dinislage/CouchPotatoServer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Libraries for Subliminal

Browse Source
This commit is contained in:
Ruud
2011-12-23 16:02:48 +01:00
parent a03c33c1c3
commit 8efc88e691
109 changed files with 26894 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
libs/enzyme/__init__.py
+56
View File
@@ -0,0 +1,56 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
import mimetypes
import os
import sys
from exceptions import *
PARSERS = [('asf', ['video/asf'], ['asf', 'wmv', 'wma']),
('flv', ['video/flv'], ['flv']),
('mkv', ['video/x-matroska', 'application/mkv'], ['mkv', 'mka', 'webm']),
('mp4', ['video/quicktime', 'video/mp4'], ['mov', 'qt', 'mp4', 'mp4a', '3gp', '3gp2', '3g2', 'mk2']),
('mpeg', ['video/mpeg'], ['mpeg', 'mpg', 'mp4', 'ts']),
('ogm', ['application/ogg'], ['ogm', 'ogg', 'ogv']),
('real', ['video/real'], ['rm', 'ra', 'ram']),
('riff', ['video/avi'], ['wav', 'avi'])]
def parse(path):
if not os.path.isfile(path):
raise ValueError('Invalid path')
extension = os.path.splitext(path)[1][1:]
mimetype = mimetypes.guess_type(path)[0]
parser_ext = None
parser_mime = None
for (parser_name, parser_mimetypes, parser_extensions) in PARSERS:
if mimetype in parser_mimetypes:
parser_mime = parser_name
if extension in parser_extensions:
parser_ext = parser_name
parser = parser_mime or parser_ext
if not parser:
raise NoParserError()
mod = __import__(parser, globals=globals(), locals=locals(), fromlist=[], level=-1)
with open(path, 'rb') as f:
p = mod.Parser(f)
return p
libs/enzyme/asf.py
+391
View File
@@ -0,0 +1,391 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import struct
import string
import logging
from exceptions import *
import core
# get logging object
log = logging.getLogger(__name__)
def _guid(input):
# Remove any '-'
s = string.join(string.split(input,'-'), '')
r = ''
if len(s) != 32:
return ''
x = ''
for i in range(0,16):
r+=chr(int(s[2*i:2*i+2],16))
guid = struct.unpack('>IHHBB6s',r)
return guid
GUIDS = {
'ASF_Header_Object' : _guid('75B22630-668E-11CF-A6D9-00AA0062CE6C'),
'ASF_Data_Object' : _guid('75B22636-668E-11CF-A6D9-00AA0062CE6C'),
'ASF_Simple_Index_Object' : _guid('33000890-E5B1-11CF-89F4-00A0C90349CB'),
'ASF_Index_Object' : _guid('D6E229D3-35DA-11D1-9034-00A0C90349BE'),
'ASF_Media_Object_Index_Object' : _guid('FEB103F8-12AD-4C64-840F-2A1D2F7AD48C'),
'ASF_Timecode_Index_Object' : _guid('3CB73FD0-0C4A-4803-953D-EDF7B6228F0C'),
'ASF_File_Properties_Object' : _guid('8CABDCA1-A947-11CF-8EE4-00C00C205365'),
'ASF_Stream_Properties_Object' : _guid('B7DC0791-A9B7-11CF-8EE6-00C00C205365'),
'ASF_Header_Extension_Object' : _guid('5FBF03B5-A92E-11CF-8EE3-00C00C205365'),
'ASF_Codec_List_Object' : _guid('86D15240-311D-11D0-A3A4-00A0C90348F6'),
'ASF_Script_Command_Object' : _guid('1EFB1A30-0B62-11D0-A39B-00A0C90348F6'),
'ASF_Marker_Object' : _guid('F487CD01-A951-11CF-8EE6-00C00C205365'),
'ASF_Bitrate_Mutual_Exclusion_Object' : _guid('D6E229DC-35DA-11D1-9034-00A0C90349BE'),
'ASF_Error_Correction_Object' : _guid('75B22635-668E-11CF-A6D9-00AA0062CE6C'),
'ASF_Content_Description_Object' : _guid('75B22633-668E-11CF-A6D9-00AA0062CE6C'),
'ASF_Extended_Content_Description_Object' : _guid('D2D0A440-E307-11D2-97F0-00A0C95EA850'),
'ASF_Content_Branding_Object' : _guid('2211B3FA-BD23-11D2-B4B7-00A0C955FC6E'),
'ASF_Stream_Bitrate_Properties_Object' : _guid('7BF875CE-468D-11D1-8D82-006097C9A2B2'),
'ASF_Content_Encryption_Object' : _guid('2211B3FB-BD23-11D2-B4B7-00A0C955FC6E'),
'ASF_Extended_Content_Encryption_Object' : _guid('298AE614-2622-4C17-B935-DAE07EE9289C'),
'ASF_Alt_Extended_Content_Encryption_Obj' : _guid('FF889EF1-ADEE-40DA-9E71-98704BB928CE'),
'ASF_Digital_Signature_Object' : _guid('2211B3FC-BD23-11D2-B4B7-00A0C955FC6E'),
'ASF_Padding_Object' : _guid('1806D474-CADF-4509-A4BA-9AABCB96AAE8'),
'ASF_Extended_Stream_Properties_Object' : _guid('14E6A5CB-C672-4332-8399-A96952065B5A'),
'ASF_Advanced_Mutual_Exclusion_Object' : _guid('A08649CF-4775-4670-8A16-6E35357566CD'),
'ASF_Group_Mutual_Exclusion_Object' : _guid('D1465A40-5A79-4338-B71B-E36B8FD6C249'),
'ASF_Stream_Prioritization_Object' : _guid('D4FED15B-88D3-454F-81F0-ED5C45999E24'),
'ASF_Bandwidth_Sharing_Object' : _guid('A69609E6-517B-11D2-B6AF-00C04FD908E9'),
'ASF_Language_List_Object' : _guid('7C4346A9-EFE0-4BFC-B229-393EDE415C85'),
'ASF_Metadata_Object' : _guid('C5F8CBEA-5BAF-4877-8467-AA8C44FA4CCA'),
'ASF_Metadata_Library_Object' : _guid('44231C94-9498-49D1-A141-1D134E457054'),
'ASF_Index_Parameters_Object' : _guid('D6E229DF-35DA-11D1-9034-00A0C90349BE'),
'ASF_Media_Object_Index_Parameters_Obj' : _guid('6B203BAD-3F11-4E84-ACA8-D7613DE2CFA7'),
'ASF_Timecode_Index_Parameters_Object' : _guid('F55E496D-9797-4B5D-8C8B-604DFE9BFB24'),
'ASF_Audio_Media' : _guid('F8699E40-5B4D-11CF-A8FD-00805F5C442B'),
'ASF_Video_Media' : _guid('BC19EFC0-5B4D-11CF-A8FD-00805F5C442B'),
'ASF_Command_Media' : _guid('59DACFC0-59E6-11D0-A3AC-00A0C90348F6'),
'ASF_JFIF_Media' : _guid('B61BE100-5B4E-11CF-A8FD-00805F5C442B'),
'ASF_Degradable_JPEG_Media' : _guid('35907DE0-E415-11CF-A917-00805F5C442B'),
'ASF_File_Transfer_Media' : _guid('91BD222C-F21C-497A-8B6D-5AA86BFC0185'),
'ASF_Binary_Media' : _guid('3AFB65E2-47EF-40F2-AC2C-70A90D71D343'),
'ASF_Web_Stream_Media_Subtype' : _guid('776257D4-C627-41CB-8F81-7AC7FF1C40CC'),
'ASF_Web_Stream_Format' : _guid('DA1E6B13-8359-4050-B398-388E965BF00C'),
'ASF_No_Error_Correction' : _guid('20FB5700-5B55-11CF-A8FD-00805F5C442B'),
'ASF_Audio_Spread' : _guid('BFC3CD50-618F-11CF-8BB2-00AA00B4E220'),
}
class Asf(core.AVContainer):
"""
ASF video parser. The ASF format is also used for Microsft Windows
Media files like wmv.
"""
def __init__(self, file):
core.AVContainer.__init__(self)
self.mime = 'video/x-ms-asf'
self.type = 'asf format'
self._languages = []
self._extinfo = {}
h = file.read(30)
if len(h) < 30:
raise ParseError()
(guidstr, objsize, objnum, reserved1, \
reserved2) = struct.unpack('<16sQIBB',h)
guid = self._parseguid(guidstr)
if (guid != GUIDS['ASF_Header_Object']):
raise ParseError()
if reserved1 != 0x01 or reserved2 != 0x02:
raise ParseError()
log.debug("asf header size: %d / %d objects" % (objsize,objnum))
header = file.read(objsize-30)
for i in range(0,objnum):
h = self._getnextheader(header)
header = header[h[1]:]
del self._languages
del self._extinfo
def _findstream(self, id):
for stream in self.video + self.audio:
if stream.id == id:
return stream
def _apply_extinfo(self, streamid):
stream = self._findstream(streamid)
if not stream or streamid not in self._extinfo:
return
stream.bitrate, stream.fps, langid, metadata = self._extinfo[streamid]
if langid is not None and langid >= 0 and langid < len(self._languages):
stream.language = self._languages[langid]
if metadata:
stream._appendtable('ASFMETADATA', metadata)
def _parseguid(self,string):
return struct.unpack('<IHHBB6s', string[:16])
def _parsekv(self,s):
pos = 0
(descriptorlen,) = struct.unpack('<H', s[pos:pos+2])
pos += 2
descriptorname = s[pos:pos+descriptorlen]
pos += descriptorlen
descriptortype, valuelen = struct.unpack('<HH', s[pos:pos+4])
pos += 4
descriptorvalue = s[pos:pos+valuelen]
pos += valuelen
value = None
if descriptortype == 0x0000:
# Unicode string
value = descriptorvalue
elif descriptortype == 0x0001:
# Byte Array
value = descriptorvalue
elif descriptortype == 0x0002:
# Bool (?)
value = struct.unpack('<I', descriptorvalue)[0] != 0
elif descriptortype == 0x0003:
# DWORD
value = struct.unpack('<I', descriptorvalue)[0]
elif descriptortype == 0x0004:
# QWORD
value = struct.unpack('<Q', descriptorvalue)[0]
elif descriptortype == 0x0005:
# WORD
value = struct.unpack('<H', descriptorvalue)[0]
else:
log.debug("Unknown Descriptor Type %d" % descriptortype)
return (pos,descriptorname,value)
def _parsekv2(self,s):
pos = 0
strno, descriptorlen, descriptortype, valuelen = struct.unpack('<2xHHHI', s[pos:pos+12])
pos += 12
descriptorname = s[pos:pos+descriptorlen]
pos += descriptorlen
descriptorvalue = s[pos:pos+valuelen]
pos += valuelen
value = None
if descriptortype == 0x0000:
# Unicode string
value = descriptorvalue
elif descriptortype == 0x0001:
# Byte Array
value = descriptorvalue
elif descriptortype == 0x0002:
# Bool
value = struct.unpack('<H', descriptorvalue)[0] != 0
pass
elif descriptortype == 0x0003:
# DWORD
value = struct.unpack('<I', descriptorvalue)[0]
elif descriptortype == 0x0004:
# QWORD
value = struct.unpack('<Q', descriptorvalue)[0]
elif descriptortype == 0x0005:
# WORD
value = struct.unpack('<H', descriptorvalue)[0]
else:
log.debug("Unknown Descriptor Type %d" % descriptortype)
return (pos,descriptorname,value,strno)
def _getnextheader(self,s):
r = struct.unpack('<16sQ',s[:24])
(guidstr,objsize) = r
guid = self._parseguid(guidstr)
if guid == GUIDS['ASF_File_Properties_Object']:
log.debug("File Properties Object")
val = struct.unpack('<16s6Q4I',s[24:24+80])
(fileid, size, date, packetcount, duration, \
senddur, preroll, flags, minpack, maxpack, maxbr) = \
val
# FIXME: parse date to timestamp
self.length = duration/10000000.0
elif guid == GUIDS['ASF_Stream_Properties_Object']:
log.debug("Stream Properties Object [%d]" % objsize)
streamtype = self._parseguid(s[24:40])
errortype = self._parseguid(s[40:56])
offset, typelen, errorlen, flags = struct.unpack('<QIIH', s[56:74])
strno = flags & 0x7f
encrypted = flags >> 15
if encrypted:
self._set('encrypted', True)
if streamtype == GUIDS['ASF_Video_Media']:
vi = core.VideoStream()
vi.width, vi.height, depth, codec, = struct.unpack('<4xII2xH4s', s[89:89+20])
vi.codec = codec
vi.id = strno
self.video.append(vi)
elif streamtype == GUIDS['ASF_Audio_Media']:
ai = core.AudioStream()
twocc, ai.channels, ai.samplerate, bitrate, block, \
ai.samplebits, = struct.unpack('<HHIIHH', s[78:78+16])
ai.bitrate = 8*bitrate
ai.codec = twocc
ai.id = strno
self.audio.append(ai)
self._apply_extinfo(strno)
elif guid == GUIDS['ASF_Extended_Stream_Properties_Object']:
streamid, langid, frametime = struct.unpack('<HHQ',s[72:84])
(bitrate,) = struct.unpack('<I', s[40:40+4])
if streamid not in self._extinfo:
self._extinfo[streamid] = [None, None, None, {}]
if frametime == 0:
# Problaby VFR, report as 1000fps (which is what MPlayer does)
frametime = 10000.0
self._extinfo[streamid][:3] = [bitrate, 10000000.0 / frametime, langid]
self._apply_extinfo(streamid)
elif guid == GUIDS['ASF_Header_Extension_Object']:
log.debug("ASF_Header_Extension_Object %d" % objsize)
size = struct.unpack('<I',s[42:46])[0]
data = s[46:46+size]
while len(data):
log.debug("Sub:")
h = self._getnextheader(data)
data = data[h[1]:]
elif guid == GUIDS['ASF_Codec_List_Object']:
log.debug("List Object")
pass
elif guid == GUIDS['ASF_Error_Correction_Object']:
log.debug("Error Correction")
pass
elif guid == GUIDS['ASF_Content_Description_Object']:
log.debug("Content Description Object")
val = struct.unpack('<5H', s[24:24+10])
pos = 34
strings = []
for i in val:
ss = s[pos:pos+i].replace('\0', '').lstrip().rstrip()
strings.append(ss)
pos+=i
# Set empty strings to None
strings = [x or None for x in strings]
self.title, self.artist, self.copyright, self.caption, rating = strings
elif guid == GUIDS['ASF_Extended_Content_Description_Object']:
(count,) = struct.unpack('<H', s[24:26])
pos = 26
descriptor = {}
for i in range(0, count):
# Read additional content descriptors
d = self._parsekv(s[pos:])
pos += d[0]
descriptor[d[1]] = d[2]
self._appendtable('ASFDESCRIPTOR', descriptor)
elif guid == GUIDS['ASF_Metadata_Object']:
(count,) = struct.unpack('<H', s[24:26])
pos = 26
streams = {}
for i in range(0, count):
# Read additional content descriptors
size,key,value,strno = self._parsekv2(s[pos:])
if strno not in streams:
streams[strno] = {}
streams[strno][key] = value
pos += size
for strno, metadata in streams.items():
if strno not in self._extinfo:
self._extinfo[strno] = [None, None, None, {}]
self._extinfo[strno][3].update(metadata)
self._apply_extinfo(strno)
elif guid == GUIDS['ASF_Language_List_Object']:
count = struct.unpack('<H', s[24:26])[0]
pos = 26
for i in range(0, count):
idlen = struct.unpack('<B', s[pos:pos+1])[0]
idstring = s[pos+1:pos+1+idlen]
idstring = unicode(idstring, 'utf-16').replace('\0', '')
log.debug("Language: %d/%d: %s" % (i+1, count, idstring))
self._languages.append(idstring)
pos += 1+idlen
elif guid == GUIDS['ASF_Stream_Bitrate_Properties_Object']:
# This record contains stream bitrate with payload overhead. For
# audio streams, we should have the average bitrate from
# ASF_Stream_Properties_Object. For video streams, we get it from
# ASF_Extended_Stream_Properties_Object. So this record is not
# used.
pass
elif guid == GUIDS['ASF_Content_Encryption_Object'] or \
guid == GUIDS['ASF_Extended_Content_Encryption_Object']:
self._set('encrypted', True)
else:
# Just print the type:
for h in GUIDS.keys():
if GUIDS[h] == guid:
log.debug("Unparsed %s [%d]" % (h,objsize))
break
else:
u = "%.8X-%.4X-%.4X-%.2X%.2X-%s" % guid
log.debug("unknown: len=%d [%d]" % (len(u), objsize))
return r
class AsfAudio(core.AudioStream):
"""
ASF audio parser for wma files.
"""
def __init__(self):
core.AudioStream.__init__(self)
self.mime = 'audio/x-ms-asf'
self.type = 'asf format'
def Parser(file):
"""
Wrapper around audio and av content.
"""
asf = Asf(file)
if not len(asf.audio) or len(asf.video):
# AV container
return asf
# No video but audio streams. Handle has audio core
audio = AsfAudio()
for key in audio._keys:
if key in asf._keys:
if not getattr(audio, key, None):
setattr(audio, key, getattr(asf, key))
return audio
libs/enzyme/core.py
+457
View File
@@ -0,0 +1,457 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
import re
import logging
import fourcc
import language
from exceptions import *
from strutils import str_to_unicode, unicode_to_str
UNPRINTABLE_KEYS = ['thumbnail', 'url', 'codec_private']
EXTENSION_DEVICE = 'device'
EXTENSION_DIRECTORY = 'directory'
EXTENSION_STREAM = 'stream'
MEDIACORE = ['title', 'caption', 'comment', 'size', 'type', 'subtype', 'timestamp',
'keywords', 'country', 'language', 'langcode', 'url', 'artist',
'mime', 'datetime', 'tags', 'hash']
AUDIOCORE = ['channels', 'samplerate', 'length', 'encoder', 'codec', 'format',
'samplebits', 'bitrate', 'fourcc', 'trackno', 'id', 'userdate',
'enabled', 'default', 'codec_private']
MUSICCORE = ['trackof', 'album', 'genre', 'discs', 'thumbnail']
VIDEOCORE = ['length', 'encoder', 'bitrate', 'samplerate', 'codec', 'format',
'samplebits', 'width', 'height', 'fps', 'aspect', 'trackno',
'fourcc', 'id', 'enabled', 'default', 'codec_private']
AVCORE = ['length', 'encoder', 'trackno', 'trackof', 'copyright', 'product',
'genre', 'writer', 'producer', 'studio', 'rating', 'actors', 'thumbnail',
'delay', 'image', 'video', 'audio', 'subtitles', 'chapters', 'software',
'summary', 'synopsis', 'season', 'episode', 'series']
# get logging object
log = logging.getLogger(__name__)
class Media(object):
media = None
"""
Media is the base class to all Media Metadata Containers. It defines
the basic structures that handle metadata. Media and its derivates
contain a common set of metadata attributes that is listed in keys.
Specific derivates contain additional keys to the dublin core set that is
defined in Media.
"""
_keys = MEDIACORE
table_mapping = {}
def __init__(self, hash=None):
if hash is not None:
# create Media based on dict
for key, value in hash.items():
if isinstance(value, list) and value and isinstance(value[0], dict):
value = [Media(x) for x in value]
self._set(key, value)
return
self._keys = self._keys[:]
self.tables = {}
# Tags, unlike tables, are more well-defined dicts whose values are
# either Tag objects, other dicts (for nested tags), or lists of either
# (for multiple instances of the tag, e.g. actor). Where possible,
# parsers should transform tag names to conform to the Official
# Matroska tags defined at http://www.matroska.org/technical/specs/tagging/index.html
# All tag names will be lower-cased.
self.tags = Tags()
for key in set(self._keys) - set(['media', 'tags']):
setattr(self, key, None)
#
# unicode and string convertion for debugging
#
#TODO: Fix that mess
def __unicode__(self):
result = u''
# print normal attributes
lists = []
for key in self._keys:
value = getattr(self, key, None)
if value == None or key == 'url':
continue
if isinstance(value, list):
if not value:
continue
elif isinstance(value[0], basestring):
# Just a list of strings (keywords?), so don't treat it specially.
value = u', '.join(value)
else:
lists.append((key, value))
continue
elif isinstance(value, dict):
# Tables or tags treated separately.
continue
if key in UNPRINTABLE_KEYS:
value = '<unprintable data, size=%d>' % len(value)
result += u'| %10s: %s\n' % (unicode(key), unicode(value))
# print tags (recursively, to support nested tags).
def print_tags(tags, suffix, show_label):
result = ''
for n, (name, tag) in enumerate(tags.items()):
result += u'| %12s%s%s = ' % (u'tags: ' if n == 0 and show_label else '', suffix, name)
if isinstance(tag, list):
# TODO: doesn't support lists/dicts within lists.
result += u'%s\n' % ', '.join(subtag.value for subtag in tag)
else:
result += u'%s\n' % (tag.value or '')
if isinstance(tag, dict):
result += print_tags(tag, ' ', False)
return result
result += print_tags(self.tags, '', True)
# print lists
for key, l in lists:
for n, item in enumerate(l):
label = '+-- ' + key.rstrip('s').capitalize()
if key not in ['tracks', 'subtitles', 'chapters']:
label += ' Track'
result += u'%s #%d\n' % (label, n + 1)
result += '| ' + re.sub(r'\n(.)', r'\n| \1', unicode(item))
# print tables
if log.level >= 10:
for name, table in self.tables.items():
result += '+-- Table %s\n' % str(name)
for key, value in table.items():
try:
value = unicode(value)
if len(value) > 50:
value = u'<unprintable data, size=%d>' % len(value)
except (UnicodeDecodeError, TypeError), e:
try:
value = u'<unprintable data, size=%d>' % len(value)
except AttributeError:
value = u'<unprintable data>'
result += u'| | %s: %s\n' % (unicode(key), value)
return result
def __str__(self):
return unicode(self).encode()
def __repr__(self):
if hasattr(self, 'url'):
return '<%s %s>' % (str(self.__class__)[8:-2], self.url)
else:
return '<%s>' % (str(self.__class__)[8:-2])
#
# internal functions
#
def _appendtable(self, name, hashmap):
"""
Appends a tables of additional metadata to the Object.
If such a table already exists, the given tables items are
added to the existing one.
"""
if name not in self.tables:
self.tables[name] = hashmap
else:
# Append to the already existing table
for k in hashmap.keys():
self.tables[name][k] = hashmap[k]
def _set(self, key, value):
"""
Set key to value and add the key to the internal keys list if
missing.
