.
"""
if filename is None:
filename = self.filename
fileobj = open(filename, "rb+")
try:
try:
self.tags._inject(fileobj)
except error as e:
reraise(self._Error, e, sys.exc_info()[2])
except EOFError:
raise self._Error("no appropriate stream found")
finally:
fileobj.close()

147
lib/mutagen/oggflac.py Normal file
View File

@@ -0,0 +1,147 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg FLAC comments.
This module handles FLAC files wrapped in an Ogg bitstream. The first
FLAC stream found is used. For 'naked' FLACs, see mutagen.flac.
This module is based off the specification at
http://flac.sourceforge.net/ogg_mapping.html.
"""
__all__ = ["OggFLAC", "Open", "delete"]
import struct
from ._compat import cBytesIO
from mutagen.flac import StreamInfo, VCFLACDict, StrictFileObject
from mutagen.ogg import OggPage, OggFileType, error as OggError
class error(OggError):
pass
class OggFLACHeaderError(error):
pass
class OggFLACStreamInfo(StreamInfo):
"""Ogg FLAC general header and stream info.
This encompasses the Ogg wrapper for the FLAC STREAMINFO metadata
block, as well as the Ogg codec setup that precedes it.
Attributes (in addition to StreamInfo's):
* packets -- number of metadata packets
* serial -- Ogg logical stream serial number
"""
packets = 0
serial = 0
def load(self, data):
# Ogg expects file objects that don't raise on read
if isinstance(data, StrictFileObject):
data = data._fileobj
page = OggPage(data)
while not page.packets[0].startswith(b"\x7FFLAC"):
page = OggPage(data)
major, minor, self.packets, flac = struct.unpack(
">BBH4s", page.packets[0][5:13])
if flac != b"fLaC":
raise OggFLACHeaderError("invalid FLAC marker (%r)" % flac)
elif (major, minor) != (1, 0):
raise OggFLACHeaderError(
"unknown mapping version: %d.%d" % (major, minor))
self.serial = page.serial
# Skip over the block header.
stringobj = StrictFileObject(cBytesIO(page.packets[0][17:]))
super(OggFLACStreamInfo, self).load(stringobj)
def _post_tags(self, fileobj):
if self.length:
return
page = OggPage.find_last(fileobj, self.serial)
self.length = page.position / float(self.sample_rate)
def pprint(self):
return u"Ogg " + super(OggFLACStreamInfo, self).pprint()
class OggFLACVComment(VCFLACDict):
def load(self, data, info, errors='replace'):
# data should be pointing at the start of an Ogg page, after
# the first FLAC page.
pages = []
complete = False
while not complete:
page = OggPage(data)
if page.serial == info.serial:
pages.append(page)
complete = page.complete or (len(page.packets) > 1)
comment = cBytesIO(OggPage.to_packets(pages)[0][4:])
super(OggFLACVComment, self).load(comment, errors=errors)
def _inject(self, fileobj):
"""Write tag data into the FLAC Vorbis comment packet/page."""
# Ogg FLAC has no convenient data marker like Vorbis, but the
# second packet - and second page - must be the comment data.
fileobj.seek(0)
page = OggPage(fileobj)
while not page.packets[0].startswith(b"\x7FFLAC"):
page = OggPage(fileobj)
first_page = page
while not (page.sequence == 1 and page.serial == first_page.serial):
page = OggPage(fileobj)
old_pages = [page]
while not (old_pages[-1].complete or len(old_pages[-1].packets) > 1):
page = OggPage(fileobj)
if page.serial == first_page.serial:
old_pages.append(page)
packets = OggPage.to_packets(old_pages, strict=False)
# Set the new comment block.
data = self.write()
data = packets[0][:1] + struct.pack(">I", len(data))[-3:] + data
packets[0] = data
new_pages = OggPage.from_packets(packets, old_pages[0].sequence)
OggPage.replace(fileobj, old_pages, new_pages)
class OggFLAC(OggFileType):
"""An Ogg FLAC file."""
_Info = OggFLACStreamInfo
_Tags = OggFLACVComment
_Error = OggFLACHeaderError
_mimes = ["audio/x-oggflac"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"OggS") * (
(b"FLAC" in header) + (b"fLaC" in header)))
Open = OggFLAC
def delete(filename):
"""Remove tags from a file."""
OggFLAC(filename).delete()