"""
if value is None and getattr(self, key, None) is None:
return
if isinstance(value, str):
value = str_to_unicode(value)
setattr(self, key, value)
if not key in self._keys:
self._keys.append(key)
def _set_url(self, url):
"""
Set the URL of the source
"""
self.url = url
def _finalize(self):
"""
Correct same data based on specific rules
"""
# make sure all strings are unicode
for key in self._keys:
if key in UNPRINTABLE_KEYS:
continue
value = getattr(self, key)
if value is None:
continue
if key == 'image':
if isinstance(value, unicode):
setattr(self, key, unicode_to_str(value))
continue
if isinstance(value, str):
setattr(self, key, str_to_unicode(value))
if isinstance(value, unicode):
setattr(self, key, value.strip().rstrip().replace(u'\0', u''))
if isinstance(value, list) and value and isinstance(value[0], Media):
for submenu in value:
submenu._finalize()
# copy needed tags from tables
for name, table in self.tables.items():
mapping = self.table_mapping.get(name, {})
for tag, attr in mapping.items():
if self.get(attr):
continue
value = table.get(tag, None)
if value is not None:
if not isinstance(value, (str, unicode)):
value = str_to_unicode(str(value))
elif isinstance(value, str):
value = str_to_unicode(value)
value = value.strip().rstrip().replace(u'\0', u'')
setattr(self, attr, value)
if 'fourcc' in self._keys and 'codec' in self._keys and self.codec is not None:
# Codec may be a fourcc, in which case we resolve it to its actual
# name and set the fourcc attribute.
self.fourcc, self.codec = fourcc.resolve(self.codec)
if 'language' in self._keys:
self.langcode, self.language = language.resolve(self.language)
#
# data access
#
def __contains__(self, key):
"""
Test if key exists in the dict
"""
return hasattr(self, key)
def get(self, attr, default=None):
"""
Returns the given attribute. If the attribute is not set by
the parser return 'default'.
"""
return getattr(self, attr, default)
def __getitem__(self, attr):
"""
Get the value of the given attribute
"""
return getattr(self, attr, None)
def __setitem__(self, key, value):
"""
Set the value of 'key' to 'value'
"""
setattr(self, key, value)
def has_key(self, key):
"""
Check if the object has an attribute 'key'
"""
return hasattr(self, key)
def convert(self):
"""
Convert Media to dict.
"""
result = {}
for k in self._keys:
value = getattr(self, k, None)
if isinstance(value, list) and value and isinstance(value[0], Media):
value = [x.convert() for x in value]
result[k] = value
return result
def keys(self):
"""
Return all keys for the attributes set by the parser.
"""
return self._keys
class Collection(Media):
"""
Collection of Digial Media like CD, DVD, Directory, Playlist
"""
_keys = Media._keys + ['id', 'tracks']
def __init__(self):
Media.__init__(self)
self.tracks = []
class Tag(object):
"""
An individual tag, which will be a value stored in a Tags object.
Tag values are strings (for binary data), unicode objects, or datetime
objects for tags that represent dates or times.
"""
def __init__(self, value=None, langcode='und', binary=False):
super(Tag, self).__init__()
self.value = value
self.langcode = langcode
self.binary = binary
def __unicode__(self):
return unicode(self.value)
def __str__(self):
return str(self.value)
def __repr__(self):
if not self.binary:
return '<Tag object: %s>' % repr(self.value)
else:
return '<Binary Tag object: size=%d>' % len(self.value)
@property
def langcode(self):
return self._langcode
@langcode.setter
def langcode(self, code):
self._langcode, self.language = language.resolve(code)
class Tags(dict, Tag):
"""
A dictionary containing Tag objects. Values can be other Tags objects
(for nested tags), lists, or Tag objects.
A Tags object is more or less a dictionary but it also contains a value.
This is necessary in order to represent this kind of tag specification
(e.g. for Matroska)::
<Simple>
<Name>LAW_RATING</Name>
<String>PG</String>
<Simple>
<Name>COUNTRY</Name>
<String>US</String>
</Simple>
</Simple>
The attribute RATING has a value (PG), but it also has a child tag
COUNTRY that specifies the country code the rating belongs to.
"""
def __init__(self, value=None, langcode='und', binary=False):
super(Tags, self).__init__()
self.value = value
self.langcode = langcode
self.binary = False
class AudioStream(Media):
"""
Audio Tracks in a Multiplexed Container.
"""
_keys = Media._keys + AUDIOCORE
class Music(AudioStream):
"""
Digital Music.
"""
_keys = AudioStream._keys + MUSICCORE
def _finalize(self):
"""
Correct same data based on specific rules
"""
AudioStream._finalize(self)
if self.trackof:
try:
# XXX Why is this needed anyway?
if int(self.trackno) < 10:
self.trackno = u'0%s' % int(self.trackno)
except (AttributeError, ValueError):
pass
class VideoStream(Media):
"""
Video Tracks in a Multiplexed Container.
"""
_keys = Media._keys + VIDEOCORE
class Chapter(Media):
"""
Chapter in a Multiplexed Container.
"""
_keys = ['enabled', 'name', 'pos', 'id']
def __init__(self, name=None, pos=0):
Media.__init__(self)
self.name = name
self.pos = pos
self.enabled = True
class Subtitle(Media):
"""
Subtitle Tracks in a Multiplexed Container.
"""
_keys = ['enabled', 'default', 'langcode', 'language', 'trackno', 'title',
'id', 'codec']
def __init__(self, language=None):
Media.__init__(self)
self.language = language
class AVContainer(Media):
"""
Container for Audio and Video streams. This is the Container Type for
all media, that contain more than one stream.
"""
_keys = Media._keys + AVCORE
def __init__(self):
Media.__init__(self)
self.audio = []
self.video = []
self.subtitles = []
self.chapters = []
def _finalize(self):
"""
Correct same data based on specific rules
"""
Media._finalize(self)
if not self.length and len(self.video) and self.video[0].length:
self.length = 0
# Length not specified for container, so use the largest length
# of its tracks as container length.
for track in self.video + self.audio:
if track.length:
self.length = max(self.length, track.length)
libs/enzyme/exceptions.py
+31
View File
@@ -0,0 +1,31 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
class Error(Exception):
pass
class NoParserError(Error):
pass
class ParseError(Error):
pass
libs/enzyme/flv.py
+182
View File
@@ -0,0 +1,182 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
from exceptions import *
import core
import logging
import struct
# get logging object
log = logging.getLogger(__name__)
FLV_TAG_TYPE_AUDIO = 0x08
FLV_TAG_TYPE_VIDEO = 0x09
FLV_TAG_TYPE_META = 0x12
# audio flags
FLV_AUDIO_CHANNEL_MASK = 0x01
FLV_AUDIO_SAMPLERATE_MASK = 0x0c
FLV_AUDIO_CODECID_MASK = 0xf0
FLV_AUDIO_SAMPLERATE_OFFSET = 2
FLV_AUDIO_CODECID_OFFSET = 4
FLV_AUDIO_CODECID = (0x0001, 0x0002, 0x0055, 0x0001)
# video flags
FLV_VIDEO_CODECID_MASK = 0x0f
FLV_VIDEO_CODECID = ( 'FLV1', 'MSS1', 'VP60') # wild guess
FLV_DATA_TYPE_NUMBER = 0x00
FLV_DATA_TYPE_BOOL = 0x01
FLV_DATA_TYPE_STRING = 0x02
FLV_DATA_TYPE_OBJECT = 0x03
FLC_DATA_TYPE_CLIP = 0x04
FLV_DATA_TYPE_REFERENCE = 0x07
FLV_DATA_TYPE_ECMARRAY = 0x08
FLV_DATA_TYPE_ENDOBJECT = 0x09
FLV_DATA_TYPE_ARRAY = 0x0a
FLV_DATA_TYPE_DATE = 0x0b
FLV_DATA_TYPE_LONGSTRING = 0x0c
FLVINFO = {
'creator': 'copyright',
}
class FlashVideo(core.AVContainer):
"""
Experimental parser for Flash videos. It requires certain flags to
be set to report video resolutions and in most cases it does not
provide that information.
"""
table_mapping = { 'FLVINFO' : FLVINFO }
def __init__(self,file):
core.AVContainer.__init__(self)
self.mime = 'video/flv'
self.type = 'Flash Video'
data = file.read(13)
if len(data) < 13 or struct.unpack('>3sBBII', data)[0] != 'FLV':
raise ParseError()
for i in range(10):
if self.audio and self.video:
break
data = file.read(11)
if len(data) < 11:
break
chunk = struct.unpack('>BH4BI', data)
size = (chunk[1] << 8) + chunk[2]
if chunk[0] == FLV_TAG_TYPE_AUDIO:
flags = ord(file.read(1))
if not self.audio:
a = core.AudioStream()
a.channels = (flags & FLV_AUDIO_CHANNEL_MASK) + 1
srate = (flags & FLV_AUDIO_SAMPLERATE_MASK)
a.samplerate = (44100 << (srate >> FLV_AUDIO_SAMPLERATE_OFFSET) >> 3)
codec = (flags & FLV_AUDIO_CODECID_MASK) >> FLV_AUDIO_CODECID_OFFSET
if codec < len(FLV_AUDIO_CODECID):
a.codec = FLV_AUDIO_CODECID[codec]
self.audio.append(a)
file.seek(size - 1, 1)
elif chunk[0] == FLV_TAG_TYPE_VIDEO:
flags = ord(file.read(1))
if not self.video:
v = core.VideoStream()
codec = (flags & FLV_VIDEO_CODECID_MASK) - 2
if codec < len(FLV_VIDEO_CODECID):
v.codec = FLV_VIDEO_CODECID[codec]
# width and height are in the meta packet, but I have
# no file with such a packet inside. So maybe we have
# to decode some parts of the video.
self.video.append(v)
file.seek(size - 1, 1)
elif chunk[0] == FLV_TAG_TYPE_META:
log.info('metadata %s', str(chunk))
metadata = file.read(size)
try:
while metadata:
length, value = self._parse_value(metadata)
if isinstance(value, dict):
log.info('metadata: %s', value)
if value.get('creator'):
self.copyright = value.get('creator')
if value.get('width'):
self.width = value.get('width')
if value.get('height'):
self.height = value.get('height')
if value.get('duration'):
self.length = value.get('duration')
self._appendtable('FLVINFO', value)
if not length:
# parse error
break
metadata = metadata[length:]
except (IndexError, struct.error, TypeError):
pass
else:
log.info('unkown %s', str(chunk))
file.seek(size, 1)
file.seek(4, 1)
def _parse_value(self, data):
"""
Parse the next metadata value.
"""
if ord(data[0]) == FLV_DATA_TYPE_NUMBER:
value = struct.unpack('>d', data[1:9])[0]
return 9, value
if ord(data[0]) == FLV_DATA_TYPE_BOOL:
return 2, bool(data[1])
if ord(data[0]) == FLV_DATA_TYPE_STRING:
length = (ord(data[1]) << 8) + ord(data[2])
return length + 3, data[3:length+3]
if ord(data[0]) == FLV_DATA_TYPE_ECMARRAY:
init_length = len(data)
num = struct.unpack('>I', data[1:5])[0]
data = data[5:]
result = {}
for i in range(num):
length = (ord(data[0]) << 8) + ord(data[1])
key = data[2:length+2]
data = data[length + 2:]
length, value = self._parse_value(data)
if not length:
return 0, result
result[key] = value
data = data[length:]
return init_length - len(data), result
log.info('unknown code: %x. Stop metadata parser', ord(data[0]))
return 0, None
Parser = FlashVideo
libs/enzyme/fourcc.py
+853
View File
@@ -0,0 +1,853 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
import string
import re
import struct
__all__ = ['resolve']
def resolve(code):
"""
Transform a twocc or fourcc code into a name. Returns a 2-tuple of (cc,
codec) where both are strings and cc is a string in the form '0xXX' if it's
a twocc, or 'ABCD' if it's a fourcc. If the given code is not a known
twocc or fourcc, the return value will be (None, 'Unknown'), unless the
code is otherwise a printable string in which case it will be returned as
the codec.
"""
if isinstance(code, basestring):
codec = u'Unknown'
# Check for twocc
if re.match(r'^0x[\da-f]{1,4}$', code, re.I):
# Twocc in hex form
return code, TWOCC.get(int(code, 16), codec)
elif code.isdigit() and 0 <= int(code) <= 0xff:
# Twocc in decimal form
return hex(int(code)), TWOCC.get(int(code), codec)
elif len(code) == 2:
code = struct.unpack('H', code)[0]
return hex(code), TWOCC.get(code, codec)
elif len(code) != 4 and len([x for x in code if x not in string.printable]) == 0:
# Code is a printable string.
codec = unicode(code)
if code[:2] == 'MS' and code[2:].upper() in FOURCC:
code = code[2:]
if code.upper() in FOURCC:
return code.upper(), unicode(FOURCC[code.upper()])
return None, codec
elif isinstance(code, (int, long)):
return hex(code), TWOCC.get(code, u'Unknown')
return None, u'Unknown'
TWOCC = {
0x0000: 'Unknown Wave Format',
0x0001: 'PCM',
0x0002: 'Microsoft ADPCM',
0x0003: 'IEEE Float',
0x0004: 'Compaq Computer VSELP',
0x0005: 'IBM CVSD',
0x0006: 'A-Law',
0x0007: 'mu-Law',
0x0008: 'Microsoft DTS',
0x0009: 'Microsoft DRM',
0x0010: 'OKI ADPCM',
0x0011: 'Intel DVI/IMA ADPCM',
0x0012: 'Videologic MediaSpace ADPCM',
0x0013: 'Sierra Semiconductor ADPCM',
0x0014: 'Antex Electronics G.723 ADPCM',
0x0015: 'DSP Solutions DigiSTD',
0x0016: 'DSP Solutions DigiFIX',
0x0017: 'Dialogic OKI ADPCM',
0x0018: 'MediaVision ADPCM',
0x0019: 'Hewlett-Packard CU',
0x0020: 'Yamaha ADPCM',
0x0021: 'Speech Compression Sonarc',
0x0022: 'DSP Group TrueSpeech',
0x0023: 'Echo Speech EchoSC1',
0x0024: 'Audiofile AF36',
0x0025: 'Audio Processing Technology APTX',
0x0026: 'AudioFile AF10',
0x0027: 'Prosody 1612',
0x0028: 'LRC',
0x0030: 'Dolby AC2',
0x0031: 'Microsoft GSM 6.10',
0x0032: 'MSNAudio',
0x0033: 'Antex Electronics ADPCME',
0x0034: 'Control Resources VQLPC',
0x0035: 'DSP Solutions DigiREAL',
0x0036: 'DSP Solutions DigiADPCM',
0x0037: 'Control Resources CR10',
0x0038: 'Natural MicroSystems VBXADPCM',
0x0039: 'Crystal Semiconductor IMA ADPCM',
0x003A: 'EchoSC3',
0x003B: 'Rockwell ADPCM',
0x003C: 'Rockwell Digit LK',
0x003D: 'Xebec',
0x0040: 'Antex Electronics G.721 ADPCM',
0x0041: 'G.728 CELP',
0x0042: 'MSG723',
0x0043: 'IBM AVC ADPCM',
0x0045: 'ITU-T G.726 ADPCM',
0x0050: 'MPEG 1, Layer 1,2',
0x0052: 'RT24',
0x0053: 'PAC',
0x0055: 'MPEG Layer 3',
0x0059: 'Lucent G.723',
0x0060: 'Cirrus',
0x0061: 'ESPCM',
0x0062: 'Voxware',
0x0063: 'Canopus Atrac',
0x0064: 'G.726 ADPCM',
0x0065: 'G.722 ADPCM',
0x0066: 'DSAT',
0x0067: 'DSAT Display',
0x0069: 'Voxware Byte Aligned',
0x0070: 'Voxware AC8',
0x0071: 'Voxware AC10',
0x0072: 'Voxware AC16',
0x0073: 'Voxware AC20',
0x0074: 'Voxware MetaVoice',
0x0075: 'Voxware MetaSound',
0x0076: 'Voxware RT29HW',
0x0077: 'Voxware VR12',
0x0078: 'Voxware VR18',
0x0079: 'Voxware TQ40',
0x0080: 'Softsound',
0x0081: 'Voxware TQ60',
0x0082: 'MSRT24',
0x0083: 'G.729A',
0x0084: 'MVI MV12',
0x0085: 'DF G.726',
0x0086: 'DF GSM610',
0x0088: 'ISIAudio',
0x0089: 'Onlive',
0x0091: 'SBC24',
0x0092: 'Dolby AC3 SPDIF',
0x0093: 'MediaSonic G.723',
0x0094: 'Aculab PLC Prosody 8KBPS',
0x0097: 'ZyXEL ADPCM',
0x0098: 'Philips LPCBB',
0x0099: 'Packed',
0x00A0: 'Malden Electronics PHONYTALK',
0x00FF: 'AAC',
0x0100: 'Rhetorex ADPCM',
0x0101: 'IBM mu-law',
0x0102: 'IBM A-law',
0x0103: 'IBM AVC Adaptive Differential Pulse Code Modulation',
0x0111: 'Vivo G.723',
0x0112: 'Vivo Siren',
0x0123: 'Digital G.723',
0x0125: 'Sanyo LD ADPCM',
0x0130: 'Sipro Lab Telecom ACELP.net',
0x0131: 'Sipro Lab Telecom ACELP.4800',
0x0132: 'Sipro Lab Telecom ACELP.8V3',
0x0133: 'Sipro Lab Telecom ACELP.G.729',
0x0134: 'Sipro Lab Telecom ACELP.G.729A',
0x0135: 'Sipro Lab Telecom ACELP.KELVIN',
0x0140: 'Windows Media Video V8',
0x0150: 'Qualcomm PureVoice',
0x0151: 'Qualcomm HalfRate',
0x0155: 'Ring Zero Systems TUB GSM',
0x0160: 'Windows Media Audio V1 / DivX audio (WMA)',
0x0161: 'Windows Media Audio V7 / V8 / V9',
0x0162: 'Windows Media Audio Professional V9',
0x0163: 'Windows Media Audio Lossless V9',
0x0170: 'UNISYS NAP ADPCM',
0x0171: 'UNISYS NAP ULAW',
0x0172: 'UNISYS NAP ALAW',
0x0173: 'UNISYS NAP 16K',
0x0200: 'Creative Labs ADPCM',
0x0202: 'Creative Labs Fastspeech8',
0x0203: 'Creative Labs Fastspeech10',
0x0210: 'UHER Informatic ADPCM',
0x0215: 'Ulead DV ACM',
0x0216: 'Ulead DV ACM',
0x0220: 'Quarterdeck',
0x0230: 'I-link Worldwide ILINK VC',
0x0240: 'Aureal Semiconductor RAW SPORT',
0x0241: 'ESST AC3',
0x0250: 'Interactive Products HSX',
0x0251: 'Interactive Products RPELP',
0x0260: 'Consistent Software CS2',
0x0270: 'Sony ATRAC3 (SCX, same as MiniDisk LP2)',
0x0300: 'Fujitsu FM Towns Snd',
0x0400: 'BTV Digital',
0x0401: 'Intel Music Coder (IMC)',
0x0402: 'Ligos Indeo Audio',
0x0450: 'QDesign Music',
0x0680: 'VME VMPCM',
0x0681: 'AT&T Labs TPC',
0x0700: 'YMPEG Alpha',
0x08AE: 'ClearJump LiteWave',
0x1000: 'Olivetti GSM',
0x1001: 'Olivetti ADPCM',
0x1002: 'Olivetti CELP',
0x1003: 'Olivetti SBC',
0x1004: 'Olivetti OPR',
0x1100: 'Lernout & Hauspie LH Codec',
0x1101: 'Lernout & Hauspie CELP codec',
0x1102: 'Lernout & Hauspie SBC codec',
0x1103: 'Lernout & Hauspie SBC codec',
0x1104: 'Lernout & Hauspie SBC codec',
0x1400: 'Norris',
0x1401: 'AT&T ISIAudio',
0x1500: 'Soundspace Music Compression',
0x181C: 'VoxWare RT24 speech codec',
0x181E: 'Lucent elemedia AX24000P Music codec',
0x1C07: 'Lucent SX8300P speech codec',
0x1C0C: 'Lucent SX5363S G.723 compliant codec',
0x1F03: 'CUseeMe DigiTalk (ex-Rocwell)',
0x1FC4: 'NCT Soft ALF2CD ACM',
0x2000: 'AC3',
0x2001: 'Dolby DTS (Digital Theater System)',
0x2002: 'RealAudio 1 / 2 14.4',
0x2003: 'RealAudio 1 / 2 28.8',
0x2004: 'RealAudio G2 / 8 Cook (low bitrate)',
0x2005: 'RealAudio 3 / 4 / 5 Music (DNET)',
0x2006: 'RealAudio 10 AAC (RAAC)',
0x2007: 'RealAudio 10 AAC+ (RACP)',
0x3313: 'makeAVIS',
0x4143: 'Divio MPEG-4 AAC audio',
0x434C: 'LEAD Speech',
0x564C: 'LEAD Vorbis',
0x674F: 'Ogg Vorbis (mode 1)',
0x6750: 'Ogg Vorbis (mode 2)',
0x6751: 'Ogg Vorbis (mode 3)',
0x676F: 'Ogg Vorbis (mode 1+)',
0x6770: 'Ogg Vorbis (mode 2+)',
0x6771: 'Ogg Vorbis (mode 3+)',
0x7A21: 'GSM-AMR (CBR, no SID)',
0x7A22: 'GSM-AMR (VBR, including SID)',
0xDFAC: 'DebugMode SonicFoundry Vegas FrameServer ACM Codec',
0xF1AC: 'Free Lossless Audio Codec FLAC',
0xFFFE: 'Extensible wave format',
0xFFFF: 'development'
}
FOURCC = {
'1978': 'A.M.Paredes predictor (LossLess)',
'2VUY': 'Optibase VideoPump 8-bit 4:2:2 Component YCbCr',
'3IV0': 'MPEG4-based codec 3ivx',
'3IV1': '3ivx v1',
'3IV2': '3ivx v2',
'3IVD': 'FFmpeg DivX ;-) (MS MPEG-4 v3)',
'3IVX': 'MPEG4-based codec 3ivx',
'8BPS': 'Apple QuickTime Planar RGB with Alpha-channel',
'AAS4': 'Autodesk Animator codec (RLE)',
'AASC': 'Autodesk Animator',
'ABYR': 'Kensington ABYR',
'ACTL': 'Streambox ACT-L2',
'ADV1': 'Loronix WaveCodec',
'ADVJ': 'Avid M-JPEG Avid Technology Also known as AVRn',
'AEIK': 'Intel Indeo Video 3.2',
'AEMI': 'Array VideoONE MPEG1-I Capture',
'AFLC': 'Autodesk Animator FLC',
'AFLI': 'Autodesk Animator FLI',
'AHDV': 'CineForm 10-bit Visually Perfect HD',
'AJPG': '22fps JPEG-based codec for digital cameras',
'AMPG': 'Array VideoONE MPEG',
'ANIM': 'Intel RDX (ANIM)',
'AP41': 'AngelPotion Definitive',
'AP42': 'AngelPotion Definitive',
'ASLC': 'AlparySoft Lossless Codec',
'ASV1': 'Asus Video v1',
'ASV2': 'Asus Video v2',
'ASVX': 'Asus Video 2.0 (audio)',
'ATM4': 'Ahead Nero Digital MPEG-4 Codec',
'AUR2': 'Aura 2 Codec - YUV 4:2:2',
'AURA': 'Aura 1 Codec - YUV 4:1:1',
'AV1X': 'Avid 1:1x (Quick Time)',
'AVC1': 'H.264 AVC',
'AVD1': 'Avid DV (Quick Time)',
'AVDJ': 'Avid Meridien JFIF with Alpha-channel',
'AVDN': 'Avid DNxHD (Quick Time)',
'AVDV': 'Avid DV',
'AVI1': 'MainConcept Motion JPEG Codec',
'AVI2': 'MainConcept Motion JPEG Codec',
'AVID': 'Avid Motion JPEG',
'AVIS': 'Wrapper for AviSynth',
'AVMP': 'Avid IMX (Quick Time)',
'AVR ': 'Avid ABVB/NuVista MJPEG with Alpha-channel',
'AVRN': 'Avid Motion JPEG',
'AVUI': 'Avid Meridien Uncompressed with Alpha-channel',
'AVUP': 'Avid 10bit Packed (Quick Time)',
'AYUV': '4:4:4 YUV (AYUV)',
'AZPR': 'Quicktime Apple Video',
'AZRP': 'Quicktime Apple Video',
'BGR ': 'Uncompressed BGR32 8:8:8:8',
'BGR(15)': 'Uncompressed BGR15 5:5:5',
'BGR(16)': 'Uncompressed BGR16 5:6:5',
'BGR(24)': 'Uncompressed BGR24 8:8:8',
'BHIV': 'BeHere iVideo',
'BINK': 'RAD Game Tools Bink Video',
'BIT ': 'BI_BITFIELDS (Raw RGB)',
'BITM': 'Microsoft H.261',
'BLOX': 'Jan Jezabek BLOX MPEG Codec',
'BLZ0': 'DivX for Blizzard Decoder Filter',
'BT20': 'Conexant Prosumer Video',
'BTCV': 'Conexant Composite Video Codec',
'BTVC': 'Conexant Composite Video',
'BW00': 'BergWave (Wavelet)',
'BW10': 'Data Translation Broadway MPEG Capture',
'BXBG': 'BOXX BGR',
'BXRG': 'BOXX RGB',
'BXY2': 'BOXX 10-bit YUV',
'BXYV': 'BOXX YUV',
'CC12': 'Intel YUV12',
'CDV5': 'Canopus SD50/DVHD',
'CDVC': 'Canopus DV',
'CDVH': 'Canopus SD50/DVHD',
'CFCC': 'Digital Processing Systems DPS Perception',
'CFHD': 'CineForm 10-bit Visually Perfect HD',
'CGDI': 'Microsoft Office 97 Camcorder Video',
'CHAM': 'Winnov Caviara Champagne',
'CJPG': 'Creative WebCam JPEG',
'CLJR': 'Cirrus Logic YUV 4 pixels',
'CLLC': 'Canopus LossLess',
'CLPL': 'YV12',
'CMYK': 'Common Data Format in Printing',
'COL0': 'FFmpeg DivX ;-) (MS MPEG-4 v3)',
'COL1': 'FFmpeg DivX ;-) (MS MPEG-4 v3)',
'CPLA': 'Weitek 4:2:0 YUV Planar',
'CRAM': 'Microsoft Video 1 (CRAM)',
'CSCD': 'RenderSoft CamStudio lossless Codec',
'CTRX': 'Citrix Scalable Video Codec',
'CUVC': 'Canopus HQ',
'CVID': 'Radius Cinepak',
'CWLT': 'Microsoft Color WLT DIB',
'CYUV': 'Creative Labs YUV',
'CYUY': 'ATI YUV',
'D261': 'H.261',
'D263': 'H.263',
'DAVC': 'Dicas MPEGable H.264/MPEG-4 AVC base profile codec',
'DC25': 'MainConcept ProDV Codec',
'DCAP': 'Pinnacle DV25 Codec',
'DCL1': 'Data Connection Conferencing Codec',
'DCT0': 'WniWni Codec',
'DFSC': 'DebugMode FrameServer VFW Codec',
'DIB ': 'Full Frames (Uncompressed)',
'DIV1': 'FFmpeg-4 V1 (hacked MS MPEG-4 V1)',
'DIV2': 'MS MPEG-4 V2',
'DIV3': 'DivX v3 MPEG-4 Low-Motion',
'DIV4': 'DivX v3 MPEG-4 Fast-Motion',
'DIV5': 'DIV5',
'DIV6': 'DivX MPEG-4',
'DIVX': 'DivX',
'DM4V': 'Dicas MPEGable MPEG-4',
'DMB1': 'Matrox Rainbow Runner hardware MJPEG',
'DMB2': 'Paradigm MJPEG',
'DMK2': 'ViewSonic V36 PDA Video',
'DP02': 'DynaPel MPEG-4',
'DPS0': 'DPS Reality Motion JPEG',
'DPSC': 'DPS PAR Motion JPEG',
'DRWX': 'Pinnacle DV25 Codec',
'DSVD': 'DSVD',
'DTMT': 'Media-100 Codec',
'DTNT': 'Media-100 Codec',
'DUCK': 'Duck True Motion 1.0',
'DV10': 'BlueFish444 (lossless RGBA, YUV 10-bit)',
'DV25': 'Matrox DVCPRO codec',
'DV50': 'Matrox DVCPRO50 codec',
'DVAN': 'DVAN',
'DVC ': 'Apple QuickTime DV (DVCPRO NTSC)',
'DVCP': 'Apple QuickTime DV (DVCPRO PAL)',
'DVCS': 'MainConcept DV Codec',
'DVE2': 'InSoft DVE-2 Videoconferencing',
'DVH1': 'Pinnacle DVHD100',
'DVHD': 'DV 1125 lines at 30.00 Hz or 1250 lines at 25.00 Hz',
'DVIS': 'VSYNC DualMoon Iris DV codec',
'DVL ': 'Radius SoftDV 16:9 NTSC',
'DVLP': 'Radius SoftDV 16:9 PAL',
'DVMA': 'Darim Vision DVMPEG',
'DVOR': 'BlueFish444 (lossless RGBA, YUV 10-bit)',
'DVPN': 'Apple QuickTime DV (DV NTSC)',
'DVPP': 'Apple QuickTime DV (DV PAL)',
'DVR1': 'TARGA2000 Codec',
'DVRS': 'VSYNC DualMoon Iris DV codec',
'DVSD': 'DV',
'DVSL': 'DV compressed in SD (SDL)',
'DVX1': 'DVX1000SP Video Decoder',
'DVX2': 'DVX2000S Video Decoder',
'DVX3': 'DVX3000S Video Decoder',
'DX50': 'DivX v5',
'DXGM': 'Electronic Arts Game Video codec',
'DXSB': 'DivX Subtitles Codec',
'DXT1': 'Microsoft DirectX Compressed Texture (DXT1)',
'DXT2': 'Microsoft DirectX Compressed Texture (DXT2)',
'DXT3': 'Microsoft DirectX Compressed Texture (DXT3)',
'DXT4': 'Microsoft DirectX Compressed Texture (DXT4)',
'DXT5': 'Microsoft DirectX Compressed Texture (DXT5)',
'DXTC': 'Microsoft DirectX Compressed Texture (DXTC)',
'DXTN': 'Microsoft DirectX Compressed Texture (DXTn)',
'EKQ0': 'Elsa EKQ0',
'ELK0': 'Elsa ELK0',
'EM2V': 'Etymonix MPEG-2 I-frame',
'EQK0': 'Elsa graphics card quick codec',
'ESCP': 'Eidos Escape',
'ETV1': 'eTreppid Video ETV1',
'ETV2': 'eTreppid Video ETV2',
'ETVC': 'eTreppid Video ETVC',
'FFDS': 'FFDShow supported',
'FFV1': 'FFDShow supported',
'FFVH': 'FFVH codec',
'FLIC': 'Autodesk FLI/FLC Animation',
'FLJP': 'D-Vision Field Encoded Motion JPEG',
'FLV1': 'FLV1 codec',
'FMJP': 'D-Vision fieldbased ISO MJPEG',
'FRLE': 'SoftLab-NSK Y16 + Alpha RLE',
'FRWA': 'SoftLab-Nsk Forward Motion JPEG w/ alpha channel',
'FRWD': 'SoftLab-Nsk Forward Motion JPEG',
'FRWT': 'SoftLab-NSK Vision Forward Motion JPEG with Alpha-channel',
'FRWU': 'SoftLab-NSK Vision Forward Uncompressed',
'FVF1': 'Iterated Systems Fractal Video Frame',
'FVFW': 'ff MPEG-4 based on XviD codec',
'GEPJ': 'White Pine (ex Paradigm Matrix) Motion JPEG Codec',
'GJPG': 'Grand Tech GT891x Codec',
'GLCC': 'GigaLink AV Capture codec',
'GLZW': 'Motion LZW',
'GPEG': 'Motion JPEG',
'GPJM': 'Pinnacle ReelTime MJPEG Codec',
'GREY': 'Apparently a duplicate of Y800',
'GWLT': 'Microsoft Greyscale WLT DIB',
'H260': 'H.260',
'H261': 'H.261',
'H262': 'H.262',
'H263': 'H.263',
'H264': 'H.264 AVC',
'H265': 'H.265',
'H266': 'H.266',
'H267': 'H.267',
'H268': 'H.268',
'H269': 'H.269',
'HD10': 'BlueFish444 (lossless RGBA, YUV 10-bit)',
'HDX4': 'Jomigo HDX4',
'HFYU': 'Huffman Lossless Codec',
'HMCR': 'Rendition Motion Compensation Format (HMCR)',
'HMRR': 'Rendition Motion Compensation Format (HMRR)',
'I263': 'Intel ITU H.263 Videoconferencing (i263)',
'I420': 'Intel Indeo 4',
'IAN ': 'Intel RDX',
'ICLB': 'InSoft CellB Videoconferencing',
'IDM0': 'IDM Motion Wavelets 2.0',
'IF09': 'Microsoft H.261',
'IGOR': 'Power DVD',
'IJPG': 'Intergraph JPEG',
'ILVC': 'Intel Layered Video',
'ILVR': 'ITU-T H.263+',
'IMC1': 'IMC1',
'IMC2': 'IMC2',
'IMC3': 'IMC3',
'IMC4': 'IMC4',
'IMJG': 'Accom SphereOUS MJPEG with Alpha-channel',
'IPDV': 'I-O Data Device Giga AVI DV Codec',
'IPJ2': 'Image Power JPEG2000',
'IR21': 'Intel Indeo 2.1',
'IRAW': 'Intel YUV Uncompressed',
'IUYV': 'Interlaced version of UYVY (line order 0,2,4 then 1,3,5 etc)',
'IV30': 'Ligos Indeo 3.0',
'IV31': 'Ligos Indeo 3.1',
'IV32': 'Ligos Indeo 3.2',
'IV33': 'Ligos Indeo 3.3',
'IV34': 'Ligos Indeo 3.4',
'IV35': 'Ligos Indeo 3.5',
'IV36': 'Ligos Indeo 3.6',
'IV37': 'Ligos Indeo 3.7',
'IV38': 'Ligos Indeo 3.8',
'IV39': 'Ligos Indeo 3.9',
'IV40': 'Ligos Indeo Interactive 4.0',
'IV41': 'Ligos Indeo Interactive 4.1',
'IV42': 'Ligos Indeo Interactive 4.2',
'IV43': 'Ligos Indeo Interactive 4.3',
'IV44': 'Ligos Indeo Interactive 4.4',
'IV45': 'Ligos Indeo Interactive 4.5',
'IV46': 'Ligos Indeo Interactive 4.6',
'IV47': 'Ligos Indeo Interactive 4.7',
'IV48': 'Ligos Indeo Interactive 4.8',
'IV49': 'Ligos Indeo Interactive 4.9',
'IV50': 'Ligos Indeo Interactive 5.0',
'IY41': 'Interlaced version of Y41P (line order 0,2,4,...,1,3,5...)',
'IYU1': '12 bit format used in mode 2 of the IEEE 1394 Digital Camera 1.04 spec',
'IYU2': '24 bit format used in mode 2 of the IEEE 1394 Digital Camera 1.04 spec',
'IYUV': 'Intel Indeo iYUV 4:2:0',
'JBYR': 'Kensington JBYR',
'JFIF': 'Motion JPEG (FFmpeg)',
'JPEG': 'Still Image JPEG DIB',
'JPG ': 'JPEG compressed',
'JPGL': 'Webcam JPEG Light',
'KMVC': 'Karl Morton\'s Video Codec',
'KPCD': 'Kodak Photo CD',
'L261': 'Lead Technologies H.261',
'L263': 'Lead Technologies H.263',
'LAGS': 'Lagarith LossLess',
'LBYR': 'Creative WebCam codec',
'LCMW': 'Lead Technologies Motion CMW Codec',
'LCW2': 'LEADTools MCMW 9Motion Wavelet)',
'LEAD': 'LEAD Video Codec',
'LGRY': 'Lead Technologies Grayscale Image',
'LJ2K': 'LEADTools JPEG2000',
'LJPG': 'LEAD MJPEG Codec',
'LMP2': 'LEADTools MPEG2',
'LOCO': 'LOCO Lossless Codec',
'LSCR': 'LEAD Screen Capture',
'LSVM': 'Vianet Lighting Strike Vmail (Streaming)',
'LZO1': 'LZO compressed (lossless codec)',
'M261': 'Microsoft H.261',
'M263': 'Microsoft H.263',
'M4CC': 'ESS MPEG4 Divio codec',
'M4S2': 'Microsoft MPEG-4 (M4S2)',
'MC12': 'ATI Motion Compensation Format (MC12)',
'MC24': 'MainConcept Motion JPEG Codec',
'MCAM': 'ATI Motion Compensation Format (MCAM)',
'MCZM': 'Theory MicroCosm Lossless 64bit RGB with Alpha-channel',
'MDVD': 'Alex MicroDVD Video (hacked MS MPEG-4)',
'MDVF': 'Pinnacle DV/DV50/DVHD100',
'MHFY': 'A.M.Paredes mhuffyYUV (LossLess)',
'MJ2C': 'Morgan Multimedia Motion JPEG2000',
'MJPA': 'Pinnacle ReelTime MJPG hardware codec',
'MJPB': 'Motion JPEG codec',
'MJPG': 'Motion JPEG DIB',
'MJPX': 'Pegasus PICVideo Motion JPEG',
'MMES': 'Matrox MPEG-2 I-frame',
'MNVD': 'MindBend MindVid LossLess',
'MP2A': 'MPEG-2 Audio',
'MP2T': 'MPEG-2 Transport Stream',
'MP2V': 'MPEG-2 Video',
'MP41': 'Microsoft MPEG-4 V1 (enhansed H263)',
'MP42': 'Microsoft MPEG-4 (low-motion)',
'MP43': 'Microsoft MPEG-4 (fast-motion)',
'MP4A': 'MPEG-4 Audio',
'MP4S': 'Microsoft MPEG-4 (MP4S)',
'MP4T': 'MPEG-4 Transport Stream',
'MP4V': 'Apple QuickTime MPEG-4 native',
'MPEG': 'MPEG-1',
'MPG1': 'FFmpeg-1',
'MPG2': 'FFmpeg-1',
'MPG3': 'Same as Low motion DivX MPEG-4',
'MPG4': 'Microsoft MPEG-4 Video High Speed Compressor',
'MPGI': 'Sigma Designs MPEG',
'MPNG': 'Motion PNG codec',
'MRCA': 'Martin Regen Codec',
'MRLE': 'Run Length Encoding',
'MSS1': 'Windows Screen Video',
'MSS2': 'Windows Media 9',
'MSUC': 'MSU LossLess',
'MSVC': 'Microsoft Video 1',
'MSZH': 'Lossless codec (ZIP compression)',
'MTGA': 'Motion TGA images (24, 32 bpp)',
'MTX1': 'Matrox MTX1',
'MTX2': 'Matrox MTX2',
'MTX3': 'Matrox MTX3',
'MTX4': 'Matrox MTX4',
'MTX5': 'Matrox MTX5',
'MTX6': 'Matrox MTX6',
'MTX7': 'Matrox MTX7',
'MTX8': 'Matrox MTX8',
'MTX9': 'Matrox MTX9',
'MV12': 'MV12',
'MVI1': 'Motion Pixels MVI',
'MVI2': 'Motion Pixels MVI',
'MWV1': 'Aware Motion Wavelets',
'MYUV': 'Media-100 844/X Uncompressed',
'NAVI': 'nAVI',
'NDIG': 'Ahead Nero Digital MPEG-4 Codec',
'NHVU': 'NVidia Texture Format (GEForce 3)',
'NO16': 'Theory None16 64bit uncompressed RAW',
'NT00': 'NewTek LigtWave HDTV YUV with Alpha-channel',
'NTN1': 'Nogatech Video Compression 1',
'NTN2': 'Nogatech Video Compression 2 (GrabBee hardware coder)',
'NUV1': 'NuppelVideo',
'NV12': '8-bit Y plane followed by an interleaved U/V plane with 2x2 subsampling',
'NV21': 'As NV12 with U and V reversed in the interleaved plane',
'NVDS': 'nVidia Texture Format',
'NVHS': 'NVidia Texture Format (GEForce 3)',
'NVS0': 'nVidia GeForce Texture',
'NVS1': 'nVidia GeForce Texture',
'NVS2': 'nVidia GeForce Texture',
'NVS3': 'nVidia GeForce Texture',
'NVS4': 'nVidia GeForce Texture',
'NVS5': 'nVidia GeForce Texture',
'NVT0': 'nVidia GeForce Texture',
'NVT1': 'nVidia GeForce Texture',
'NVT2': 'nVidia GeForce Texture',
'NVT3': 'nVidia GeForce Texture',
'NVT4': 'nVidia GeForce Texture',
'NVT5': 'nVidia GeForce Texture',
'PDVC': 'I-O Data Device Digital Video Capture DV codec',
'PGVV': 'Radius Video Vision',
'PHMO': 'IBM Photomotion',
'PIM1': 'Pegasus Imaging',
'PIM2': 'Pegasus Imaging',
'PIMJ': 'Pegasus Imaging Lossless JPEG',
'PIXL': 'MiroVideo XL (Motion JPEG)',
'PNG ': 'Apple PNG',
'PNG1': 'Corecodec.org CorePNG Codec',
'PVEZ': 'Horizons Technology PowerEZ',
'PVMM': 'PacketVideo Corporation MPEG-4',
'PVW2': 'Pegasus Imaging Wavelet Compression',
'PVWV': 'Pegasus Imaging Wavelet 2000',
'PXLT': 'Apple Pixlet (Wavelet)',
'Q1.0': 'Q-Team QPEG 1.0 (www.q-team.de)',
'Q1.1': 'Q-Team QPEG 1.1 (www.q-team.de)',
'QDGX': 'Apple QuickDraw GX',
'QPEG': 'Q-Team QPEG 1.0',
'QPEQ': 'Q-Team QPEG 1.1',
'R210': 'BlackMagic YUV (Quick Time)',
'R411': 'Radius DV NTSC YUV',
'R420': 'Radius DV PAL YUV',
'RAVI': 'GroupTRON ReferenceAVI codec (dummy for MPEG compressor)',
'RAV_': 'GroupTRON ReferenceAVI codec (dummy for MPEG compressor)',
'RAW ': 'Full Frames (Uncompressed)',
'RGB ': 'Full Frames (Uncompressed)',
'RGB(15)': 'Uncompressed RGB15 5:5:5',
'RGB(16)': 'Uncompressed RGB16 5:6:5',
'RGB(24)': 'Uncompressed RGB24 8:8:8',
'RGB1': 'Uncompressed RGB332 3:3:2',
'RGBA': 'Raw RGB with alpha',
'RGBO': 'Uncompressed RGB555 5:5:5',
'RGBP': 'Uncompressed RGB565 5:6:5',
'RGBQ': 'Uncompressed RGB555X 5:5:5 BE',
'RGBR': 'Uncompressed RGB565X 5:6:5 BE',
'RGBT': 'Computer Concepts 32-bit support',
'RL4 ': 'RLE 4bpp RGB',
'RL8 ': 'RLE 8bpp RGB',
'RLE ': 'Microsoft Run Length Encoder',
'RLE4': 'Run Length Encoded 4',
'RLE8': 'Run Length Encoded 8',
'RMP4': 'REALmagic MPEG-4 Video Codec',
'ROQV': 'Id RoQ File Video Decoder',
'RPZA': 'Apple Video 16 bit "road pizza"',
'RT21': 'Intel Real Time Video 2.1',
'RTV0': 'NewTek VideoToaster',
'RUD0': 'Rududu video codec',
'RV10': 'RealVideo codec',
'RV13': 'RealVideo codec',
'RV20': 'RealVideo G2',
'RV30': 'RealVideo 8',
'RV40': 'RealVideo 9',
'RVX ': 'Intel RDX (RVX )',
'S263': 'Sorenson Vision H.263',
'S422': 'Tekram VideoCap C210 YUV 4:2:2',
'SAMR': 'Adaptive Multi-Rate (AMR) audio codec',
'SAN3': 'MPEG-4 codec (direct copy of DivX 3.11a)',
'SDCC': 'Sun Communication Digital Camera Codec',
'SEDG': 'Samsung MPEG-4 codec',
'SFMC': 'CrystalNet Surface Fitting Method',
'SHR0': 'BitJazz SheerVideo',
'SHR1': 'BitJazz SheerVideo',
'SHR2': 'BitJazz SheerVideo',
'SHR3': 'BitJazz SheerVideo',
'SHR4': 'BitJazz SheerVideo',