138
lib/mutagen/oggopus.py Normal file
View File

@@ -0,0 +1,138 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2012, 2013 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg Opus comments.
This module handles Opus files wrapped in an Ogg bitstream. The
first Opus stream found is used.
Based on http://tools.ietf.org/html/draft-terriberry-oggopus-01
"""
__all__ = ["OggOpus", "Open", "delete"]
import struct
from mutagen import StreamInfo
from mutagen._compat import BytesIO
from mutagen._vorbis import VCommentDict
from mutagen.ogg import OggPage, OggFileType, error as OggError
class error(OggError):
pass
class OggOpusHeaderError(error):
pass
class OggOpusInfo(StreamInfo):
"""Ogg Opus stream information.
Attributes:
* length - file length in seconds, as a float
* channels - number of channels
"""
length = 0
def __init__(self, fileobj):
page = OggPage(fileobj)
while not page.packets[0].startswith(b"OpusHead"):
page = OggPage(fileobj)
self.serial = page.serial
if not page.first:
raise OggOpusHeaderError(
"page has ID header, but doesn't start a stream")
(version, self.channels, pre_skip, orig_sample_rate, output_gain,
channel_map) = struct.unpack("<BBHIhB", page.packets[0][8:19])
self.__pre_skip = pre_skip
# only the higher 4 bits change on incombatible changes
major = version >> 4
if major != 0:
raise OggOpusHeaderError("version %r unsupported" % major)
def _post_tags(self, fileobj):
page = OggPage.find_last(fileobj, self.serial)
self.length = (page.position - self.__pre_skip) / float(48000)
def pprint(self):
return u"Ogg Opus, %.2f seconds" % (self.length)
class OggOpusVComment(VCommentDict):
"""Opus comments embedded in an Ogg bitstream."""
def __get_comment_pages(self, fileobj, info):
# find the first tags page with the right serial
page = OggPage(fileobj)
while ((info.serial != page.serial) or
not page.packets[0].startswith(b"OpusTags")):
page = OggPage(fileobj)
# get all comment pages
pages = [page]
while not (pages[-1].complete or len(pages[-1].packets) > 1):
page = OggPage(fileobj)
if page.serial == pages[0].serial:
pages.append(page)
return pages
def __init__(self, fileobj, info):
pages = self.__get_comment_pages(fileobj, info)
data = OggPage.to_packets(pages)[0][8:] # Strip OpusTags
fileobj = BytesIO(data)
super(OggOpusVComment, self).__init__(fileobj, framing=False)
# in case the LSB of the first byte after v-comment is 1, preserve the
# following data
padding_flag = fileobj.read(1)
if padding_flag and ord(padding_flag) & 0x1:
self._pad_data = padding_flag + fileobj.read()
else:
self._pad_data = b""
def _inject(self, fileobj):
fileobj.seek(0)
info = OggOpusInfo(fileobj)
old_pages = self.__get_comment_pages(fileobj, info)
packets = OggPage.to_packets(old_pages)
packets[0] = b"OpusTags" + self.write(framing=False) + self._pad_data
new_pages = OggPage.from_packets(packets, old_pages[0].sequence)
OggPage.replace(fileobj, old_pages, new_pages)
class OggOpus(OggFileType):
"""An Ogg Opus file."""
_Info = OggOpusInfo
_Tags = OggOpusVComment
_Error = OggOpusHeaderError
_mimes = ["audio/ogg", "audio/ogg; codecs=opus"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"OggS") * (b"OpusHead" in header))
Open = OggOpus
def delete(filename):
"""Remove tags from a file."""
OggOpus(filename).delete()