'SHR5': 'BitJazz SheerVideo',
'SHR6': 'BitJazz SheerVideo',
'SHR7': 'BitJazz SheerVideo',
'SJPG': 'CUseeMe Networks Codec',
'SL25': 'SoftLab-NSK DVCPRO',
'SL50': 'SoftLab-NSK DVCPRO50',
'SLDV': 'SoftLab-NSK Forward DV Draw codec',
'SLIF': 'SoftLab-NSK MPEG2 I-frames',
'SLMJ': 'SoftLab-NSK Forward MJPEG',
'SMC ': 'Apple Graphics (SMC) codec (256 color)',
'SMSC': 'Radius SMSC',
'SMSD': 'Radius SMSD',
'SMSV': 'WorldConnect Wavelet Video',
'SNOW': 'SNOW codec',
'SP40': 'SunPlus YUV',
'SP44': 'SunPlus Aiptek MegaCam Codec',
'SP53': 'SunPlus Aiptek MegaCam Codec',
'SP54': 'SunPlus Aiptek MegaCam Codec',
'SP55': 'SunPlus Aiptek MegaCam Codec',
'SP56': 'SunPlus Aiptek MegaCam Codec',
'SP57': 'SunPlus Aiptek MegaCam Codec',
'SP58': 'SunPlus Aiptek MegaCam Codec',
'SPIG': 'Radius Spigot',
'SPLC': 'Splash Studios ACM Audio Codec',
'SPRK': 'Sorenson Spark',
'SQZ2': 'Microsoft VXTreme Video Codec V2',
'STVA': 'ST CMOS Imager Data (Bayer)',
'STVB': 'ST CMOS Imager Data (Nudged Bayer)',
'STVC': 'ST CMOS Imager Data (Bunched)',
'STVX': 'ST CMOS Imager Data (Extended CODEC Data Format)',
'STVY': 'ST CMOS Imager Data (Extended CODEC Data Format with Correction Data)',
'SV10': 'Sorenson Video R1',
'SVQ1': 'Sorenson Video R3',
'SVQ3': 'Sorenson Video 3 (Apple Quicktime 5)',
'SWC1': 'MainConcept Motion JPEG Codec',
'T420': 'Toshiba YUV 4:2:0',
'TGA ': 'Apple TGA (with Alpha-channel)',
'THEO': 'FFVFW Supported Codec',
'TIFF': 'Apple TIFF (with Alpha-channel)',
'TIM2': 'Pinnacle RAL DVI',
'TLMS': 'TeraLogic Motion Intraframe Codec (TLMS)',
'TLST': 'TeraLogic Motion Intraframe Codec (TLST)',
'TM20': 'Duck TrueMotion 2.0',
'TM2A': 'Duck TrueMotion Archiver 2.0',
'TM2X': 'Duck TrueMotion 2X',
'TMIC': 'TeraLogic Motion Intraframe Codec (TMIC)',
'TMOT': 'Horizons Technology TrueMotion S',
'TR20': 'Duck TrueMotion RealTime 2.0',
'TRLE': 'Akula Alpha Pro Custom AVI (LossLess)',
'TSCC': 'TechSmith Screen Capture Codec',
'TV10': 'Tecomac Low-Bit Rate Codec',
'TVJP': 'TrueVision Field Encoded Motion JPEG',
'TVMJ': 'Truevision TARGA MJPEG Hardware Codec',
'TY0N': 'Trident TY0N',
'TY2C': 'Trident TY2C',
'TY2N': 'Trident TY2N',
'U263': 'UB Video StreamForce H.263',
'U<Y ': 'Discreet UC YUV 4:2:2:4 10 bit',
'U<YA': 'Discreet UC YUV 4:2:2:4 10 bit (with Alpha-channel)',
'UCOD': 'eMajix.com ClearVideo',
'ULTI': 'IBM Ultimotion',
'UMP4': 'UB Video MPEG 4',
'UYNV': 'UYVY',
'UYVP': 'YCbCr 4:2:2',
'UYVU': 'SoftLab-NSK Forward YUV codec',
'UYVY': 'UYVY 4:2:2 byte ordering',
'V210': 'Optibase VideoPump 10-bit 4:2:2 Component YCbCr',
'V261': 'Lucent VX2000S',
'V422': '24 bit YUV 4:2:2 Format',
'V655': '16 bit YUV 4:2:2 Format',
'VBLE': 'MarcFD VBLE Lossless Codec',
'VCR1': 'ATI VCR 1.0',
'VCR2': 'ATI VCR 2.0',
'VCR3': 'ATI VCR 3.0',
'VCR4': 'ATI VCR 4.0',
'VCR5': 'ATI VCR 5.0',
'VCR6': 'ATI VCR 6.0',
'VCR7': 'ATI VCR 7.0',
'VCR8': 'ATI VCR 8.0',
'VCR9': 'ATI VCR 9.0',
'VDCT': 'Video Maker Pro DIB',
'VDOM': 'VDOnet VDOWave',
'VDOW': 'VDOnet VDOLive (H.263)',
'VDST': 'VirtualDub remote frameclient ICM driver',
'VDTZ': 'Darim Vison VideoTizer YUV',
'VGPX': 'VGPixel Codec',
'VIDM': 'DivX 5.0 Pro Supported Codec',
'VIDS': 'YUV 4:2:2 CCIR 601 for V422',
'VIFP': 'VIFP',
'VIV1': 'Vivo H.263',
'VIV2': 'Vivo H.263',
'VIVO': 'Vivo H.263 v2.00',
'VIXL': 'Miro Video XL',
'VLV1': 'Videologic VLCAP.DRV',
'VP30': 'On2 VP3.0',
'VP31': 'On2 VP3.1',
'VP40': 'On2 TrueCast VP4',
'VP50': 'On2 TrueCast VP5',
'VP60': 'On2 TrueCast VP6',
'VP61': 'On2 TrueCast VP6.1',
'VP62': 'On2 TrueCast VP6.2',
'VP70': 'On2 TrueMotion VP7',
'VQC1': 'Vector-quantised codec 1',
'VQC2': 'Vector-quantised codec 2',
'VR21': 'BlackMagic YUV (Quick Time)',
'VSSH': 'Vanguard VSS H.264',
'VSSV': 'Vanguard Software Solutions Video Codec',
'VSSW': 'Vanguard VSS H.264',
'VTLP': 'Alaris VideoGramPixel Codec',
'VX1K': 'VX1000S Video Codec',
'VX2K': 'VX2000S Video Codec',
'VXSP': 'VX1000SP Video Codec',
'VYU9': 'ATI Technologies YUV',
'VYUY': 'ATI Packed YUV Data',
'WBVC': 'Winbond W9960',
'WHAM': 'Microsoft Video 1 (WHAM)',
'WINX': 'Winnov Software Compression',
'WJPG': 'AverMedia Winbond JPEG',
'WMV1': 'Windows Media Video V7',
'WMV2': 'Windows Media Video V8',
'WMV3': 'Windows Media Video V9',
'WMVA': 'WMVA codec',
'WMVP': 'Windows Media Video V9',
'WNIX': 'WniWni Codec',
'WNV1': 'Winnov Hardware Compression',
'WNVA': 'Winnov hw compress',
'WRLE': 'Apple QuickTime BMP Codec',
'WRPR': 'VideoTools VideoServer Client Codec',
'WV1F': 'WV1F codec',
'WVLT': 'IllusionHope Wavelet 9/7',
'WVP2': 'WVP2 codec',
'X263': 'Xirlink H.263',
'X264': 'XiWave GNU GPL x264 MPEG-4 Codec',
'XLV0': 'NetXL Video Decoder',
'XMPG': 'Xing MPEG (I-Frame only)',
'XVID': 'XviD MPEG-4',
'XVIX': 'Based on XviD MPEG-4 codec',
'XWV0': 'XiWave Video Codec',
'XWV1': 'XiWave Video Codec',
'XWV2': 'XiWave Video Codec',
'XWV3': 'XiWave Video Codec (Xi-3 Video)',
'XWV4': 'XiWave Video Codec',
'XWV5': 'XiWave Video Codec',
'XWV6': 'XiWave Video Codec',
'XWV7': 'XiWave Video Codec',
'XWV8': 'XiWave Video Codec',
'XWV9': 'XiWave Video Codec',
'XXAN': 'XXAN',
'XYZP': 'Extended PAL format XYZ palette',
'Y211': 'YUV 2:1:1 Packed',
'Y216': 'Pinnacle TARGA CineWave YUV (Quick Time)',
'Y411': 'YUV 4:1:1 Packed',
'Y41B': 'YUV 4:1:1 Planar',
'Y41P': 'PC1 4:1:1',
'Y41T': 'PC1 4:1:1 with transparency',
'Y422': 'Y422',
'Y42B': 'YUV 4:2:2 Planar',
'Y42T': 'PCI 4:2:2 with transparency',
'Y444': 'IYU2',
'Y8 ': 'Grayscale video',
'Y800': 'Simple grayscale video',
'YC12': 'Intel YUV12 Codec',
'YMPG': 'YMPEG Alpha',
'YU12': 'ATI YV12 4:2:0 Planar',
'YU92': 'Intel - YUV',
'YUNV': 'YUNV',
'YUV2': 'Apple Component Video (YUV 4:2:2)',
'YUV8': 'Winnov Caviar YUV8',
'YUV9': 'Intel YUV9',
'YUVP': 'YCbCr 4:2:2',
'YUY2': 'Uncompressed YUV 4:2:2',
'YUYV': 'Canopus YUV',
'YV12': 'YVU12 Planar',
'YV16': 'Elecard YUV 4:2:2 Planar',
'YV92': 'Intel Smart Video Recorder YVU9',
'YVU9': 'Intel YVU9 Planar',
'YVYU': 'YVYU 4:2:2 byte ordering',
'ZLIB': 'ZLIB',
'ZPEG': 'Metheus Video Zipper',
'ZYGO': 'ZyGo Video Codec'
}
# make it fool prove
for code, value in FOURCC.items():
if not code.upper() in FOURCC:
FOURCC[code.upper()] = value
if code.endswith(' '):
FOURCC[code.strip().upper()] = value
libs/enzyme/infos.py
+26
View File
@@ -0,0 +1,26 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__title__ = 'enzyme'
__description__ = 'Video metadata parser'
__version__ = '0.1'
__author__ = 'Antoine Bertin'
__email__ = 'diaoulael@gmail.com'
__license__ = 'GPLv3'
libs/enzyme/language.py
+538
View File
@@ -0,0 +1,538 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
import re
__all__ = ['resolve']
def resolve(code):
"""
Transform the given (2- or 3-letter) language code to a human readable
language name. The return value is a 2-tuple containing the given
language code and the language name. If the language code cannot be
resolved, name will be 'Unknown (<code>)'.
"""
if not code:
return None, None
if not isinstance(code, basestring):
raise ValueError('Invalid language code specified by parser')
# Take up to 3 letters from the code.
code = re.split(r'[^a-z]', code.lower())[0][:3]
for spec in codes:
if code in spec[:-1]:
return code, spec[-1]
return code, u'Unknown (%s)' % code
# Parsed from http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
codes = (
('aar', 'aa', u'Afar'),
('abk', 'ab', u'Abkhazian'),
('ace', u'Achinese'),
('ach', u'Acoli'),
('ada', u'Adangme'),
('ady', u'Adyghe'),
('afa', u'Afro-Asiatic '),
('afh', u'Afrihili'),
('afr', 'af', u'Afrikaans'),
('ain', u'Ainu'),
('aka', 'ak', u'Akan'),
('akk', u'Akkadian'),
('alb', 'sq', u'Albanian'),
('ale', u'Aleut'),
('alg', u'Algonquian languages'),
('alt', u'Southern Altai'),
('amh', 'am', u'Amharic'),
('ang', u'English, Old '),
('anp', u'Angika'),
('apa', u'Apache languages'),
('ara', 'ar', u'Arabic'),
('arc', u'Official Aramaic '),
('arg', 'an', u'Aragonese'),
('arm', 'hy', u'Armenian'),
('arn', u'Mapudungun'),
('arp', u'Arapaho'),
('art', u'Artificial '),
('arw', u'Arawak'),
('asm', 'as', u'Assamese'),
('ast', u'Asturian'),
('ath', u'Athapascan languages'),
('aus', u'Australian languages'),
('ava', 'av', u'Avaric'),
('ave', 'ae', u'Avestan'),
('awa', u'Awadhi'),
('aym', 'ay', u'Aymara'),
('aze', 'az', u'Azerbaijani'),
('bad', u'Banda languages'),
('bai', u'Bamileke languages'),
('bak', 'ba', u'Bashkir'),
('bal', u'Baluchi'),
('bam', 'bm', u'Bambara'),
('ban', u'Balinese'),
('baq', 'eu', u'Basque'),
('bas', u'Basa'),
('bat', u'Baltic '),
('bej', u'Beja'),
('bel', 'be', u'Belarusian'),
('bem', u'Bemba'),
('ben', 'bn', u'Bengali'),
('ber', u'Berber '),
('bho', u'Bhojpuri'),
('bih', 'bh', u'Bihari'),
('bik', u'Bikol'),
('bin', u'Bini'),
('bis', 'bi', u'Bislama'),
('bla', u'Siksika'),
('bnt', u'Bantu '),
('bos', 'bs', u'Bosnian'),
('bra', u'Braj'),
('bre', 'br', u'Breton'),
('btk', u'Batak languages'),
('bua', u'Buriat'),
('bug', u'Buginese'),
('bul', 'bg', u'Bulgarian'),
('bur', 'my', u'Burmese'),
('byn', u'Blin'),
('cad', u'Caddo'),
('cai', u'Central American Indian '),
('car', u'Galibi Carib'),
('cat', 'ca', u'Catalan'),
('cau', u'Caucasian '),
('ceb', u'Cebuano'),
('cel', u'Celtic '),
('cha', 'ch', u'Chamorro'),
('chb', u'Chibcha'),
('che', 'ce', u'Chechen'),
('chg', u'Chagatai'),
('chi', 'zh', u'Chinese'),
('chk', u'Chuukese'),
('chm', u'Mari'),
('chn', u'Chinook jargon'),
('cho', u'Choctaw'),
('chp', u'Chipewyan'),
('chr', u'Cherokee'),
('chu', 'cu', u'Church Slavic'),
('chv', 'cv', u'Chuvash'),
('chy', u'Cheyenne'),
('cmc', u'Chamic languages'),
('cop', u'Coptic'),
('cor', 'kw', u'Cornish'),
('cos', 'co', u'Corsican'),
('cpe', u'Creoles and pidgins, English based '),
('cpf', u'Creoles and pidgins, French-based '),
('cpp', u'Creoles and pidgins, Portuguese-based '),
('cre', 'cr', u'Cree'),
('crh', u'Crimean Tatar'),
('crp', u'Creoles and pidgins '),
('csb', u'Kashubian'),
('cus', u'Cushitic '),
('cze', 'cs', u'Czech'),
('dak', u'Dakota'),
('dan', 'da', u'Danish'),
('dar', u'Dargwa'),
('day', u'Land Dayak languages'),
('del', u'Delaware'),
('den', u'Slave '),
('dgr', u'Dogrib'),
('din', u'Dinka'),
('div', 'dv', u'Divehi'),
('doi', u'Dogri'),
('dra', u'Dravidian '),
('dsb', u'Lower Sorbian'),
('dua', u'Duala'),
('dum', u'Dutch, Middle '),
('dut', 'nl', u'Dutch'),
('dyu', u'Dyula'),
('dzo', 'dz', u'Dzongkha'),
('efi', u'Efik'),
('egy', u'Egyptian '),
('eka', u'Ekajuk'),
('elx', u'Elamite'),
('eng', 'en', u'English'),
('enm', u'English, Middle '),
('epo', 'eo', u'Esperanto'),
('est', 'et', u'Estonian'),
('ewe', 'ee', u'Ewe'),
('ewo', u'Ewondo'),
('fan', u'Fang'),
('fao', 'fo', u'Faroese'),
('fat', u'Fanti'),
('fij', 'fj', u'Fijian'),
('fil', u'Filipino'),
('fin', 'fi', u'Finnish'),
('fiu', u'Finno-Ugrian '),
('fon', u'Fon'),
('fre', 'fr', u'French'),
('frm', u'French, Middle '),
('fro', u'French, Old '),
('frr', u'Northern Frisian'),
('frs', u'Eastern Frisian'),
('fry', 'fy', u'Western Frisian'),
('ful', 'ff', u'Fulah'),
('fur', u'Friulian'),
('gaa', u'Ga'),
('gay', u'Gayo'),
('gba', u'Gbaya'),
('gem', u'Germanic '),
('geo', 'ka', u'Georgian'),
('ger', 'de', u'German'),
('gez', u'Geez'),
('gil', u'Gilbertese'),
('gla', 'gd', u'Gaelic'),
('gle', 'ga', u'Irish'),
('glg', 'gl', u'Galician'),
('glv', 'gv', u'Manx'),
('gmh', u'German, Middle High '),
('goh', u'German, Old High '),
('gon', u'Gondi'),
('gor', u'Gorontalo'),
('got', u'Gothic'),
('grb', u'Grebo'),
('grc', u'Greek, Ancient '),
('gre', 'el', u'Greek, Modern '),
('grn', 'gn', u'Guarani'),
('gsw', u'Swiss German'),
('guj', 'gu', u'Gujarati'),
('gwi', u"Gwich'in"),
('hai', u'Haida'),
('hat', 'ht', u'Haitian'),
('hau', 'ha', u'Hausa'),
('haw', u'Hawaiian'),
('heb', 'he', u'Hebrew'),
('her', 'hz', u'Herero'),
('hil', u'Hiligaynon'),
('him', u'Himachali'),
('hin', 'hi', u'Hindi'),
('hit', u'Hittite'),
('hmn', u'Hmong'),
('hmo', 'ho', u'Hiri Motu'),
('hsb', u'Upper Sorbian'),
('hun', 'hu', u'Hungarian'),
('hup', u'Hupa'),
('iba', u'Iban'),
('ibo', 'ig', u'Igbo'),
('ice', 'is', u'Icelandic'),
('ido', 'io', u'Ido'),
('iii', 'ii', u'Sichuan Yi'),
('ijo', u'Ijo languages'),
('iku', 'iu', u'Inuktitut'),
('ile', 'ie', u'Interlingue'),
('ilo', u'Iloko'),
('ina', 'ia', u'Interlingua '),
('inc', u'Indic '),
('ind', 'id', u'Indonesian'),
('ine', u'Indo-European '),
('inh', u'Ingush'),
('ipk', 'ik', u'Inupiaq'),
('ira', u'Iranian '),
('iro', u'Iroquoian languages'),
('ita', 'it', u'Italian'),
('jav', 'jv', u'Javanese'),
('jbo', u'Lojban'),
('jpn', 'ja', u'Japanese'),
('jpr', u'Judeo-Persian'),
('jrb', u'Judeo-Arabic'),
('kaa', u'Kara-Kalpak'),
('kab', u'Kabyle'),
('kac', u'Kachin'),
('kal', 'kl', u'Kalaallisut'),
('kam', u'Kamba'),
('kan', 'kn', u'Kannada'),
('kar', u'Karen languages'),
('kas', 'ks', u'Kashmiri'),
('kau', 'kr', u'Kanuri'),
('kaw', u'Kawi'),
('kaz', 'kk', u'Kazakh'),
('kbd', u'Kabardian'),
('kha', u'Khasi'),
('khi', u'Khoisan '),
('khm', 'km', u'Central Khmer'),
('kho', u'Khotanese'),
('kik', 'ki', u'Kikuyu'),
('kin', 'rw', u'Kinyarwanda'),
('kir', 'ky', u'Kirghiz'),
('kmb', u'Kimbundu'),
('kok', u'Konkani'),
('kom', 'kv', u'Komi'),
('kon', 'kg', u'Kongo'),
('kor', 'ko', u'Korean'),
('kos', u'Kosraean'),
('kpe', u'Kpelle'),
('krc', u'Karachay-Balkar'),
('krl', u'Karelian'),
('kro', u'Kru languages'),
('kru', u'Kurukh'),
('kua', 'kj', u'Kuanyama'),
('kum', u'Kumyk'),
('kur', 'ku', u'Kurdish'),
('kut', u'Kutenai'),
('lad', u'Ladino'),
('lah', u'Lahnda'),
('lam', u'Lamba'),
('lao', 'lo', u'Lao'),
('lat', 'la', u'Latin'),
('lav', 'lv', u'Latvian'),
('lez', u'Lezghian'),
('lim', 'li', u'Limburgan'),
('lin', 'ln', u'Lingala'),
('lit', 'lt', u'Lithuanian'),
('lol', u'Mongo'),
('loz', u'Lozi'),
('ltz', 'lb', u'Luxembourgish'),
('lua', u'Luba-Lulua'),
('lub', 'lu', u'Luba-Katanga'),
('lug', 'lg', u'Ganda'),
('lui', u'Luiseno'),
('lun', u'Lunda'),
('luo', u'Luo '),
('lus', u'Lushai'),
('mac', 'mk', u'Macedonian'),
('mad', u'Madurese'),
('mag', u'Magahi'),
('mah', 'mh', u'Marshallese'),
('mai', u'Maithili'),
('mak', u'Makasar'),
('mal', 'ml', u'Malayalam'),
('man', u'Mandingo'),
('mao', 'mi', u'Maori'),
('map', u'Austronesian '),
('mar', 'mr', u'Marathi'),
('mas', u'Masai'),
('may', 'ms', u'Malay'),
('mdf', u'Moksha'),
('mdr', u'Mandar'),
('men', u'Mende'),
('mga', u'Irish, Middle '),
('mic', u"Mi'kmaq"),
('min', u'Minangkabau'),
('mis', u'Uncoded languages'),
('mkh', u'Mon-Khmer '),
('mlg', 'mg', u'Malagasy'),
('mlt', 'mt', u'Maltese'),
('mnc', u'Manchu'),
('mni', u'Manipuri'),
('mno', u'Manobo languages'),
('moh', u'Mohawk'),
('mol', 'mo', u'Moldavian'),
('mon', 'mn', u'Mongolian'),
('mos', u'Mossi'),
('mul', u'Multiple languages'),
('mun', u'Munda languages'),
('mus', u'Creek'),
('mwl', u'Mirandese'),
('mwr', u'Marwari'),
('myn', u'Mayan languages'),
('myv', u'Erzya'),
('nah', u'Nahuatl languages'),
('nai', u'North American Indian'),
('nap', u'Neapolitan'),
('nau', 'na', u'Nauru'),
('nav', 'nv', u'Navajo'),
('nbl', 'nr', u'Ndebele, South'),
('nde', 'nd', u'Ndebele, North'),
('ndo', 'ng', u'Ndonga'),
('nds', u'Low German'),
('nep', 'ne', u'Nepali'),
('new', u'Nepal Bhasa'),
('nia', u'Nias'),
('nic', u'Niger-Kordofanian '),
('niu', u'Niuean'),
('nno', 'nn', u'Norwegian Nynorsk'),
('nob', 'nb', u'Bokm\xe5l, Norwegian'),
('nog', u'Nogai'),
('non', u'Norse, Old'),
('nor', 'no', u'Norwegian'),
('nqo', u"N'Ko"),
('nso', u'Pedi'),
('nub', u'Nubian languages'),
('nwc', u'Classical Newari'),
('nya', 'ny', u'Chichewa'),
('nym', u'Nyamwezi'),
('nyn', u'Nyankole'),
('nyo', u'Nyoro'),
('nzi', u'Nzima'),
('oci', 'oc', u'Occitan '),
('oji', 'oj', u'Ojibwa'),
('ori', 'or', u'Oriya'),
('orm', 'om', u'Oromo'),
('osa', u'Osage'),
('oss', 'os', u'Ossetian'),
('ota', u'Turkish, Ottoman '),
('oto', u'Otomian languages'),
('paa', u'Papuan '),
('pag', u'Pangasinan'),
('pal', u'Pahlavi'),
('pam', u'Pampanga'),
('pan', 'pa', u'Panjabi'),
('pap', u'Papiamento'),
('pau', u'Palauan'),
('peo', u'Persian, Old '),