138
lib/mutagen/oggspeex.py Normal file
View File

@@ -0,0 +1,138 @@
# -*- coding: utf-8 -*-
# Copyright 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg Speex comments.
This module handles Speex files wrapped in an Ogg bitstream. The
first Speex stream found is used.
Read more about Ogg Speex at http://www.speex.org/. This module is
based on the specification at http://www.speex.org/manual2/node7.html
and clarifications after personal communication with Jean-Marc,
http://lists.xiph.org/pipermail/speex-dev/2006-July/004676.html.
"""
__all__ = ["OggSpeex", "Open", "delete"]
from mutagen import StreamInfo
from mutagen._vorbis import VCommentDict
from mutagen.ogg import OggPage, OggFileType, error as OggError
from mutagen._util import cdata
class error(OggError):
pass
class OggSpeexHeaderError(error):
pass
class OggSpeexInfo(StreamInfo):
"""Ogg Speex stream information.
Attributes:
* bitrate - nominal bitrate in bits per second
* channels - number of channels
* length - file length in seconds, as a float
The reference encoder does not set the bitrate; in this case,
the bitrate will be 0.
"""
length = 0
def __init__(self, fileobj):
page = OggPage(fileobj)
while not page.packets[0].startswith(b"Speex "):
page = OggPage(fileobj)
if not page.first:
raise OggSpeexHeaderError(
"page has ID header, but doesn't start a stream")
self.sample_rate = cdata.uint_le(page.packets[0][36:40])
self.channels = cdata.uint_le(page.packets[0][48:52])
self.bitrate = max(0, cdata.int_le(page.packets[0][52:56]))
self.serial = page.serial
def _post_tags(self, fileobj):
page = OggPage.find_last(fileobj, self.serial)
self.length = page.position / float(self.sample_rate)
def pprint(self):
return u"Ogg Speex, %.2f seconds" % self.length
class OggSpeexVComment(VCommentDict):
"""Speex comments embedded in an Ogg bitstream."""
def __init__(self, fileobj, info):
pages = []
complete = False
while not complete:
page = OggPage(fileobj)
if page.serial == info.serial:
pages.append(page)
complete = page.complete or (len(page.packets) > 1)
data = OggPage.to_packets(pages)[0] + b"\x01"
super(OggSpeexVComment, self).__init__(data, framing=False)
def _inject(self, fileobj):
"""Write tag data into the Speex comment packet/page."""
fileobj.seek(0)
# Find the first header page, with the stream info.
# Use it to get the serial number.
page = OggPage(fileobj)
while not page.packets[0].startswith(b"Speex "):
page = OggPage(fileobj)
# Look for the next page with that serial number, it'll start
# the comment packet.
serial = page.serial
page = OggPage(fileobj)
while page.serial != serial:
page = OggPage(fileobj)
# Then find all the pages with the comment packet.
old_pages = [page]
while not (old_pages[-1].complete or len(old_pages[-1].packets) > 1):
page = OggPage(fileobj)
if page.serial == old_pages[0].serial:
old_pages.append(page)
packets = OggPage.to_packets(old_pages, strict=False)
# Set the new comment packet.
packets[0] = self.write(framing=False)
new_pages = OggPage.from_packets(packets, old_pages[0].sequence)
OggPage.replace(fileobj, old_pages, new_pages)
class OggSpeex(OggFileType):
"""An Ogg Speex file."""
_Info = OggSpeexInfo
_Tags = OggSpeexVComment
_Error = OggSpeexHeaderError
_mimes = ["audio/x-speex"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"OggS") * (b"Speex " in header))
Open = OggSpeex
def delete(filename):
"""Remove tags from a file."""
OggSpeex(filename).delete()

131
lib/mutagen/oggtheora.py Normal file
View File

@@ -0,0 +1,131 @@
# -*- coding: utf-8 -*-
# Copyright 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg Theora comments.
This module handles Theora files wrapped in an Ogg bitstream. The
first Theora stream found is used.
Based on the specification at http://theora.org/doc/Theora_I_spec.pdf.
"""
__all__ = ["OggTheora", "Open", "delete"]
import struct
from mutagen import StreamInfo
from mutagen._vorbis import VCommentDict
from mutagen._util import cdata
from mutagen.ogg import OggPage, OggFileType, error as OggError
class error(OggError):
pass
class OggTheoraHeaderError(error):
pass
class OggTheoraInfo(StreamInfo):
"""Ogg Theora stream information.
Attributes:
* length - file length in seconds, as a float
* fps - video frames per second, as a float
"""
length = 0
def __init__(self, fileobj):
page = OggPage(fileobj)
while not page.packets[0].startswith(b"\x80theora"):
page = OggPage(fileobj)
if not page.first:
raise OggTheoraHeaderError(
"page has ID header, but doesn't start a stream")
data = page.packets[0]
vmaj, vmin = struct.unpack("2B", data[7:9])
if (vmaj, vmin) != (3, 2):
raise OggTheoraHeaderError(
"found Theora version %d.%d != 3.2" % (vmaj, vmin))
fps_num, fps_den = struct.unpack(">2I", data[22:30])
self.fps = fps_num / float(fps_den)
self.bitrate = cdata.uint_be(b"\x00" + data[37:40])
self.granule_shift = (cdata.ushort_be(data[40:42]) >> 5) & 0x1F
self.serial = page.serial
def _post_tags(self, fileobj):
page = OggPage.find_last(fileobj, self.serial)
position = page.position
mask = (1 << self.granule_shift) - 1
frames = (position >> self.granule_shift) + (position & mask)
self.length = frames / float(self.fps)
def pprint(self):
return "Ogg Theora, %.2f seconds, %d bps" % (self.length, self.bitrate)
class OggTheoraCommentDict(VCommentDict):
"""Theora comments embedded in an Ogg bitstream."""
def __init__(self, fileobj, info):
pages = []
complete = False
while not complete:
page = OggPage(fileobj)
if page.serial == info.serial:
pages.append(page)
complete = page.complete or (len(page.packets) > 1)
data = OggPage.to_packets(pages)[0][7:]
super(OggTheoraCommentDict, self).__init__(data + b"\x01")
def _inject(self, fileobj):
"""Write tag data into the Theora comment packet/page."""
fileobj.seek(0)
page = OggPage(fileobj)
while not page.packets[0].startswith(b"\x81theora"):
page = OggPage(fileobj)
old_pages = [page]
while not (old_pages[-1].complete or len(old_pages[-1].packets) > 1):
page = OggPage(fileobj)
if page.serial == old_pages[0].serial:
old_pages.append(page)
packets = OggPage.to_packets(old_pages, strict=False)
packets[0] = b"\x81theora" + self.write(framing=False)
new_pages = OggPage.from_packets(packets, old_pages[0].sequence)
OggPage.replace(fileobj, old_pages, new_pages)
class OggTheora(OggFileType):
"""An Ogg Theora file."""
_Info = OggTheoraInfo
_Tags = OggTheoraCommentDict
_Error = OggTheoraHeaderError
_mimes = ["video/x-theora"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"OggS") *
((b"\x80theora" in header) + (b"\x81theora" in header)) * 2)
Open = OggTheora
def delete(filename):
"""Remove tags from a file."""
OggTheora(filename).delete()

139
lib/mutagen/oggvorbis.py Normal file
View File

@@ -0,0 +1,139 @@
# -*- coding: utf-8 -*-
# Copyright 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Read and write Ogg Vorbis comments.
This module handles Vorbis files wrapped in an Ogg bitstream. The
first Vorbis stream found is used.
Read more about Ogg Vorbis at http://vorbis.com/. This module is based
on the specification at http://www.xiph.org/vorbis/doc/Vorbis_I_spec.html.
"""
__all__ = ["OggVorbis", "Open", "delete"]
import struct
from mutagen import StreamInfo
from mutagen._vorbis import VCommentDict
from mutagen.ogg import OggPage, OggFileType, error as OggError
class error(OggError):
pass
class OggVorbisHeaderError(error):
pass
class OggVorbisInfo(StreamInfo):
"""Ogg Vorbis stream information.
Attributes:
* length - file length in seconds, as a float
* bitrate - nominal ('average') bitrate in bits per second, as an int
"""
length = 0
def __init__(self, fileobj):
page = OggPage(fileobj)
while not page.packets[0].startswith(b"\x01vorbis"):
page = OggPage(fileobj)
if not page.first:
raise OggVorbisHeaderError(
"page has ID header, but doesn't start a stream")
(self.channels, self.sample_rate, max_bitrate, nominal_bitrate,
min_bitrate) = struct.unpack("<B4i", page.packets[0][11:28])
self.serial = page.serial
max_bitrate = max(0, max_bitrate)
min_bitrate = max(0, min_bitrate)
nominal_bitrate = max(0, nominal_bitrate)
if nominal_bitrate == 0:
self.bitrate = (max_bitrate + min_bitrate) // 2
elif max_bitrate and max_bitrate < nominal_bitrate:
# If the max bitrate is less than the nominal, we know
# the nominal is wrong.
self.bitrate = max_bitrate
elif min_bitrate > nominal_bitrate:
self.bitrate = min_bitrate
else:
self.bitrate = nominal_bitrate
def _post_tags(self, fileobj):
page = OggPage.find_last(fileobj, self.serial)
self.length = page.position / float(self.sample_rate)
def pprint(self):
return u"Ogg Vorbis, %.2f seconds, %d bps" % (
self.length, self.bitrate)
class OggVCommentDict(VCommentDict):
"""Vorbis comments embedded in an Ogg bitstream."""
def __init__(self, fileobj, info):
pages = []
complete = False
while not complete:
page = OggPage(fileobj)
if page.serial == info.serial:
pages.append(page)
complete = page.complete or (len(page.packets) > 1)
data = OggPage.to_packets(pages)[0][7:] # Strip off "\x03vorbis".
super(OggVCommentDict, self).__init__(data)
def _inject(self, fileobj):
"""Write tag data into the Vorbis comment packet/page."""
# Find the old pages in the file; we'll need to remove them,
# plus grab any stray setup packet data out of them.
fileobj.seek(0)
page = OggPage(fileobj)
while not page.packets[0].startswith(b"\x03vorbis"):
page = OggPage(fileobj)
old_pages = [page]
while not (old_pages[-1].complete or len(old_pages[-1].packets) > 1):
page = OggPage(fileobj)
if page.serial == old_pages[0].serial:
old_pages.append(page)
packets = OggPage.to_packets(old_pages, strict=False)
# Set the new comment packet.
packets[0] = b"\x03vorbis" + self.write()
new_pages = OggPage.from_packets(packets, old_pages[0].sequence)
OggPage.replace(fileobj, old_pages, new_pages)
class OggVorbis(OggFileType):
"""An Ogg Vorbis file."""
_Info = OggVorbisInfo
_Tags = OggVCommentDict
_Error = OggVorbisHeaderError
_mimes = ["audio/vorbis", "audio/x-vorbis"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"OggS") * (b"\x01vorbis" in header))
Open = OggVorbis
def delete(filename):
"""Remove tags from a file."""
OggVorbis(filename).delete()