('per', 'fa', u'Persian'),
('phi', u'Philippine '),
('phn', u'Phoenician'),
('pli', 'pi', u'Pali'),
('pol', 'pl', u'Polish'),
('pon', u'Pohnpeian'),
('por', 'pt', u'Portuguese'),
('pra', u'Prakrit languages'),
('pro', u'Proven\xe7al, Old '),
('pus', 'ps', u'Pushto'),
('qaa-qtz', u'Reserved for local use'),
('que', 'qu', u'Quechua'),
('raj', u'Rajasthani'),
('rap', u'Rapanui'),
('rar', u'Rarotongan'),
('roa', u'Romance '),
('roh', 'rm', u'Romansh'),
('rom', u'Romany'),
('rum', 'ro', u'Romanian'),
('run', 'rn', u'Rundi'),
('rup', u'Aromanian'),
('rus', 'ru', u'Russian'),
('sad', u'Sandawe'),
('sag', 'sg', u'Sango'),
('sah', u'Yakut'),
('sai', u'South American Indian '),
('sal', u'Salishan languages'),
('sam', u'Samaritan Aramaic'),
('san', 'sa', u'Sanskrit'),
('sas', u'Sasak'),
('sat', u'Santali'),
('scc', 'sr', u'Serbian'),
('scn', u'Sicilian'),
('sco', u'Scots'),
('scr', 'hr', u'Croatian'),
('sel', u'Selkup'),
('sem', u'Semitic '),
('sga', u'Irish, Old '),
('sgn', u'Sign Languages'),
('shn', u'Shan'),
('sid', u'Sidamo'),
('sin', 'si', u'Sinhala'),
('sio', u'Siouan languages'),
('sit', u'Sino-Tibetan '),
('sla', u'Slavic '),
('slo', 'sk', u'Slovak'),
('slv', 'sl', u'Slovenian'),
('sma', u'Southern Sami'),
('sme', 'se', u'Northern Sami'),
('smi', u'Sami languages '),
('smj', u'Lule Sami'),
('smn', u'Inari Sami'),
('smo', 'sm', u'Samoan'),
('sms', u'Skolt Sami'),
('sna', 'sn', u'Shona'),
('snd', 'sd', u'Sindhi'),
('snk', u'Soninke'),
('sog', u'Sogdian'),
('som', 'so', u'Somali'),
('son', u'Songhai languages'),
('sot', 'st', u'Sotho, Southern'),
('spa', 'es', u'Spanish'),
('srd', 'sc', u'Sardinian'),
('srn', u'Sranan Tongo'),
('srr', u'Serer'),
('ssa', u'Nilo-Saharan '),
('ssw', 'ss', u'Swati'),
('suk', u'Sukuma'),
('sun', 'su', u'Sundanese'),
('sus', u'Susu'),
('sux', u'Sumerian'),
('swa', 'sw', u'Swahili'),
('swe', 'sv', u'Swedish'),
('syc', u'Classical Syriac'),
('syr', u'Syriac'),
('tah', 'ty', u'Tahitian'),
('tai', u'Tai '),
('tam', 'ta', u'Tamil'),
('tat', 'tt', u'Tatar'),
('tel', 'te', u'Telugu'),
('tem', u'Timne'),
('ter', u'Tereno'),
('tet', u'Tetum'),
('tgk', 'tg', u'Tajik'),
('tgl', 'tl', u'Tagalog'),
('tha', 'th', u'Thai'),
('tib', 'bo', u'Tibetan'),
('tig', u'Tigre'),
('tir', 'ti', u'Tigrinya'),
('tiv', u'Tiv'),
('tkl', u'Tokelau'),
('tlh', u'Klingon'),
('tli', u'Tlingit'),
('tmh', u'Tamashek'),
('tog', u'Tonga '),
('ton', 'to', u'Tonga '),
('tpi', u'Tok Pisin'),
('tsi', u'Tsimshian'),
('tsn', 'tn', u'Tswana'),
('tso', 'ts', u'Tsonga'),
('tuk', 'tk', u'Turkmen'),
('tum', u'Tumbuka'),
('tup', u'Tupi languages'),
('tur', 'tr', u'Turkish'),
('tut', u'Altaic '),
('tvl', u'Tuvalu'),
('twi', 'tw', u'Twi'),
('tyv', u'Tuvinian'),
('udm', u'Udmurt'),
('uga', u'Ugaritic'),
('uig', 'ug', u'Uighur'),
('ukr', 'uk', u'Ukrainian'),
('umb', u'Umbundu'),
('und', u'Undetermined'),
('urd', 'ur', u'Urdu'),
('uzb', 'uz', u'Uzbek'),
('vai', u'Vai'),
('ven', 've', u'Venda'),
('vie', 'vi', u'Vietnamese'),
('vol', 'vo', u'Volap\xfck'),
('vot', u'Votic'),
('wak', u'Wakashan languages'),
('wal', u'Walamo'),
('war', u'Waray'),
('was', u'Washo'),
('wel', 'cy', u'Welsh'),
('wen', u'Sorbian languages'),
('wln', 'wa', u'Walloon'),
('wol', 'wo', u'Wolof'),
('xal', u'Kalmyk'),
('xho', 'xh', u'Xhosa'),
('yao', u'Yao'),
('yap', u'Yapese'),
('yid', 'yi', u'Yiddish'),
('yor', 'yo', u'Yoruba'),
('ypk', u'Yupik languages'),
('zap', u'Zapotec'),
('zbl', u'Blissymbols'),
('zen', u'Zenaga'),
('zha', 'za', u'Zhuang'),
('znd', u'Zande languages'),
('zul', 'zu', u'Zulu'),
('zun', u'Zuni'),
('zxx', u'No linguistic content'),
('zza', u'Zaza'),
)
libs/enzyme/mkv.py
+841
View File
@@ -0,0 +1,841 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
# Copyright (C) 2003-2006 Jason Tackaberry <tack@urandom.ca>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
from datetime import datetime
from exceptions import *
from struct import unpack
import core
import logging
import re
# get logging object
log = logging.getLogger(__name__)
# Main IDs for the Matroska streams
MATROSKA_VIDEO_TRACK = 0x01
MATROSKA_AUDIO_TRACK = 0x02
MATROSKA_SUBTITLES_TRACK = 0x11
MATROSKA_HEADER_ID = 0x1A45DFA3
MATROSKA_TRACKS_ID = 0x1654AE6B
MATROSKA_CUES_ID = 0x1C53BB6B
MATROSKA_SEGMENT_ID = 0x18538067
MATROSKA_SEGMENT_INFO_ID = 0x1549A966
MATROSKA_CLUSTER_ID = 0x1F43B675
MATROSKA_VOID_ID = 0xEC
MATROSKA_CRC_ID = 0xBF
MATROSKA_TIMECODESCALE_ID = 0x2AD7B1
MATROSKA_DURATION_ID = 0x4489
MATROSKA_CRC32_ID = 0xBF
MATROSKA_TIMECODESCALE_ID = 0x2AD7B1
MATROSKA_MUXING_APP_ID = 0x4D80
MATROSKA_WRITING_APP_ID = 0x5741
MATROSKA_CODEC_ID = 0x86
MATROSKA_CODEC_PRIVATE_ID = 0x63A2
MATROSKA_FRAME_DURATION_ID = 0x23E383
MATROSKA_VIDEO_SETTINGS_ID = 0xE0
MATROSKA_VIDEO_WIDTH_ID = 0xB0
MATROSKA_VIDEO_HEIGHT_ID = 0xBA
MATROSKA_VIDEO_INTERLACED_ID = 0x9A
MATROSKA_VIDEO_DISPLAY_WIDTH_ID = 0x54B0
MATROSKA_VIDEO_DISPLAY_HEIGHT_ID = 0x54BA
MATROSKA_AUDIO_SETTINGS_ID = 0xE1
MATROSKA_AUDIO_SAMPLERATE_ID = 0xB5
MATROSKA_AUDIO_CHANNELS_ID = 0x9F
MATROSKA_TRACK_UID_ID = 0x73C5
MATROSKA_TRACK_NUMBER_ID = 0xD7
MATROSKA_TRACK_TYPE_ID = 0x83
MATROSKA_TRACK_LANGUAGE_ID = 0x22B59C
MATROSKA_TRACK_OFFSET = 0x537F
MATROSKA_TRACK_FLAG_DEFAULT_ID = 0x88
MATROSKA_TRACK_FLAG_ENABLED_ID = 0xB9
MATROSKA_TITLE_ID = 0x7BA9
MATROSKA_DATE_UTC_ID = 0x4461
MATROSKA_NAME_ID = 0x536E
MATROSKA_CHAPTERS_ID = 0x1043A770
MATROSKA_CHAPTER_UID_ID = 0x73C4
MATROSKA_EDITION_ENTRY_ID = 0x45B9
MATROSKA_CHAPTER_ATOM_ID = 0xB6
MATROSKA_CHAPTER_TIME_START_ID = 0x91
MATROSKA_CHAPTER_TIME_END_ID = 0x92
MATROSKA_CHAPTER_FLAG_ENABLED_ID = 0x4598
MATROSKA_CHAPTER_DISPLAY_ID = 0x80
MATROSKA_CHAPTER_LANGUAGE_ID = 0x437C
MATROSKA_CHAPTER_STRING_ID = 0x85
MATROSKA_ATTACHMENTS_ID = 0x1941A469
MATROSKA_ATTACHED_FILE_ID = 0x61A7
MATROSKA_FILE_DESC_ID = 0x467E
MATROSKA_FILE_NAME_ID = 0x466E
MATROSKA_FILE_MIME_TYPE_ID = 0x4660
MATROSKA_FILE_DATA_ID = 0x465C
MATROSKA_SEEKHEAD_ID = 0x114D9B74
MATROSKA_SEEK_ID = 0x4DBB
MATROSKA_SEEKID_ID = 0x53AB
MATROSKA_SEEK_POSITION_ID = 0x53AC
MATROSKA_TAGS_ID = 0x1254C367
MATROSKA_TAG_ID = 0x7373
MATROSKA_TARGETS_ID = 0x63C0
MATROSKA_TARGET_TYPE_VALUE_ID = 0x68CA
MATROSKA_TARGET_TYPE_ID = 0x63CA
MATRSOKA_TAGS_TRACK_UID_ID = 0x63C5
MATRSOKA_TAGS_EDITION_UID_ID = 0x63C9
MATRSOKA_TAGS_CHAPTER_UID_ID = 0x63C4
MATRSOKA_TAGS_ATTACHMENT_UID_ID = 0x63C6
MATROSKA_SIMPLE_TAG_ID = 0x67C8
MATROSKA_TAG_NAME_ID = 0x45A3
MATROSKA_TAG_LANGUAGE_ID = 0x447A
MATROSKA_TAG_STRING_ID = 0x4487
MATROSKA_TAG_BINARY_ID = 0x4485
# See mkv spec for details:
# http://www.matroska.org/technical/specs/index.html
# Map to convert to well known codes
# http://haali.cs.msu.ru/mkv/codecs.pdf
FOURCCMap = {
'V_THEORA': 'THEO',
'V_SNOW': 'SNOW',
'V_MPEG4/ISO/ASP': 'MP4V',
'V_MPEG4/ISO/AVC': 'AVC1',
'A_AC3': 0x2000,
'A_MPEG/L3': 0x0055,
'A_MPEG/L2': 0x0050,
'A_MPEG/L1': 0x0050,
'A_DTS': 0x2001,
'A_PCM/INT/LIT': 0x0001,
'A_PCM/FLOAT/IEEE': 0x003,
'A_TTA1': 0x77a1,
'A_WAVPACK4': 0x5756,
'A_VORBIS': 0x6750,
'A_FLAC': 0xF1AC,
'A_AAC': 0x00ff,
'A_AAC/': 0x00ff
}
def matroska_date_to_datetime(date):
"""
Converts a date in Matroska's date format to a python datetime object.
Returns the given date string if it could not be converted.
"""
# From the specs:
# The fields with dates should have the following format: YYYY-MM-DD
# HH:MM:SS.MSS [...] To store less accuracy, you remove items starting
# from the right. To store only the year, you would use, "2004". To store
# a specific day such as May 1st, 2003, you would use "2003-05-01".
format = re.split(r'([-:. ])', '%Y-%m-%d %H:%M:%S.%f')
while format:
try:
return datetime.strptime(date, ''.join(format))
except ValueError:
format = format[:-2]
return date
def matroska_bps_to_bitrate(bps):
"""
Tries to convert a free-form bps string into a bitrate (bits per second).
"""
m = re.search('([\d.]+)\s*(\D.*)', bps)
if m:
bps, suffix = m.groups()
if 'kbit' in suffix:
return float(bps) * 1024
elif 'kbyte' in suffix:
return float(bps) * 1024 * 8
elif 'byte' in suffix:
return float(bps) * 8
elif 'bps' in suffix or 'bit' in suffix:
return float(bps)
if bps.replace('.', '').isdigit():
if float(bps) < 30000:
# Assume kilobits and convert to bps
return float(bps) * 1024
return float(bps)
# Used to convert the official matroska tag names (only lower-cased) to core
# attributes. tag name -> attr, filter
TAGS_MAP = {
# From Media core
u'title': ('title', None),
u'subtitle': ('caption', None),
u'comment': ('comment', None),
u'url': ('url', None),
u'artist': ('artist', None),
u'keywords': ('keywords', lambda s: [word.strip() for word in s.split(',')]),
u'composer_nationality': ('country', None),
u'date_released': ('datetime', None),
u'date_recorded': ('datetime', None),
u'date_written': ('datetime', None),
# From Video core
u'encoder': ('encoder', None),
u'bps': ('bitrate', matroska_bps_to_bitrate),
u'part_number': ('trackno', int),
u'total_parts': ('trackof', int),
u'copyright': ('copyright', None),
u'genre': ('genre', None),
u'actor': ('actors', None),
u'written_by': ('writer', None),
u'producer': ('producer', None),
u'production_studio': ('studio', None),
u'law_rating': ('rating', None),
u'summary': ('summary', None),
u'synopsis': ('synopsis', None),
}
class EbmlEntity:
"""
This is class that is responsible to handle one Ebml entity as described in
the Matroska/Ebml spec
"""
def __init__(self, inbuf):
# Compute the EBML id
# Set the CRC len to zero
self.crc_len = 0
# Now loop until we find an entity without CRC
try:
self.build_entity(inbuf)
except IndexError:
raise ParseError()
while self.get_id() == MATROSKA_CRC32_ID:
self.crc_len += self.get_total_len()
inbuf = inbuf[self.get_total_len():]
self.build_entity(inbuf)
def build_entity(self, inbuf):
self.compute_id(inbuf)
if self.id_len == 0:
log.error("EBML entity not found, bad file format")
raise ParseError()
self.entity_len, self.len_size = self.compute_len(inbuf[self.id_len:])
self.entity_data = inbuf[self.get_header_len() : self.get_total_len()]
self.ebml_length = self.entity_len
self.entity_len = min(len(self.entity_data), self.entity_len)
# if the data size is 8 or less, it could be a numeric value
self.value = 0
if self.entity_len <= 8:
for pos, shift in zip(range(self.entity_len), range((self.entity_len-1)*8, -1, -8)):
self.value |= ord(self.entity_data[pos]) << shift
def add_data(self, data):
maxlen = self.ebml_length - len(self.entity_data)
if maxlen <= 0:
return
self.entity_data += data[:maxlen]
self.entity_len = len(self.entity_data)
def compute_id(self, inbuf):
self.id_len = 0
if len(inbuf) < 1:
return 0
first = ord(inbuf[0])
if first & 0x80:
self.id_len = 1
self.entity_id = first
elif first & 0x40:
if len(inbuf) < 2:
return 0
self.id_len = 2
self.entity_id = ord(inbuf[0])<<8 | ord(inbuf[1])
elif first & 0x20:
if len(inbuf) < 3:
return 0
self.id_len = 3
self.entity_id = (ord(inbuf[0])<<16) | (ord(inbuf[1])<<8) | \
(ord(inbuf[2]))
elif first & 0x10:
if len(inbuf) < 4:
return 0
self.id_len = 4
self.entity_id = (ord(inbuf[0])<<24) | (ord(inbuf[1])<<16) | \
(ord(inbuf[2])<<8) | (ord(inbuf[3]))
self.entity_str = inbuf[0:self.id_len]
def compute_len(self, inbuf):
if not inbuf:
return 0, 0
i = num_ffs = 0
len_mask = 0x80
len = ord(inbuf[0])
while not len & len_mask:
i += 1
len_mask >>= 1
if i >= 8:
return 0, 0
len &= len_mask - 1
if len == len_mask - 1:
num_ffs += 1
for p in range(i):
len = (len << 8) | ord(inbuf[p + 1])
if len & 0xff == 0xff:
num_ffs += 1
if num_ffs == i + 1:
len = 0
return len, i + 1
def get_crc_len(self):
return self.crc_len
def get_value(self):
return self.value
def get_float_value(self):
if len(self.entity_data) == 4:
return unpack('!f', self.entity_data)[0]
elif len(self.entity_data) == 8:
return unpack('!d', self.entity_data)[0]
return 0.0
def get_data(self):
return self.entity_data
def get_utf8(self):
return unicode(self.entity_data, 'utf-8', 'replace')
def get_str(self):
return unicode(self.entity_data, 'ascii', 'replace')
def get_id(self):
return self.entity_id
def get_str_id(self):
return self.entity_str
def get_len(self):
return self.entity_len
def get_total_len(self):
return self.entity_len + self.id_len + self.len_size
def get_header_len(self):
return self.id_len + self.len_size
class Matroska(core.AVContainer):
"""
Matroska video and audio parser. If at least one video stream is
detected it will set the type to MEDIA_AV.
"""
def __init__(self, file):
core.AVContainer.__init__(self)
self.samplerate = 1
self.file = file
# Read enough that we're likely to get the full seekhead (FIXME: kludge)
buffer = file.read(2000)
if len(buffer) == 0:
# Regular File end
raise ParseError()
# Check the Matroska header
header = EbmlEntity(buffer)
if header.get_id() != MATROSKA_HEADER_ID:
raise ParseError()
log.debug("HEADER ID found %08X" % header.get_id() )
self.mime = 'video/x-matroska'
self.type = 'Matroska'
self.has_idx = False
self.objects_by_uid = {}
# Now get the segment
self.segment = segment = EbmlEntity(buffer[header.get_total_len():])
# Record file offset of segment data for seekheads
self.segment.offset = header.get_total_len() + segment.get_header_len()
if segment.get_id() != MATROSKA_SEGMENT_ID:
log.debug("SEGMENT ID not found %08X" % segment.get_id())
return
log.debug("SEGMENT ID found %08X" % segment.get_id())
try:
for elem in self.process_one_level(segment):
if elem.get_id() == MATROSKA_SEEKHEAD_ID:
self.process_elem(elem)
except ParseError:
pass
if not self.has_idx:
log.warning('File has no index')
self._set('corrupt', True)
def process_elem(self, elem):
elem_id = elem.get_id()
log.debug('BEGIN: process element %s' % hex(elem_id))
if elem_id == MATROSKA_SEGMENT_INFO_ID:
duration = 0
scalecode = 1000000.0
for ielem in self.process_one_level(elem):
ielem_id = ielem.get_id()
if ielem_id == MATROSKA_TIMECODESCALE_ID:
scalecode = ielem.get_value()
elif ielem_id == MATROSKA_DURATION_ID:
duration = ielem.get_float_value()
elif ielem_id == MATROSKA_TITLE_ID:
self.title = ielem.get_utf8()
elif ielem_id == MATROSKA_DATE_UTC_ID:
timestamp = unpack('!q', ielem.get_data())[0] / 10.0**9
# Date is offset 2001-01-01 00:00:00 (timestamp 978307200.0)
self.timestamp = int(timestamp + 978307200)
self.length = duration * scalecode / 1000000000.0
elif elem_id == MATROSKA_TRACKS_ID:
self.process_tracks(elem)
elif elem_id == MATROSKA_CHAPTERS_ID:
self.process_chapters(elem)
elif elem_id == MATROSKA_ATTACHMENTS_ID:
self.process_attachments(elem)
elif elem_id == MATROSKA_SEEKHEAD_ID:
self.process_seekhead(elem)
elif elem_id == MATROSKA_TAGS_ID:
self.process_tags(elem)
elif elem_id == MATROSKA_CUES_ID:
self.has_idx = True
log.debug('END: process element %s' % hex(elem_id))
return True
def process_seekhead(self, elem):
for seek_elem in self.process_one_level(elem):
if seek_elem.get_id() != MATROSKA_SEEK_ID:
continue
for sub_elem in self.process_one_level(seek_elem):
if sub_elem.get_id() == MATROSKA_SEEKID_ID:
if sub_elem.get_value() == MATROSKA_CLUSTER_ID:
# Not interested in these.
return
elif sub_elem.get_id() == MATROSKA_SEEK_POSITION_ID:
self.file.seek(self.segment.offset + sub_elem.get_value())
buffer = self.file.read(100)
try:
elem = EbmlEntity(buffer)
except ParseError:
continue
# Fetch all data necessary for this element.
elem.add_data(self.file.read(elem.ebml_length))
self.process_elem(elem)
def process_tracks(self, tracks):
tracksbuf = tracks.get_data()
index = 0
while index < tracks.get_len():
trackelem = EbmlEntity(tracksbuf[index:])
log.debug ("ELEMENT %X found" % trackelem.get_id())
self.process_track(trackelem)
index += trackelem.get_total_len() + trackelem.get_crc_len()
def process_one_level(self, item):
buf = item.get_data()
index = 0
while index < item.get_len():
if len(buf[index:]) == 0:
break
elem = EbmlEntity(buf[index:])
yield elem
index += elem.get_total_len() + elem.get_crc_len()
def set_track_defaults(self, track):
track.language = 'eng'
def process_track(self, track):
# Collapse generator into a list since we need to iterate over it
# twice.
elements = [x for x in self.process_one_level(track)]
track_type = [x.get_value() for x in elements if x.get_id() == MATROSKA_TRACK_TYPE_ID]
if not track_type:
log.debug('Bad track: no type id found')
return
track_type = track_type[0]
track = None
if track_type == MATROSKA_VIDEO_TRACK:
log.debug("Video track found")
track = self.process_video_track(elements)
elif track_type == MATROSKA_AUDIO_TRACK:
log.debug("Audio track found")
track = self.process_audio_track(elements)
elif track_type == MATROSKA_SUBTITLES_TRACK:
log.debug("Subtitle track found")
track = core.Subtitle()
self.set_track_defaults(track)
track.id = len(self.subtitles)
self.subtitles.append(track)
for elem in elements:
self.process_track_common(elem, track)
def process_track_common(self, elem, track):
elem_id = elem.get_id()
if elem_id == MATROSKA_TRACK_LANGUAGE_ID:
track.language = elem.get_str()
log.debug("Track language found: %s" % track.language)
elif elem_id == MATROSKA_NAME_ID:
track.title = elem.get_utf8()
elif elem_id == MATROSKA_TRACK_NUMBER_ID:
track.trackno = elem.get_value()
elif elem_id == MATROSKA_TRACK_FLAG_ENABLED_ID:
track.enabled = bool(elem.get_value())
elif elem_id == MATROSKA_TRACK_FLAG_DEFAULT_ID:
track.default = bool(elem.get_value())
elif elem_id == MATROSKA_CODEC_ID:
track.codec = elem.get_str()
elif elem_id == MATROSKA_CODEC_PRIVATE_ID:
track.codec_private = elem.get_data()
elif elem_id == MATROSKA_TRACK_UID_ID:
self.objects_by_uid[elem.get_value()] = track
def process_video_track(self, elements):
track = core.VideoStream()
# Defaults
track.codec = u'Unknown'
track.fps = 0
self.set_track_defaults(track)
for elem in elements:
elem_id = elem.get_id()
if elem_id == MATROSKA_CODEC_ID:
track.codec = elem.get_str()
elif elem_id == MATROSKA_FRAME_DURATION_ID:
try:
track.fps = 1 / (pow(10, -9) * (elem.get_value()))
except ZeroDivisionError:
pass
elif elem_id == MATROSKA_VIDEO_SETTINGS_ID:
d_width = d_height = None
for settings_elem in self.process_one_level(elem):
settings_elem_id = settings_elem.get_id()
if settings_elem_id == MATROSKA_VIDEO_WIDTH_ID:
track.width = settings_elem.get_value()
elif settings_elem_id == MATROSKA_VIDEO_HEIGHT_ID:
track.height = settings_elem.get_value()
elif settings_elem_id == MATROSKA_VIDEO_DISPLAY_WIDTH_ID:
d_width = settings_elem.get_value()
elif settings_elem_id == MATROSKA_VIDEO_DISPLAY_HEIGHT_ID:
d_height = settings_elem.get_value()
elif settings_elem_id == MATROSKA_VIDEO_INTERLACED_ID:
value = int(settings_elem.get_value())
self._set('interlaced', value)
if None not in [d_width, d_height]:
track.aspect = float(d_width) / d_height
else:
self.process_track_common(elem, track)
# convert codec information
# http://haali.cs.msu.ru/mkv/codecs.pdf
if track.codec in FOURCCMap:
track.codec = FOURCCMap[track.codec]
elif '/' in track.codec and track.codec.split('/')[0] + '/' in FOURCCMap:
track.codec = FOURCCMap[track.codec.split('/')[0] + '/']
elif track.codec.endswith('FOURCC') and len(track.codec_private or '') == 40:
track.codec = track.codec_private[16:20]
elif track.codec.startswith('V_REAL/'):
track.codec = track.codec[7:]
elif track.codec.startswith('V_'):
# FIXME: add more video codecs here
track.codec = track.codec[2:]
track.id = len(self.video)
self.video.append(track)
return track
def process_audio_track(self, elements):
track = core.AudioStream()
track.codec = u'Unknown'
self.set_track_defaults(track)
for elem in elements:
elem_id = elem.get_id()
if elem_id == MATROSKA_CODEC_ID:
track.codec = elem.get_str()
elif elem_id == MATROSKA_AUDIO_SETTINGS_ID:
for settings_elem in self.process_one_level(elem):
settings_elem_id = settings_elem.get_id()
if settings_elem_id == MATROSKA_AUDIO_SAMPLERATE_ID:
track.samplerate = settings_elem.get_float_value()
elif settings_elem_id == MATROSKA_AUDIO_CHANNELS_ID:
track.channels = settings_elem.get_value()
else:
self.process_track_common(elem, track)
if track.codec in FOURCCMap:
track.codec = FOURCCMap[track.codec]
elif '/' in track.codec and track.codec.split('/')[0] + '/' in FOURCCMap:
track.codec = FOURCCMap[track.codec.split('/')[0] + '/']
elif track.codec.startswith('A_'):
track.codec = track.codec[2:]
track.id = len(self.audio)
self.audio.append(track)
return track
def process_chapters(self, chapters):
elements = self.process_one_level(chapters)
for elem in elements:
if elem.get_id() == MATROSKA_EDITION_ENTRY_ID:
buf = elem.get_data()
index = 0
while index < elem.get_len():
sub_elem = EbmlEntity(buf[index:])
if sub_elem.get_id() == MATROSKA_CHAPTER_ATOM_ID:
self.process_chapter_atom(sub_elem)
index += sub_elem.get_total_len() + sub_elem.get_crc_len()
def process_chapter_atom(self, atom):
elements = self.process_one_level(atom)
chap = core.Chapter()
for elem in elements:
elem_id = elem.get_id()
if elem_id == MATROSKA_CHAPTER_TIME_START_ID:
# Scale timecode to seconds (float)
chap.pos = elem.get_value() / 1000000 / 1000.0
elif elem_id == MATROSKA_CHAPTER_FLAG_ENABLED_ID:
chap.enabled = elem.get_value()
elif elem_id == MATROSKA_CHAPTER_DISPLAY_ID:
# Matroska supports multiple (chapter name, language) pairs for
# each chapter, so chapter names can be internationalized. This
# logic will only take the last one in the list.
for display_elem in self.process_one_level(elem):
if display_elem.get_id() == MATROSKA_CHAPTER_STRING_ID:
chap.name = display_elem.get_utf8()
elif elem_id == MATROSKA_CHAPTER_UID_ID:
self.objects_by_uid[elem.get_value()] = chap
log.debug('Chapter "%s" found', chap.name)
chap.id = len(self.chapters)
self.chapters.append(chap)
def process_attachments(self, attachments):
buf = attachments.get_data()
index = 0
while index < attachments.get_len():
elem = EbmlEntity(buf[index:])
if elem.get_id() == MATROSKA_ATTACHED_FILE_ID:
self.process_attachment(elem)
index += elem.get_total_len() + elem.get_crc_len()
def process_attachment(self, attachment):
elements = self.process_one_level(attachment)
name = desc = mimetype = ""
data = None
for elem in elements:
elem_id = elem.get_id()
if elem_id == MATROSKA_FILE_NAME_ID:
name = elem.get_utf8()
elif elem_id == MATROSKA_FILE_DESC_ID:
desc = elem.get_utf8()
elif elem_id == MATROSKA_FILE_MIME_TYPE_ID:
mimetype = elem.get_data()
elif elem_id == MATROSKA_FILE_DATA_ID:
data = elem.get_data()
# Right now we only support attachments that could be cover images.
# Make a guess to see if this attachment is a cover image.
if mimetype.startswith("image/") and u"cover" in (name+desc).lower() and data:
self.thumbnail = data
log.debug('Attachment "%s" found' % name)
def process_tags(self, tags):
# Tags spec: http://www.matroska.org/technical/specs/tagging/index.html
# Iterate over Tags children. Tags element children is a
# Tag element (whose children are SimpleTags) and a Targets element
# whose children specific what objects the tags apply to.
for tag_elem in self.process_one_level(tags):
# Start a new dict to hold all SimpleTag elements.
tags_dict = core.Tags()
# A list of target uids this tags dict applies too. If empty,
# tags are global.
targets = []
for sub_elem in self.process_one_level(tag_elem):
if sub_elem.get_id() == MATROSKA_SIMPLE_TAG_ID:
self.process_simple_tag(sub_elem, tags_dict)
elif sub_elem.get_id() == MATROSKA_TARGETS_ID:
# Targets element: if there is no uid child (track uid,
# chapter uid, etc.) then the tags dict applies to the
# whole file (top-level Media object).
for target_elem in self.process_one_level(sub_elem):
target_elem_id = target_elem.get_id()
if target_elem_id in (MATRSOKA_TAGS_TRACK_UID_ID, MATRSOKA_TAGS_EDITION_UID_ID,
MATRSOKA_TAGS_CHAPTER_UID_ID, MATRSOKA_TAGS_ATTACHMENT_UID_ID):
targets.append(target_elem.get_value())
elif target_elem_id == MATROSKA_TARGET_TYPE_VALUE_ID:
# Target types not supported for now. (Unclear how this
# would fit with kaa.metadata.)
pass
if targets:
# Assign tags to all listed uids
for target in targets:
try:
self.objects_by_uid[target].tags.update(tags_dict)
self.tags_to_attributes(self.objects_by_uid[target], tags_dict)
except KeyError:
log.warning('Tags assigned to unknown/unsupported target uid %d', target)
else:
self.tags.update(tags_dict)
self.tags_to_attributes(self, tags_dict)
def process_simple_tag(self, simple_tag_elem, tags_dict):
"""
Returns a dict representing the Tag element.
"""
name = lang = value = children = None
binary = False
for elem in self.process_one_level(simple_tag_elem):
elem_id = elem.get_id()
if elem_id == MATROSKA_TAG_NAME_ID:
name = elem.get_utf8().lower()
elif elem_id == MATROSKA_TAG_STRING_ID:
value = elem.get_utf8()
elif elem_id == MATROSKA_TAG_BINARY_ID:
value = elem.get_data()
binary = True
elif elem_id == MATROSKA_TAG_LANGUAGE_ID:
lang = elem.get_utf8()
elif elem_id == MATROSKA_SIMPLE_TAG_ID:
if children is None:
children = core.Tags()
self.process_simple_tag(elem, children)
if children:
# Convert ourselves to a Tags object.
children.value = value
children.langcode = lang
value = children
else:
if name.startswith('date_'):
# Try to convert date to a datetime object.
value = matroska_date_to_datetime(value)
value = core.Tag(value, lang, binary)
if name in tags_dict:
# Multiple items of this tag name.
if not isinstance(tags_dict[name], list):
# Convert to a list
tags_dict[name] = [tags_dict[name]]
# Append to list
tags_dict[name].append(value)
else:
tags_dict[name] = value
def tags_to_attributes(self, obj, tags):
# Convert tags to core attributes.
for name, tag in tags.items():
if isinstance(tag, dict):
# Nested tags dict, recurse.
self.tags_to_attributes(obj, tag)
continue
elif name not in TAGS_MAP:
continue
attr, filter = TAGS_MAP[name]
if attr not in obj._keys and attr not in self._keys:
# Tag is not in any core attribute for this object or global,
# so skip.
continue
# Pull value out of Tag object or list of Tag objects.
value = [item.value for item in tag] if isinstance(tag, list) else tag.value
if filter:
try:
value = [filter(item) for item in value] if isinstance(value, list) else filter(value)
except Exception, e:
log.warning('Failed to convert tag to core attribute: %s', e)
# Special handling for tv series recordings. The 'title' tag
# can be used for both the series and the episode name. The
# same is true for trackno which may refer to the season
# and the episode number. Therefore, if we find these
# attributes already set we try some guessing.
if attr == 'trackno' and getattr(self, attr) is not None:
# delete trackno and save season and episode
self.season = self.trackno
self.episode = value
self.trackno = None
continue
if attr == 'title' and getattr(self, attr) is not None:
# store current value of title as series and use current
# value of title as title
self.series = self.title
if attr in obj._keys:
setattr(obj, attr, value)
else:
setattr(self, attr, value)
Parser = Matroska
libs/enzyme/mp4.py
+476
View File
@@ -0,0 +1,476 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2007 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2007 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import zlib
import logging
import StringIO
import struct
from exceptions import *
import core
# get logging object
log = logging.getLogger(__name__)
# http://developer.apple.com/documentation/QuickTime/QTFF/index.html
# http://developer.apple.com/documentation/QuickTime/QTFF/QTFFChap4/\
# chapter_5_section_2.html#//apple_ref/doc/uid/TP40000939-CH206-BBCBIICE
# Note: May need to define custom log level to work like ATOM_DEBUG did here
QTUDTA = {
'nam': 'title',
'aut': 'artist',
'cpy': 'copyright'
}
QTLANGUAGES = {
0: "en",
1: "fr",
2: "de",
3: "it",
4: "nl",
5: "sv",
6: "es",
7: "da",
8: "pt",
9: "no",
10: "he",
11: "ja",
12: "ar",
13: "fi",
14: "el",
15: "is",
16: "mt",
17: "tr",
18: "hr",
19: "Traditional Chinese",
20: "ur",
21: "hi",
22: "th",
23: "ko",
24: "lt",
25: "pl",
26: "hu",
27: "et",
28: "lv",
29: "Lappish",
30: "fo",
31: "Farsi",
32: "ru",
33: "Simplified Chinese",
34: "Flemish",
35: "ga",
36: "sq",
37: "ro",
38: "cs",
39: "sk",
40: "sl",
41: "yi",
42: "sr",
43: "mk",
44: "bg",
45: "uk",
46: "be",
47: "uz",
48: "kk",
49: "az",
50: "AzerbaijanAr",
51: "hy",
52: "ka",
53: "mo",
54: "ky",
55: "tg",
56: "tk",
57: "mn",
58: "MongolianCyr",
59: "ps",
60: "ku",
61: "ks",
62: "sd",
63: "bo",
64: "ne",
65: "sa",
66: "mr",
67: "bn",
68: "as",
69: "gu",
70: "pa",
71: "or",
72: "ml",
73: "kn",
74: "ta",
75: "te",
76: "si",
77: "my",
78: "Khmer",
79: "lo",
80: "vi",
81: "id",
82: "tl",
83: "MalayRoman",
84: "MalayArabic",
85: "am",
86: "ti",
87: "om",
88: "so",
89: "sw",
90: "Ruanda",
91: "Rundi",
92: "Chewa",
93: "mg",
94: "eo",
128: "cy",
129: "eu",
130: "ca",
131: "la",
132: "qu",
133: "gn",
134: "ay",
135: "tt",
136: "ug",
137: "Dzongkha",
138: "JavaneseRom",
}
class MPEG4(core.AVContainer):
"""
Parser for the MP4 container format. This format is mostly
identical to Apple Quicktime and 3GP files. It maps to mp4, mov,
qt and some other extensions.
"""
table_mapping = {'QTUDTA': QTUDTA}
def __init__(self, file):
core.AVContainer.__init__(self)
self._references = []
self.mime = 'video/quicktime'
self.type = 'Quicktime Video'
h = file.read(8)
try:
(size, type) = struct.unpack('>I4s',h)
except struct.error:
# EOF.
raise ParseError()
if type == 'ftyp':
# file type information
if size >= 12:
# this should always happen
if file.read(4) != 'qt ':
# not a quicktime movie, it is a mpeg4 container
self.mime = 'video/mp4'
self.type = 'MPEG-4 Video'
size -= 4
file.seek(size-8, 1)
h = file.read(8)
(size, type) = struct.unpack('>I4s',h)
while type in ['mdat', 'skip']:
# movie data at the beginning, skip
file.seek(size-8, 1)
h = file.read(8)
(size, type) = struct.unpack('>I4s',h)
if not type in ['moov', 'wide', 'free']:
log.debug('invalid header: %r' % type)
raise ParseError()
# Extended size
if size == 1:
size = struct.unpack('>Q', file.read(8))
# Back over the atom header we just read, since _readatom expects the
# file position to be at the start of an atom.
file.seek(-8, 1)
while self._readatom(file):
pass
if self._references:
self._set('references', self._references)
def _readatom(self, file):
s = file.read(8)
if len(s) < 8:
return 0
atomsize,atomtype = struct.unpack('>I4s', s)
if not str(atomtype).decode('latin1').isalnum():
# stop at nonsense data
return 0
log.debug('%s [%X]' % (atomtype,atomsize))
if atomtype == 'udta':
# Userdata (Metadata)
pos = 0
tabl = {}
i18ntabl = {}
atomdata = file.read(atomsize-8)
while pos < atomsize-12:
(datasize, datatype) = struct.unpack('>I4s', atomdata[pos:pos+8])
if ord(datatype[0]) == 169:
# i18n Metadata...
mypos = 8+pos
while mypos + 4 < datasize+pos:
# first 4 Bytes are i18n header
(tlen, lang) = struct.unpack('>HH', atomdata[mypos:mypos+4])
i18ntabl[lang] = i18ntabl.get(lang, {})
l = atomdata[mypos+4:mypos+tlen+4]
i18ntabl[lang][datatype[1:]] = l
mypos += tlen+4
elif datatype == 'WLOC':
# Drop Window Location
pass
else:
if ord(atomdata[pos+8:pos+datasize][0]) > 1:
tabl[datatype] = atomdata[pos+8:pos+datasize]
pos += datasize
if len(i18ntabl.keys()) > 0:
for k in i18ntabl.keys():
if QTLANGUAGES.has_key(k) and QTLANGUAGES[k] == 'en':
self._appendtable('QTUDTA', i18ntabl[k])
self._appendtable('QTUDTA', tabl)
else:
log.debug('NO i18')
self._appendtable('QTUDTA', tabl)
elif atomtype == 'trak':
atomdata = file.read(atomsize-8)
pos = 0
trackinfo = {}
tracktype = None
while pos < atomsize-8:
(datasize, datatype) = struct.unpack('>I4s', atomdata[pos:pos+8])
if datatype == 'tkhd':
tkhd = struct.unpack('>6I8x4H36xII', atomdata[pos+8:pos+datasize])
trackinfo['width'] = tkhd[10] >> 16
trackinfo['height'] = tkhd[11] >> 16
trackinfo['id'] = tkhd[3]
try:
# XXX Timestamp of Seconds is since January 1st 1904!
# XXX 2082844800 is the difference between Unix and
# XXX Apple time. FIXME to work on Apple, too
self.timestamp = int(tkhd[1]) - 2082844800
except Exception, e:
log.exception('There was trouble extracting timestamp')
elif datatype == 'mdia':
pos += 8
datasize -= 8
log.debug('--> mdia information')
while datasize:
mdia = struct.unpack('>I4s', atomdata[pos:pos+8])
if mdia[1] == 'mdhd':
# Parse based on version of mdhd header. See
# http://wiki.multimedia.cx/index.php?title=QuickTime_container#mdhd
ver = ord(atomdata[pos + 8])
if ver == 0:
mdhd = struct.unpack('>IIIIIhh', atomdata[pos+8:pos+8+24])
elif ver == 1:
mdhd = struct.unpack('>IQQIQhh', atomdata[pos+8:pos+8+36])
else:
mdhd = None
if mdhd:
# duration / time scale
trackinfo['length'] = mdhd[4] / mdhd[3]
if mdhd[5] in QTLANGUAGES:
trackinfo['language'] = QTLANGUAGES[mdhd[5]]
# mdhd[6] == quality
self.length = max(self.length, mdhd[4] / mdhd[3])
elif mdia[1] == 'minf':
# minf has only atoms inside
pos -= (mdia[0] - 8)
datasize += (mdia[0] - 8)
elif mdia[1] == 'stbl':
# stbl has only atoms inside
pos -= (mdia[0] - 8)
datasize += (mdia[0] - 8)
elif mdia[1] == 'hdlr':
hdlr = struct.unpack('>I4s4s', atomdata[pos+8:pos+8+12])
if hdlr[1] == 'mhlr':
if hdlr[2] == 'vide':
tracktype = 'video'
if hdlr[2] == 'soun':
tracktype = 'audio'
elif mdia[1] == 'stsd':
stsd = struct.unpack('>2I', atomdata[pos+8:pos+8+8])
if stsd[1] > 0:
codec = atomdata[pos+16:pos+16+8]
codec = struct.unpack('>I4s', codec)
trackinfo['codec'] = codec[1]
if codec[1] == 'jpeg':
tracktype = 'image'
elif mdia[1] == 'dinf':
dref = struct.unpack('>I4s', atomdata[pos+8:pos+8+8])
log.debug(' --> %s, %s (useless)' % mdia)
if dref[1] == 'dref':
num = struct.unpack('>I', atomdata[pos+20:pos+20+4])[0]
rpos = pos+20+4
for ref in range(num):
# FIXME: do somthing if this references
ref = struct.unpack('>I3s', atomdata[rpos:rpos+7])
data = atomdata[rpos+7:rpos+ref[0]]
rpos += ref[0]
else:
if mdia[1].startswith('st'):
log.debug(' --> %s, %s (sample)' % mdia)
elif mdia[1] == 'vmhd' and not tracktype:
# indicates that this track is video
tracktype = 'video'
elif mdia[1] in ['vmhd', 'smhd'] and not tracktype:
# indicates that this track is audio
tracktype = 'audio'
else:
log.debug(' --> %s, %s (unknown)' % mdia)
pos += mdia[0]
datasize -= mdia[0]
elif datatype == 'udta':
log.debug(struct.unpack('>I4s', atomdata[:8]))
else:
if datatype == 'edts':
log.debug('--> %s [%d] (edit list)' % \
(datatype, datasize))
else:
log.debug('--> %s [%d] (unknown)' % \
(datatype, datasize))
pos += datasize
info = None
if tracktype == 'video':
info = core.VideoStream()
self.video.append(info)
if tracktype == 'audio':
info = core.AudioStream()
self.audio.append(info)
if info:
for key, value in trackinfo.items():
setattr(info, key, value)
elif atomtype == 'mvhd':
# movie header
mvhd = struct.unpack('>6I2h', file.read(28))
self.length = max(self.length, mvhd[4] / mvhd[3])
self.volume = mvhd[6]
file.seek(atomsize-8-28,1)
elif atomtype == 'cmov':
# compressed movie
datasize, atomtype = struct.unpack('>I4s', file.read(8))
if not atomtype == 'dcom':
return atomsize
method = struct.unpack('>4s', file.read(datasize-8))[0]
datasize, atomtype = struct.unpack('>I4s', file.read(8))
if not atomtype == 'cmvd':
return atomsize
if method == 'zlib':
data = file.read(datasize-8)
try:
decompressed = zlib.decompress(data)
except Exception, e:
try:
decompressed = zlib.decompress(data[4:])
except Exception, e:
log.exception('There was a proble decompressiong atom')
return atomsize
decompressedIO = StringIO.StringIO(decompressed)
while self._readatom(decompressedIO):
pass
else:
log.info('unknown compression %s' % method)
# unknown compression method
file.seek(datasize-8,1)
elif atomtype == 'moov':
# decompressed movie info
while self._readatom(file):
pass
elif atomtype == 'mdat':
pos = file.tell() + atomsize - 8
# maybe there is data inside the mdat
log.info('parsing mdat')
while self._readatom(file):
pass
log.info('end of mdat')
file.seek(pos, 0)
elif atomtype == 'rmra':
# reference list
while self._readatom(file):
pass
elif atomtype == 'rmda':
# reference
atomdata = file.read(atomsize-8)
pos = 0
url = ''
quality = 0
datarate = 0
while pos < atomsize-8:
(datasize, datatype) = struct.unpack('>I4s', atomdata[pos:pos+8])
if datatype == 'rdrf':
rflags, rtype, rlen = struct.unpack('>I4sI', atomdata[pos+8:pos+20])
if rtype == 'url ':
url = atomdata[pos+20:pos+20+rlen]
if url.find('\0') > 0:
url = url[:url.find('\0')]
elif datatype == 'rmqu':
quality = struct.unpack('>I', atomdata[pos+8:pos+12])[0]
elif datatype == 'rmdr':
datarate = struct.unpack('>I', atomdata[pos+12:pos+16])[0]
pos += datasize
if url:
self._references.append((url, quality, datarate))
else:
if not atomtype in ['wide', 'free']:
log.info('unhandled base atom %s' % atomtype)
# Skip unknown atoms
try:
file.seek(atomsize-8,1)
except IOError:
return 0
return atomsize
Parser = MPEG4
libs/enzyme/mpeg.py
+915
View File
@@ -0,0 +1,915 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import os
import struct
import logging
import stat
from exceptions import *
import core
# get logging object
log = logging.getLogger(__name__)
##------------------------------------------------------------------------
## START_CODE
##
## Start Codes, with 'slice' occupying 0x01..0xAF
##------------------------------------------------------------------------
START_CODE = {
0x00 : 'picture_start_code',
0xB0 : 'reserved',
0xB1 : 'reserved',
0xB2 : 'user_data_start_code',
0xB3 : 'sequence_header_code',
0xB4 : 'sequence_error_code',
0xB5 : 'extension_start_code',
0xB6 : 'reserved',
0xB7 : 'sequence end',
0xB8 : 'group of pictures',
}
for i in range(0x01,0xAF):
START_CODE[i] = 'slice_start_code'
##------------------------------------------------------------------------
## START CODES
##------------------------------------------------------------------------
PICTURE = 0x00
USERDATA = 0xB2
SEQ_HEAD = 0xB3
SEQ_ERR = 0xB4
EXT_START = 0xB5
SEQ_END = 0xB7
GOP = 0xB8
SEQ_START_CODE = 0xB3
PACK_PKT = 0xBA
SYS_PKT = 0xBB
PADDING_PKT = 0xBE
AUDIO_PKT = 0xC0
VIDEO_PKT = 0xE0
PRIVATE_STREAM1 = 0xBD
PRIVATE_STREAM2 = 0xBf
TS_PACKET_LENGTH = 188
TS_SYNC = 0x47
##------------------------------------------------------------------------
## FRAME_RATE
##
## A lookup table of all the standard frame rates. Some rates adhere to
## a particular profile that ensures compatibility with VLSI capabilities
## of the early to mid 1990s.
##
## CPB
## Constrained Parameters Bitstreams, an MPEG-1 set of sampling and
## bitstream parameters designed to normalize decoder computational
## complexity, buffer size, and memory bandwidth while still addressing
## the widest possible range of applications.
##
## Main Level
## MPEG-2 Video Main Profile and Main Level is analogous to MPEG-1's
## CPB, with sampling limits at CCIR 601 parameters (720x480x30 Hz or
## 720x576x24 Hz).
##
##------------------------------------------------------------------------
FRAME_RATE = [
0,
24000.0/1001, ## 3-2 pulldown NTSC (CPB/Main Level)
24, ## Film (CPB/Main Level)
25, ## PAL/SECAM or 625/60 video
30000.0/1001, ## NTSC (CPB/Main Level)
30, ## drop-frame NTSC or component 525/60 (CPB/Main Level)
50, ## double-rate PAL
60000.0/1001, ## double-rate NTSC
60, ## double-rate, drop-frame NTSC/component 525/60 video
]
##------------------------------------------------------------------------
## ASPECT_RATIO -- INCOMPLETE?
##
## This lookup table maps the header aspect ratio index to a float value.
## These are just the defined ratios for CPB I believe. As I understand
## it, a stream that doesn't adhere to one of these aspect ratios is
## technically considered non-compliant.
##------------------------------------------------------------------------
ASPECT_RATIO = ( None, # Forbidden
1.0, # 1/1 (VGA)
4.0 / 3, # 4/3 (TV)
16.0 / 9, # 16/9 (Widescreen)
2.21 # (Cinema)
)
class MPEG(core.AVContainer):
"""
Parser for various MPEG files. This includes MPEG-1 and MPEG-2
program streams, elementary streams and transport streams. The
reported length differs from the length reported by most video
players but the provides length here is correct. An MPEG file has
no additional metadata like title, etc; only codecs, length and
resolution is reported back.