74
lib/mutagen/optimfrog.py Normal file
View File

@@ -0,0 +1,74 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Lukas Lalinsky
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""OptimFROG audio streams with APEv2 tags.
OptimFROG is a lossless audio compression program. Its main goal is to
reduce at maximum the size of audio files, while permitting bit
identical restoration for all input. It is similar with the ZIP
compression, but it is highly specialized to compress audio data.
Only versions 4.5 and higher are supported.
For more information, see http://www.losslessaudio.org/
"""
__all__ = ["OptimFROG", "Open", "delete"]
import struct
from ._compat import endswith
from mutagen import StreamInfo
from mutagen.apev2 import APEv2File, error, delete
class OptimFROGHeaderError(error):
pass
class OptimFROGInfo(StreamInfo):
"""OptimFROG stream information.
Attributes:
* channels - number of audio channels
* length - file length in seconds, as a float
* sample_rate - audio sampling rate in Hz
"""
def __init__(self, fileobj):
header = fileobj.read(76)
if (len(header) != 76 or not header.startswith(b"OFR ") or
struct.unpack("<I", header[4:8])[0] not in [12, 15]):
raise OptimFROGHeaderError("not an OptimFROG file")
(total_samples, total_samples_high, sample_type, self.channels,
self.sample_rate) = struct.unpack("<IHBBI", header[8:20])
total_samples += total_samples_high << 32
self.channels += 1
if self.sample_rate:
self.length = float(total_samples) / (self.channels *
self.sample_rate)
else:
self.length = 0.0
def pprint(self):
return "OptimFROG, %.2f seconds, %d Hz" % (self.length,
self.sample_rate)
class OptimFROG(APEv2File):
_Info = OptimFROGInfo
@staticmethod
def score(filename, fileobj, header):
filename = filename.lower()
return (header.startswith(b"OFR") + endswith(filename, b".ofr") +
endswith(filename, b".ofs"))
Open = OptimFROG

84
lib/mutagen/trueaudio.py Normal file
View File

@@ -0,0 +1,84 @@
# -*- coding: utf-8 -*-
# Copyright (C) 2006 Joe Wreschnig
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of version 2 of the GNU General Public License as
# published by the Free Software Foundation.
"""True Audio audio stream information and tags.
True Audio is a lossless format designed for real-time encoding and
decoding. This module is based on the documentation at
http://www.true-audio.com/TTA_Lossless_Audio_Codec\_-_Format_Description
True Audio files use ID3 tags.
"""
__all__ = ["TrueAudio", "Open", "delete", "EasyTrueAudio"]
from ._compat import endswith
from mutagen import StreamInfo
from mutagen.id3 import ID3FileType, delete
from mutagen._util import cdata, MutagenError
class error(RuntimeError, MutagenError):
pass
class TrueAudioHeaderError(error, IOError):
pass
class TrueAudioInfo(StreamInfo):
"""True Audio stream information.
Attributes:
* length - audio length, in seconds
* sample_rate - audio sample rate, in Hz
"""
def __init__(self, fileobj, offset):
fileobj.seek(offset or 0)
header = fileobj.read(18)
if len(header) != 18 or not header.startswith(b"TTA"):
raise TrueAudioHeaderError("TTA header not found")
self.sample_rate = cdata.int_le(header[10:14])
samples = cdata.uint_le(header[14:18])
self.length = float(samples) / self.sample_rate
def pprint(self):
return "True Audio, %.2f seconds, %d Hz." % (
self.length, self.sample_rate)
class TrueAudio(ID3FileType):
"""A True Audio file.
:ivar info: :class:`TrueAudioInfo`
:ivar tags: :class:`ID3 <mutagen.id3.ID3>`
"""
_Info = TrueAudioInfo
_mimes = ["audio/x-tta"]
@staticmethod
def score(filename, fileobj, header):
return (header.startswith(b"ID3") + header.startswith(b"TTA") +
endswith(filename.lower(), b".tta") * 2)
Open = TrueAudio
class EasyTrueAudio(TrueAudio):
"""Like MP3, but uses EasyID3 for tags.
:ivar info: :class:`TrueAudioInfo`
:ivar tags: :class:`EasyID3 <mutagen.easyid3.EasyID3>`
"""
from mutagen.easyid3 import EasyID3 as ID3
ID3 = ID3