"""
def __init__(self,file):
core.AVContainer.__init__(self)
self.sequence_header_offset = 0
self.mpeg_version = 2
# detect TS (fast scan)
if not self.isTS(file):
# detect system mpeg (many infos)
if not self.isMPEG(file):
# detect PES
if not self.isPES(file):
# Maybe it's MPEG-ES
if self.isES(file):
# If isES() succeeds, we needn't do anything further.
return
if file.name.lower().endswith('mpeg') or \
file.name.lower().endswith('mpg'):
# This has to be an mpeg file. It could be a bad
# recording from an ivtv based hardware encoder with
# same bytes missing at the beginning.
# Do some more digging...
if not self.isMPEG(file, force=True) or \
not self.video or not self.audio:
# does not look like an mpeg at all
raise ParseError()
else:
# no mpeg at all
raise ParseError()
self.mime = 'video/mpeg'
if not self.video:
self.video.append(core.VideoStream())
if self.sequence_header_offset <= 0:
return
self.progressive(file)
for vi in self.video:
vi.width, vi.height = self.dxy(file)
vi.fps, vi.aspect = self.framerate_aspect(file)
vi.bitrate = self.bitrate(file)
if self.length:
vi.length = self.length
if not self.type:
self.type = 'MPEG Video'
# set fourcc codec for video and audio
vc, ac = 'MP2V', 'MP2A'
if self.mpeg_version == 1:
vc, ac = 'MPEG', 0x0050
for v in self.video:
v.codec = vc
for a in self.audio:
if not a.codec:
a.codec = ac
def dxy(self,file):
"""
get width and height of the video
"""
file.seek(self.sequence_header_offset+4,0)
v = file.read(4)
x = struct.unpack('>H',v[:2])[0] >> 4
y = struct.unpack('>H',v[1:3])[0] & 0x0FFF
return (x,y)
def framerate_aspect(self,file):
"""
read framerate and aspect ratio
"""
file.seek(self.sequence_header_offset+7,0)
v = struct.unpack( '>B', file.read(1) )[0]
try:
fps = FRAME_RATE[v&0xf]
except IndexError:
fps = None
if v>>4 < len(ASPECT_RATIO):
aspect = ASPECT_RATIO[v>>4]
else:
aspect = None
return (fps, aspect)
def progressive(self, file):
"""
Try to find out with brute force if the mpeg is interlaced or not.
Search for the Sequence_Extension in the extension header (01B5)
"""
file.seek(0)
buffer = ''
count = 0
while 1:
if len(buffer) < 1000:
count += 1
if count > 1000:
break
buffer += file.read(1024)
if len(buffer) < 1000:
break
pos = buffer.find('\x00\x00\x01\xb5')
if pos == -1 or len(buffer) - pos < 5:
buffer = buffer[-10:]
continue
ext = (ord(buffer[pos+4]) >> 4)
if ext == 8:
pass
elif ext == 1:
if (ord(buffer[pos+5]) >> 3) & 1:
self._set('progressive', True)
else:
self._set('interlaced', True)
return True
else:
log.debug('ext', ext)
buffer = buffer[pos+4:]
return False
##------------------------------------------------------------------------
## bitrate()
##
## From the MPEG-2.2 spec:
##
## bit_rate -- This is a 30-bit integer. The lower 18 bits of the
## integer are in bit_rate_value and the upper 12 bits are in
## bit_rate_extension. The 30-bit integer specifies the bitrate of the
## bitstream measured in units of 400 bits/second, rounded upwards.
## The value zero is forbidden.
##
## So ignoring all the variable bitrate stuff for now, this 30 bit integer
## multiplied times 400 bits/sec should give the rate in bits/sec.
##
## TODO: Variable bitrates? I need one that implements this.
##
## Continued from the MPEG-2.2 spec:
##
## If the bitstream is a constant bitrate stream, the bitrate specified
## is the actual rate of operation of the VBV specified in annex C. If
## the bitstream is a variable bitrate stream, the STD specifications in
## ISO/IEC 13818-1 supersede the VBV, and the bitrate specified here is
## used to dimension the transport stream STD (2.4.2 in ITU-T Rec. xxx |
## ISO/IEC 13818-1), or the program stream STD (2.4.5 in ITU-T Rec. xxx |
## ISO/IEC 13818-1).
##
## If the bitstream is not a constant rate bitstream the vbv_delay
## field shall have the value FFFF in hexadecimal.
##
## Given the value encoded in the bitrate field, the bitstream shall be
## generated so that the video encoding and the worst case multiplex
## jitter do not cause STD buffer overflow or underflow.
##
##
##------------------------------------------------------------------------
## Some parts in the code are based on mpgtx (mpgtx.sf.net)
def bitrate(self,file):
"""
read the bitrate (most of the time broken)
"""
file.seek(self.sequence_header_offset+8,0)
t,b = struct.unpack( '>HB', file.read(3) )
vrate = t << 2 | b >> 6
return vrate * 400
def ReadSCRMpeg2(self, buffer):
"""
read SCR (timestamp) for MPEG2 at the buffer beginning (6 Bytes)
"""
if len(buffer) < 6:
return None
highbit = (ord(buffer[0])&0x20)>>5
low4Bytes= ((long(ord(buffer[0])) & 0x18) >> 3) << 30
low4Bytes |= (ord(buffer[0]) & 0x03) << 28
low4Bytes |= ord(buffer[1]) << 20
low4Bytes |= (ord(buffer[2]) & 0xF8) << 12
low4Bytes |= (ord(buffer[2]) & 0x03) << 13
low4Bytes |= ord(buffer[3]) << 5
low4Bytes |= (ord(buffer[4])) >> 3
sys_clock_ref=(ord(buffer[4]) & 0x3) << 7
sys_clock_ref|=(ord(buffer[5]) >> 1)
return (long(highbit * (1<<16) * (1<<16)) + low4Bytes) / 90000
def ReadSCRMpeg1(self, buffer):
"""
read SCR (timestamp) for MPEG1 at the buffer beginning (5 Bytes)
"""
if len(buffer) < 5:
return None
highbit = (ord(buffer[0]) >> 3) & 0x01
low4Bytes = ((long(ord(buffer[0])) >> 1) & 0x03) << 30
low4Bytes |= ord(buffer[1]) << 22;
low4Bytes |= (ord(buffer[2]) >> 1) << 15;
low4Bytes |= ord(buffer[3]) << 7;
low4Bytes |= ord(buffer[4]) >> 1;
return (long(highbit) * (1<<16) * (1<<16) + low4Bytes) / 90000;
def ReadPTS(self, buffer):
"""
read PTS (PES timestamp) at the buffer beginning (5 Bytes)
"""
high = ((ord(buffer[0]) & 0xF) >> 1)
med = (ord(buffer[1]) << 7) + (ord(buffer[2]) >> 1)
low = (ord(buffer[3]) << 7) + (ord(buffer[4]) >> 1)
return ((long(high) << 30 ) + (med << 15) + low) / 90000
def ReadHeader(self, buffer, offset):
"""
Handle MPEG header in buffer on position offset
Return None on error, new offset or 0 if the new offset can't be scanned
"""
if buffer[offset:offset+3] != '\x00\x00\x01':
return None
id = ord(buffer[offset+3])
if id == PADDING_PKT:
return offset + (ord(buffer[offset+4]) << 8) + \
ord(buffer[offset+5]) + 6
if id == PACK_PKT:
if ord(buffer[offset+4]) & 0xF0 == 0x20:
self.type = 'MPEG-1 Video'
self.get_time = self.ReadSCRMpeg1
self.mpeg_version = 1
return offset + 12
elif (ord(buffer[offset+4]) & 0xC0) == 0x40:
self.type = 'MPEG-2 Video'
self.get_time = self.ReadSCRMpeg2
return offset + (ord(buffer[offset+13]) & 0x07) + 14
else:
# I have no idea what just happened, but for some DVB
# recordings done with mencoder this points to a
# PACK_PKT describing something odd. Returning 0 here
# (let's hope there are no extensions in the header)
# fixes it.
return 0
if 0xC0 <= id <= 0xDF:
# code for audio stream
for a in self.audio:
if a.id == id:
break
else:
self.audio.append(core.AudioStream())
self.audio[-1]._set('id', id)
return 0
if 0xE0 <= id <= 0xEF:
# code for video stream
for v in self.video:
if v.id == id:
break
else:
self.video.append(core.VideoStream())
self.video[-1]._set('id', id)
return 0
if id == SEQ_HEAD:
# sequence header, remember that position for later use
self.sequence_header_offset = offset
return 0
if id in [PRIVATE_STREAM1, PRIVATE_STREAM2]:
# private stream. we don't know, but maybe we can guess later
add = ord(buffer[offset+8])
# if (ord(buffer[offset+6]) & 4) or 1:
# id = ord(buffer[offset+10+add])
if buffer[offset+11+add:offset+15+add].find('\x0b\x77') != -1:
# AC3 stream
for a in self.audio:
if a.id == id:
break
else:
self.audio.append(core.AudioStream())
self.audio[-1]._set('id', id)
self.audio[-1].codec = 0x2000 # AC3
return 0
if id == SYS_PKT:
return 0
if id == EXT_START:
return 0
return 0
# Normal MPEG (VCD, SVCD) ========================================
def isMPEG(self, file, force=False):
"""
This MPEG starts with a sequence of 0x00 followed by a PACK Header
http://dvd.sourceforge.net/dvdinfo/packhdr.html
"""
file.seek(0,0)
buffer = file.read(10000)
offset = 0
# seek until the 0 byte stop
while offset < len(buffer)-100 and buffer[offset] == '\0':
offset += 1
offset -= 2
# test for mpeg header 0x00 0x00 0x01
header = '\x00\x00\x01%s' % chr(PACK_PKT)
if offset < 0 or not buffer[offset:offset+4] == header:
if not force:
return 0
# brute force and try to find the pack header in the first
# 10000 bytes somehow
offset = buffer.find(header)
if offset < 0:
return 0
# scan the 100000 bytes of data
buffer += file.read(100000)
# scan first header, to get basic info about
# how to read a timestamp
self.ReadHeader(buffer, offset)
# store first timestamp
self.start = self.get_time(buffer[offset+4:])
while len(buffer) > offset + 1000 and \
buffer[offset:offset+3] == '\x00\x00\x01':
# read the mpeg header
new_offset = self.ReadHeader(buffer, offset)
# header scanning detected error, this is no mpeg
if new_offset == None:
return 0
if new_offset:
# we have a new offset
offset = new_offset
# skip padding 0 before a new header
while len(buffer) > offset + 10 and \
not ord(buffer[offset+2]):
offset += 1
else:
# seek to new header by brute force
offset += buffer[offset+4:].find('\x00\x00\x01') + 4
# fill in values for support functions:
self.__seek_size__ = 1000000
self.__sample_size__ = 10000
self.__search__ = self._find_timer_
self.filename = file.name
# get length of the file
self.length = self.get_length()
return 1
def _find_timer_(self, buffer):
"""
Return position of timer in buffer or None if not found.
This function is valid for 'normal' mpeg files
"""
pos = buffer.find('\x00\x00\x01%s' % chr(PACK_PKT))
if pos == -1:
return None
return pos + 4
# PES ============================================================
def ReadPESHeader(self, offset, buffer, id=0):
"""
Parse a PES header.
Since it starts with 0x00 0x00 0x01 like 'normal' mpegs, this
function will return (0, None) when it is no PES header or
(packet length, timestamp position (maybe None))
http://dvd.sourceforge.net/dvdinfo/pes-hdr.html
"""
if not buffer[0:3] == '\x00\x00\x01':
return 0, None
packet_length = (ord(buffer[4]) << 8) + ord(buffer[5]) + 6
align = ord(buffer[6]) & 4
header_length = ord(buffer[8])
# PES ID (starting with 001)
if ord(buffer[3]) & 0xE0 == 0xC0:
id = id or ord(buffer[3]) & 0x1F
for a in self.audio:
if a.id == id:
break
else:
self.audio.append(core.AudioStream())
self.audio[-1]._set('id', id)
elif ord(buffer[3]) & 0xF0 == 0xE0:
id = id or ord(buffer[3]) & 0xF
for v in self.video:
if v.id == id:
break
else:
self.video.append(core.VideoStream())
self.video[-1]._set('id', id)
# new mpeg starting
if buffer[header_length+9:header_length+13] == \
'\x00\x00\x01\xB3' and not self.sequence_header_offset:
# yes, remember offset for later use
self.sequence_header_offset = offset + header_length+9
elif ord(buffer[3]) == 189 or ord(buffer[3]) == 191:
# private stream. we don't know, but maybe we can guess later
id = id or ord(buffer[3]) & 0xF
if align and \
buffer[header_length+9:header_length+11] == '\x0b\x77':
# AC3 stream
for a in self.audio:
if a.id == id:
break
else:
self.audio.append(core.AudioStream())
self.audio[-1]._set('id', id)
self.audio[-1].codec = 0x2000 # AC3
else:
# unknown content
pass
ptsdts = ord(buffer[7]) >> 6
if ptsdts and ptsdts == ord(buffer[9]) >> 4:
if ord(buffer[9]) >> 4 != ptsdts:
log.warning('WARNING: bad PTS/DTS, please contact us')
return packet_length, None
# timestamp = self.ReadPTS(buffer[9:14])
high = ((ord(buffer[9]) & 0xF) >> 1)
med = (ord(buffer[10]) << 7) + (ord(buffer[11]) >> 1)
low = (ord(buffer[12]) << 7) + (ord(buffer[13]) >> 1)
return packet_length, 9
return packet_length, None
def isPES(self, file):
log.info('trying mpeg-pes scan')
file.seek(0,0)
buffer = file.read(3)
# header (also valid for all mpegs)
if not buffer == '\x00\x00\x01':
return 0
self.sequence_header_offset = 0
buffer += file.read(10000)
offset = 0
while offset + 1000 < len(buffer):
pos, timestamp = self.ReadPESHeader(offset, buffer[offset:])
if not pos:
return 0
if timestamp != None and not hasattr(self, 'start'):
self.get_time = self.ReadPTS
bpos = buffer[offset+timestamp:offset+timestamp+5]
self.start = self.get_time(bpos)
if self.sequence_header_offset and hasattr(self, 'start'):
# we have all informations we need
break
offset += pos
if offset + 1000 < len(buffer) and len(buffer) < 1000000 or 1:
# looks like a pes, read more
buffer += file.read(10000)
if not self.video and not self.audio:
# no video and no audio?
return 0
self.type = 'MPEG-PES'
# fill in values for support functions:
self.__seek_size__ = 10000000 # 10 MB
self.__sample_size__ = 500000 # 500 k scanning
self.__search__ = self._find_timer_PES_
self.filename = file.name
# get length of the file
self.length = self.get_length()
return 1
def _find_timer_PES_(self, buffer):
"""
Return position of timer in buffer or -1 if not found.
This function is valid for PES files
"""
pos = buffer.find('\x00\x00\x01')
offset = 0
if pos == -1 or offset + 1000 >= len(buffer):
return None
retpos = -1
ackcount = 0
while offset + 1000 < len(buffer):
pos, timestamp = self.ReadPESHeader(offset, buffer[offset:])
if timestamp != None and retpos == -1:
retpos = offset + timestamp
if pos == 0:
# Oops, that was a mpeg header, no PES header
offset += buffer[offset:].find('\x00\x00\x01')
retpos = -1
ackcount = 0
else:
offset += pos
if retpos != -1:
ackcount += 1
if ackcount > 10:
# looks ok to me
return retpos
return None
# Elementary Stream ===============================================
def isES(self, file):
file.seek(0, 0)
try:
header = struct.unpack('>LL', file.read(8))
except (struct.error, IOError):
return False
if header[0] != 0x1B3:
return False
# Is an mpeg video elementary stream
self.mime = 'video/mpeg'
video = core.VideoStream()
video.width = header[1] >> 20
video.height = (header[1] >> 8) & 0xfff
if header[1] & 0xf < len(FRAME_RATE):
video.fps = FRAME_RATE[header[1] & 0xf]
if (header[1] >> 4) & 0xf < len(ASPECT_RATIO):
# FIXME: Empirically the aspect looks like PAR rather than DAR
video.aspect = ASPECT_RATIO[(header[1] >> 4) & 0xf]
self.video.append(video)
return True
# Transport Stream ===============================================
def isTS(self, file):
file.seek(0,0)
buffer = file.read(TS_PACKET_LENGTH * 2)
c = 0
while c + TS_PACKET_LENGTH < len(buffer):
if ord(buffer[c]) == ord(buffer[c+TS_PACKET_LENGTH]) == TS_SYNC:
break
c += 1
else:
return 0
buffer += file.read(10000)
self.type = 'MPEG-TS'
while c + TS_PACKET_LENGTH < len(buffer):
start = ord(buffer[c+1]) & 0x40
# maybe load more into the buffer
if c + 2 * TS_PACKET_LENGTH > len(buffer) and c < 500000:
buffer += file.read(10000)
# wait until the ts payload contains a payload header
if not start:
c += TS_PACKET_LENGTH
continue
tsid = ((ord(buffer[c+1]) & 0x3F) << 8) + ord(buffer[c+2])
adapt = (ord(buffer[c+3]) & 0x30) >> 4
offset = 4
if adapt & 0x02:
# meta info present, skip it for now
adapt_len = ord(buffer[c+offset])
offset += adapt_len + 1
if not ord(buffer[c+1]) & 0x40:
# no new pes or psi in stream payload starting
pass
elif adapt & 0x01:
# PES
timestamp = self.ReadPESHeader(c+offset, buffer[c+offset:],
tsid)[1]
if timestamp != None:
if not hasattr(self, 'start'):
self.get_time = self.ReadPTS
timestamp = c + offset + timestamp
self.start = self.get_time(buffer[timestamp:timestamp+5])
elif not hasattr(self, 'audio_ok'):
timestamp = c + offset + timestamp
start = self.get_time(buffer[timestamp:timestamp+5])
if start is not None and self.start is not None and \
abs(start - self.start) < 10:
# looks ok
self.audio_ok = True
else:
# timestamp broken
del self.start
log.warning('Timestamp error, correcting')
if hasattr(self, 'start') and self.start and \
self.sequence_header_offset and self.video and self.audio:
break
c += TS_PACKET_LENGTH
if not self.sequence_header_offset:
return 0
# fill in values for support functions:
self.__seek_size__ = 10000000 # 10 MB
self.__sample_size__ = 100000 # 100 k scanning
self.__search__ = self._find_timer_TS_
self.filename = file.name
# get length of the file
self.length = self.get_length()
return 1
def _find_timer_TS_(self, buffer):
c = 0
while c + TS_PACKET_LENGTH < len(buffer):
if ord(buffer[c]) == ord(buffer[c+TS_PACKET_LENGTH]) == TS_SYNC:
break
c += 1
else:
return None
while c + TS_PACKET_LENGTH < len(buffer):
start = ord(buffer[c+1]) & 0x40
if not start:
c += TS_PACKET_LENGTH
continue
tsid = ((ord(buffer[c+1]) & 0x3F) << 8) + ord(buffer[c+2])
adapt = (ord(buffer[c+3]) & 0x30) >> 4
offset = 4
if adapt & 0x02:
# meta info present, skip it for now
offset += ord(buffer[c+offset]) + 1
if adapt & 0x01:
timestamp = self.ReadPESHeader(c+offset, buffer[c+offset:], tsid)[1]
if timestamp is None:
# this should not happen
log.error('bad TS')
return None
return c + offset + timestamp
c += TS_PACKET_LENGTH
return None
# Support functions ==============================================
def get_endpos(self):
"""
get the last timestamp of the mpeg, return -1 if this is not possible
"""
if not hasattr(self, 'filename') or not hasattr(self, 'start'):
return None
length = os.stat(self.filename)[stat.ST_SIZE]
if length < self.__sample_size__:
return
file = open(self.filename)
file.seek(length - self.__sample_size__)
buffer = file.read(self.__sample_size__)
end = None
while 1:
pos = self.__search__(buffer)
if pos == None:
break
end = self.get_time(buffer[pos:]) or end
buffer = buffer[pos+100:]
file.close()
return end
def get_length(self):
"""
get the length in seconds, return -1 if this is not possible
"""
end = self.get_endpos()
if end == None or self.start == None:
return None
if self.start > end:
return int(((long(1) << 33) - 1 ) / 90000) - self.start + end
return end - self.start
def seek(self, end_time):
"""
Return the byte position in the file where the time position
is 'pos' seconds. Return 0 if this is not possible
"""
if not hasattr(self, 'filename') or not hasattr(self, 'start'):
return 0
file = open(self.filename)
seek_to = 0
while 1:
file.seek(self.__seek_size__, 1)
buffer = file.read(self.__sample_size__)
if len(buffer) < 10000:
break
pos = self.__search__(buffer)
if pos != None:
# found something
nt = self.get_time(buffer[pos:])
if nt is not None and nt >= end_time:
# too much, break
break
# that wasn't enough
seek_to = file.tell()
file.close()
return seek_to
def __scan__(self):
"""
scan file for timestamps (may take a long time)
"""
if not hasattr(self, 'filename') or not hasattr(self, 'start'):
return 0
file = open(self.filename)
log.debug('scanning file...')