124
lib/mutagen/wavpack.py Normal file
View File

@@ -0,0 +1,124 @@
# -*- coding: utf-8 -*-
# Copyright 2006 Joe Wreschnig
# 2014 Christoph Reiter
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""WavPack reading and writing.
WavPack is a lossless format that uses APEv2 tags. Read
* http://www.wavpack.com/
* http://www.wavpack.com/file_format.txt
for more information.
"""
__all__ = ["WavPack", "Open", "delete"]
from mutagen import StreamInfo
from mutagen.apev2 import APEv2File, error, delete
from mutagen._util import cdata
class WavPackHeaderError(error):
pass
RATES = [6000, 8000, 9600, 11025, 12000, 16000, 22050, 24000, 32000, 44100,
48000, 64000, 88200, 96000, 192000]
class _WavPackHeader(object):
def __init__(self, block_size, version, track_no, index_no, total_samples,
block_index, block_samples, flags, crc):
self.block_size = block_size
self.version = version
self.track_no = track_no
self.index_no = index_no
self.total_samples = total_samples
self.block_index = block_index
self.block_samples = block_samples
self.flags = flags
self.crc = crc
@classmethod
def from_fileobj(cls, fileobj):
"""A new _WavPackHeader or raises WavPackHeaderError"""
header = fileobj.read(32)
if len(header) != 32 or not header.startswith(b"wvpk"):
raise WavPackHeaderError("not a WavPack header: %r" % header)
block_size = cdata.uint_le(header[4:8])
version = cdata.ushort_le(header[8:10])
track_no = ord(header[10:11])
index_no = ord(header[11:12])
samples = cdata.uint_le(header[12:16])
if samples == 2 ** 32 - 1:
samples = -1
block_index = cdata.uint_le(header[16:20])
block_samples = cdata.uint_le(header[20:24])
flags = cdata.uint_le(header[24:28])
crc = cdata.uint_le(header[28:32])
return _WavPackHeader(block_size, version, track_no, index_no,
samples, block_index, block_samples, flags, crc)
class WavPackInfo(StreamInfo):
"""WavPack stream information.
Attributes:
* channels - number of audio channels (1 or 2)
* length - file length in seconds, as a float
* sample_rate - audio sampling rate in Hz
* version - WavPack stream version
"""
def __init__(self, fileobj):
try:
header = _WavPackHeader.from_fileobj(fileobj)
except WavPackHeaderError:
raise WavPackHeaderError("not a WavPack file")
self.version = header.version
self.channels = bool(header.flags & 4) or 2
self.sample_rate = RATES[(header.flags >> 23) & 0xF]
if header.total_samples == -1 or header.block_index != 0:
# TODO: we could make this faster by using the tag size
# and search backwards for the last block, then do
# last.block_index + last.block_samples - initial.block_index
samples = header.block_samples
while 1:
fileobj.seek(header.block_size - 32 + 8, 1)
try:
header = _WavPackHeader.from_fileobj(fileobj)
except WavPackHeaderError:
break
samples += header.block_samples
else:
samples = header.total_samples
self.length = float(samples) / self.sample_rate
def pprint(self):
return "WavPack, %.2f seconds, %d Hz" % (self.length, self.sample_rate)
class WavPack(APEv2File):
_Info = WavPackInfo
_mimes = ["audio/x-wavpack"]
@staticmethod
def score(filename, fileobj, header):
return header.startswith(b"wvpk") * 2
Open = WavPack