while 1:
file.seek(self.__seek_size__ * 10, 1)
buffer = file.read(self.__sample_size__)
if len(buffer) < 10000:
break
pos = self.__search__(buffer)
if pos == None:
continue
log.debug('buffer position: %s' % self.get_time(buffer[pos:]))
file.close()
log.debug('done scanning file')
Parser = MPEG
libs/enzyme/ogm.py
+301
View File
@@ -0,0 +1,301 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import struct
import re
import stat
import os
import logging
from exceptions import *
import core
# get logging object
log = logging.getLogger(__name__)
PACKET_TYPE_HEADER = 0x01
PACKED_TYPE_METADATA = 0x03
PACKED_TYPE_SETUP = 0x05
PACKET_TYPE_BITS = 0x07
PACKET_IS_SYNCPOINT = 0x08
#VORBIS_VIDEO_PACKET_INFO = 'video'
STREAM_HEADER_VIDEO = '<4sIQQIIHII'
STREAM_HEADER_AUDIO = '<4sIQQIIHHHI'
VORBISCOMMENT = { 'TITLE': 'title',
'ALBUM': 'album',
'ARTIST': 'artist',
'COMMENT': 'comment',
'ENCODER': 'encoder',
'TRACKNUMBER': 'trackno',
'LANGUAGE': 'language',
'GENRE': 'genre',
}
# FIXME: check VORBISCOMMENT date and convert to timestamp
# Deactived tag: 'DATE': 'date',
MAXITERATIONS = 30
class Ogm(core.AVContainer):
table_mapping = { 'VORBISCOMMENT' : VORBISCOMMENT }
def __init__(self, file):
core.AVContainer.__init__(self)
self.samplerate = 1
self.all_streams = [] # used to add meta data to streams
self.all_header = []
for i in range(MAXITERATIONS):
granule, nextlen = self._parseOGGS(file)
if granule == None:
if i == 0:
# oops, bad file
raise ParseError()
break
elif granule > 0:
# ok, file started
break
# seek to the end of the stream, to avoid scanning the whole file
if (os.stat(file.name)[stat.ST_SIZE] > 50000):
file.seek(os.stat(file.name)[stat.ST_SIZE]-49000)
# read the rest of the file into a buffer
h = file.read()
# find last OggS to get length info
if len(h) > 200:
idx = h.find('OggS')
pos = -49000 + idx
if idx:
file.seek(os.stat(file.name)[stat.ST_SIZE] + pos)
while 1:
granule, nextlen = self._parseOGGS(file)
if not nextlen:
break
# Copy metadata to the streams
if len(self.all_header) == len(self.all_streams):
for i in range(len(self.all_header)):
# get meta info
for key in self.all_streams[i].keys():
if self.all_header[i].has_key(key):
self.all_streams[i][key] = self.all_header[i][key]
del self.all_header[i][key]
if self.all_header[i].has_key(key.upper()):
asi = self.all_header[i][key.upper()]
self.all_streams[i][key] = asi
del self.all_header[i][key.upper()]
# Chapter parser
if self.all_header[i].has_key('CHAPTER01') and \
not self.chapters:
while 1:
s = 'CHAPTER%02d' % (len(self.chapters) + 1)
if self.all_header[i].has_key(s) and \
self.all_header[i].has_key(s + 'NAME'):
pos = self.all_header[i][s]
try:
pos = int(pos)
except ValueError:
new_pos = 0
for v in pos.split(':'):
new_pos = new_pos * 60 + float(v)
pos = int(new_pos)
c = self.all_header[i][s + 'NAME']
c = core.Chapter(c, pos)
del self.all_header[i][s + 'NAME']
del self.all_header[i][s]
self.chapters.append(c)
else:
break
# If there are no video streams in this ogg container, it
# must be an audio file. Raise an exception to cause the
# factory to fall back to audio.ogg.
if len(self.video) == 0:
raise ParseError
# Copy Metadata from tables into the main set of attributes
for header in self.all_header:
self._appendtable('VORBISCOMMENT', header)
def _parseOGGS(self,file):
h = file.read(27)
if len(h) == 0:
# Regular File end
return None, None
elif len(h) < 27:
log.debug("%d Bytes of Garbage found after End." % len(h))
return None, None
if h[:4] != "OggS":
log.debug("Invalid Ogg")
raise ParseError()
version = ord(h[4])
if version != 0:
log.debug("Unsupported OGG/OGM Version %d." % version)
return None, None
head = struct.unpack('<BQIIIB', h[5:])
headertype, granulepos, serial, pageseqno, checksum, \
pageSegCount = head
self.mime = 'application/ogm'
self.type = 'OGG Media'
tab = file.read(pageSegCount)
nextlen = 0
for i in range(len(tab)):
nextlen += ord(tab[i])
else:
h = file.read(1)
packettype = ord(h[0]) & PACKET_TYPE_BITS
if packettype == PACKET_TYPE_HEADER:
h += file.read(nextlen-1)
self._parseHeader(h, granulepos)
elif packettype == PACKED_TYPE_METADATA:
h += file.read(nextlen-1)
self._parseMeta(h)
else:
file.seek(nextlen-1,1)
if len(self.all_streams) > serial:
stream = self.all_streams[serial]
if hasattr(stream, 'samplerate') and \
stream.samplerate:
stream.length = granulepos / stream.samplerate
elif hasattr(stream, 'bitrate') and \
stream.bitrate:
stream.length = granulepos / stream.bitrate
return granulepos, nextlen + 27 + pageSegCount
def _parseMeta(self,h):
flags = ord(h[0])
headerlen = len(h)
if headerlen >= 7 and h[1:7] == 'vorbis':
header = {}
nextlen, self.encoder = self._extractHeaderString(h[7:])
numItems = struct.unpack('<I',h[7+nextlen:7+nextlen+4])[0]
start = 7+4+nextlen
for i in range(numItems):
(nextlen, s) = self._extractHeaderString(h[start:])
start += nextlen
if s:
a = re.split('=',s)
header[(a[0]).upper()]=a[1]
# Put Header fields into info fields
self.type = 'OGG Vorbis'
self.subtype = ''
self.all_header.append(header)
def _parseHeader(self,header,granule):
headerlen = len(header)
flags = ord(header[0])
if headerlen >= 30 and header[1:7] == 'vorbis':
ai = core.AudioStream()
ai.version, ai.channels, ai.samplerate, bitrate_max, ai.bitrate, \
bitrate_min, blocksize, framing = \
struct.unpack('<IBIiiiBB',header[7:7+23])
ai.codec = 'Vorbis'
#ai.granule = granule
#ai.length = granule / ai.samplerate
self.audio.append(ai)
self.all_streams.append(ai)
elif headerlen >= 7 and header[1:7] == 'theora':
# Theora Header
# XXX Finish Me
vi = core.VideoStream()
vi.codec = 'theora'
self.video.append(vi)
self.all_streams.append(vi)
elif headerlen >= 142 and \
header[1:36] == 'Direct Show Samples embedded in Ogg':
# Old Directshow format
# XXX Finish Me
vi = core.VideoStream()
vi.codec = 'dshow'
self.video.append(vi)
self.all_streams.append(vi)
elif flags & PACKET_TYPE_BITS == PACKET_TYPE_HEADER and \
headerlen >= struct.calcsize(STREAM_HEADER_VIDEO)+1:
# New Directshow Format
htype = header[1:9]
if htype[:5] == 'video':
sh = header[9:struct.calcsize(STREAM_HEADER_VIDEO)+9]
streamheader = struct.unpack(STREAM_HEADER_VIDEO, sh)
vi = core.VideoStream()
(type, ssize, timeunit, samplerate, vi.length, buffersize, \
vi.bitrate, vi.width, vi.height) = streamheader
vi.width /= 65536
vi.height /= 65536
# XXX length, bitrate are very wrong
vi.codec = type
vi.fps = 10000000 / timeunit
self.video.append(vi)
self.all_streams.append(vi)
elif htype[:5] == 'audio':
sha = header[9:struct.calcsize(STREAM_HEADER_AUDIO)+9]
streamheader = struct.unpack(STREAM_HEADER_AUDIO, sha)
ai = core.AudioStream()
(type, ssize, timeunit, ai.samplerate, ai.length, buffersize, \
ai.bitrate, ai.channels, bloc, ai.bitrate) = streamheader
self.samplerate = ai.samplerate
log.debug("Samplerate %d" % self.samplerate)
self.audio.append(ai)
self.all_streams.append(ai)
elif htype[:4] == 'text':
subtitle = core.Subtitle()
# FIXME: add more info
self.subtitles.append(subtitle)
self.all_streams.append(subtitle)
else:
log.debug("Unknown Header")
def _extractHeaderString(self,header):
len = struct.unpack('<I', header[:4])[0]
try:
return (len+4,unicode(header[4:4+len], 'utf-8'))
except (KeyError, IndexError, UnicodeDecodeError):
return (len+4,None)
Parser = Ogm
libs/enzyme/real.py
+120
View File
@@ -0,0 +1,120 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import struct
import logging
from exceptions import *
import core
# http://www.pcisys.net/~melanson/codecs/rmff.htm
# http://www.pcisys.net/~melanson/codecs/
# get logging object
log = logging.getLogger(__name__)
class RealVideo(core.AVContainer):
def __init__(self,file):
core.AVContainer.__init__(self)
self.mime = 'video/real'
self.type = 'Real Video'
h = file.read(10)
try:
(object_id,object_size,object_version) = struct.unpack('>4sIH',h)
except struct.error:
# EOF.
raise ParseError()
if not object_id == '.RMF':
raise ParseError()
file_version, num_headers = struct.unpack('>II', file.read(8))
log.debug("size: %d, ver: %d, headers: %d" % \
(object_size, file_version,num_headers))
for i in range(0,num_headers):
try:
oi = struct.unpack('>4sIH',file.read(10))
except (struct.error, IOError):
# Header data we expected wasn't there. File may be
# only partially complete.
break
if object_id == 'DATA' and oi[0] != 'INDX':
log.debug('INDX chunk expected after DATA but not found -- file corrupt')
break
(object_id,object_size,object_version) = oi
if object_id == 'DATA':
# Seek over the data chunk rather than reading it in.
file.seek(object_size - 10, 1)
else:
self._read_header(object_id, file.read(object_size-10))
log.debug("%s [%d]" % (object_id,object_size-10))
# Read all the following headers
def _read_header(self,object_id,s):
if object_id == 'PROP':
prop = struct.unpack('>9IHH', s)
log.debug(prop)
if object_id == 'MDPR':
mdpr = struct.unpack('>H7I', s[:30])
log.debug(mdpr)
self.length = mdpr[7]/1000.0
(stream_name_size,) = struct.unpack('>B', s[30:31])
stream_name = s[31:31+stream_name_size]
pos = 31+stream_name_size
(mime_type_size,) = struct.unpack('>B', s[pos:pos+1])
mime = s[pos+1:pos+1+mime_type_size]
pos += mime_type_size+1
(type_specific_len,) = struct.unpack('>I', s[pos:pos+4])
type_specific = s[pos+4:pos+4+type_specific_len]
pos += 4+type_specific_len
if mime[:5] == 'audio':
ai = core.AudioStream()
ai.id = mdpr[0]
ai.bitrate = mdpr[2]
self.audio.append(ai)
elif mime[:5] == 'video':
vi = core.VideoStream()
vi.id = mdpr[0]
vi.bitrate = mdpr[2]
self.video.append(vi)
else:
log.debug("Unknown: %s" % mime)
if object_id == 'CONT':
pos = 0
(title_len,) = struct.unpack('>H', s[pos:pos+2])
self.title = s[2:title_len+2]
pos += title_len+2
(author_len,) = struct.unpack('>H', s[pos:pos+2])
self.artist = s[pos+2:pos+author_len+2]
pos += author_len+2
(copyright_len,) = struct.unpack('>H', s[pos:pos+2])
self.copyright = s[pos+2:pos+copyright_len+2]
pos += copyright_len+2
(comment_len,) = struct.unpack('>H', s[pos:pos+2])
self.comment = s[pos+2:pos+comment_len+2]
Parser = RealVideo
libs/enzyme/riff.py
+568
View File
@@ -0,0 +1,568 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2003-2006 Thomas Schueppel <stain@acm.org>
# Copyright (C) 2003-2006 Dirk Meyer <dischi@freevo.org>
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['Parser']
import os
import struct
import string
import logging
import time
from exceptions import *
import core
# get logging object
log = logging.getLogger(__name__)
# List of tags
# http://kibus1.narod.ru/frames_eng.htm?sof/abcavi/infotags.htm
# http://www.divx-digest.com/software/avitags_dll.html
# File Format: google for odmlff2.pdf
AVIINFO = {
'INAM': 'title',
'IART': 'artist',
'IPRD': 'product',
'ISFT': 'software',
'ICMT': 'comment',
'ILNG': 'language',
'IKEY': 'keywords',
'IPRT': 'trackno',
'IFRM': 'trackof',
'IPRO': 'producer',
'IWRI': 'writer',
'IGNR': 'genre',
'ICOP': 'copyright'
}
# Taken from libavcodec/mpeg4data.h (pixel_aspect struct)
PIXEL_ASPECT = {
1: (1, 1),
2: (12, 11),
3: (10, 11),
4: (16, 11),
5: (40, 33)
}
class Riff(core.AVContainer):
"""
AVI parser also parsing metadata like title, languages, etc.
"""
table_mapping = { 'AVIINFO' : AVIINFO }
def __init__(self,file):
core.AVContainer.__init__(self)
# read the header
h = file.read(12)
if h[:4] != "RIFF" and h[:4] != 'SDSS':
raise ParseError()
self.has_idx = False
self.header = {}
self.junkStart = None
self.infoStart = None
self.type = h[8:12]
if self.type == 'AVI ':
self.mime = 'video/avi'
elif self.type == 'WAVE':
self.mime = 'audio/wav'
try:
while self._parseRIFFChunk(file):
pass
except IOError:
log.exception('error in file, stop parsing')
self._find_subtitles(file.name)
if not self.has_idx and isinstance(self, core.AVContainer):
log.debug('WARNING: avi has no index')
self._set('corrupt', True)
def _find_subtitles(self, filename):
"""
Search for subtitle files. Right now only VobSub is supported
"""
base = os.path.splitext(filename)[0]
if os.path.isfile(base+'.idx') and \
(os.path.isfile(base+'.sub') or os.path.isfile(base+'.rar')):
file = open(base+'.idx')
if file.readline().find('VobSub index file') > 0:
for line in file.readlines():
if line.find('id') == 0:
sub = core.Subtitle()
sub.language = line[4:6]
sub.trackno = base + '.idx' # Maybe not?
self.subtitles.append(sub)
file.close()
def _parseAVIH(self,t):
retval = {}
v = struct.unpack('<IIIIIIIIIIIIII',t[0:56])
( retval['dwMicroSecPerFrame'],
retval['dwMaxBytesPerSec'],
retval['dwPaddingGranularity'],
retval['dwFlags'],
retval['dwTotalFrames'],
retval['dwInitialFrames'],
retval['dwStreams'],
retval['dwSuggestedBufferSize'],
retval['dwWidth'],
retval['dwHeight'],
retval['dwScale'],
retval['dwRate'],
retval['dwStart'],
retval['dwLength'] ) = v
if retval['dwMicroSecPerFrame'] == 0:
log.warning("ERROR: Corrupt AVI")
raise ParseError()
return retval
def _parseSTRH(self,t):
retval = {}
retval['fccType'] = t[0:4]
log.debug("_parseSTRH(%s) : %d bytes" % (retval['fccType'], len(t)))
if retval['fccType'] != 'auds':
retval['fccHandler'] = t[4:8]
v = struct.unpack('<IHHIIIIIIIII',t[8:52])
( retval['dwFlags'],
retval['wPriority'],
retval['wLanguage'],
retval['dwInitialFrames'],
retval['dwScale'],
retval['dwRate'],
retval['dwStart'],
retval['dwLength'],
retval['dwSuggestedBufferSize'],
retval['dwQuality'],
retval['dwSampleSize'],
retval['rcFrame'] ) = v
else:
try:
v = struct.unpack('<IHHIIIIIIIII',t[8:52])
( retval['dwFlags'],
retval['wPriority'],
retval['wLanguage'],
retval['dwInitialFrames'],
retval['dwScale'],
retval['dwRate'],
retval['dwStart'],
retval['dwLength'],
retval['dwSuggestedBufferSize'],
retval['dwQuality'],
retval['dwSampleSize'],
retval['rcFrame'] ) = v
self.delay = float(retval['dwStart']) / \
(float(retval['dwRate']) / retval['dwScale'])
except (KeyError, IndexError, ValueError, ZeroDivisionError):
pass
return retval
def _parseSTRF(self,t,strh):
fccType = strh['fccType']
retval = {}
if fccType == 'auds':
v = struct.unpack('<HHHHHH',t[0:12])
( retval['wFormatTag'],
retval['nChannels'],
retval['nSamplesPerSec'],
retval['nAvgBytesPerSec'],
retval['nBlockAlign'],
retval['nBitsPerSample'],
) = v
ai = core.AudioStream()
ai.samplerate = retval['nSamplesPerSec']
ai.channels = retval['nChannels']
# FIXME: Bitrate calculation is completely wrong.
#ai.samplebits = retval['nBitsPerSample']
#ai.bitrate = retval['nAvgBytesPerSec'] * 8
# TODO: set code if possible
# http://www.stats.uwa.edu.au/Internal/Specs/DXALL/FileSpec/\
# Languages
# ai.language = strh['wLanguage']
ai.codec = retval['wFormatTag']
self.audio.append(ai)
elif fccType == 'vids':
v = struct.unpack('<IIIHH',t[0:16])
( retval['biSize'],
retval['biWidth'],
retval['biHeight'],
retval['biPlanes'],
retval['biBitCount'] ) = v
v = struct.unpack('IIIII',t[20:40])
( retval['biSizeImage'],
retval['biXPelsPerMeter'],
retval['biYPelsPerMeter'],
retval['biClrUsed'],
retval['biClrImportant'] ) = v
vi = core.VideoStream()
vi.codec = t[16:20]
vi.width = retval['biWidth']
vi.height = retval['biHeight']
# FIXME: Bitrate calculation is completely wrong.
#vi.bitrate = strh['dwRate']
vi.fps = float(strh['dwRate']) / strh['dwScale']
vi.length = strh['dwLength'] / vi.fps
self.video.append(vi)
return retval
def _parseSTRL(self,t):
retval = {}
size = len(t)
i = 0
while i < len(t) - 8:
key = t[i:i+4]
sz = struct.unpack('<I',t[i+4:i+8])[0]
i+=8
value = t[i:]
if key == 'strh':
retval[key] = self._parseSTRH(value)
elif key == 'strf':
retval[key] = self._parseSTRF(value, retval['strh'])
else:
log.debug("_parseSTRL: unsupported stream tag '%s'", key)
i += sz
return retval, i
def _parseODML(self,t):
retval = {}
size = len(t)
i = 0
key = t[i:i+4]
sz = struct.unpack('<I',t[i+4:i+8])[0]
i += 8
value = t[i:]
if key != 'dmlh':
log.debug("_parseODML: Error")
i += sz - 8
return (retval, i)
def _parseVPRP(self,t):
retval = {}
v = struct.unpack('<IIIIIIIIII',t[:4*10])
( retval['VideoFormat'],
retval['VideoStandard'],
retval['RefreshRate'],
retval['HTotalIn'],
retval['VTotalIn'],
retval['FrameAspectRatio'],
retval['wPixel'],
retval['hPixel'] ) = v[1:-1]
# I need an avi with more informations
# enum {FORMAT_UNKNOWN, FORMAT_PAL_SQUARE, FORMAT_PAL_CCIR_601,
# FORMAT_NTSC_SQUARE, FORMAT_NTSC_CCIR_601,...} VIDEO_FORMAT;
# enum {STANDARD_UNKNOWN, STANDARD_PAL, STANDARD_NTSC, STANDARD_SECAM}
# VIDEO_STANDARD;
#
r = retval['FrameAspectRatio']
r = float(r >> 16) / (r & 0xFFFF)
retval['FrameAspectRatio'] = r
if self.video:
map(lambda v: setattr(v, 'aspect', r), self.video)
return (retval, v[0])
def _parseLISTmovi(self, size, file):
"""
Digs into movi list, looking for a Video Object Layer header in an
mpeg4 stream in order to determine aspect ratio.
"""
i = 0
n_dc = 0
done = False
# If the VOL header doesn't appear within 5MB or 5 video chunks,
# give up. The 5MB limit is not likely to apply except in
# pathological cases.
while i < min(1024*1024*5, size - 8) and n_dc < 5:
data = file.read(8)
if ord(data[0]) == 0:
# Eat leading nulls.
data = data[1:] + file.read(1)
i += 1
key, sz = struct.unpack('<4sI', data)
if key[2:] != 'dc' or sz > 1024*500:
# This chunk is not video or is unusually big (> 500KB);
# skip it.
file.seek(sz, 1)
i += 8 + sz
continue
n_dc += 1
# Read video chunk into memory
data = file.read(sz)
#for p in range(0,min(80, sz)):
# print "%02x " % ord(data[p]),
#print "\n\n"
# Look through the picture header for VOL startcode. The basic
# logic for this is taken from libavcodec, h263.c
pos = 0
startcode = 0xff
def bits(v, o, n):
# Returns n bits in v, offset o bits.
return (v & 2**n-1 << (64-n-o)) >> 64-n-o
while pos < sz:
startcode = ((startcode << 8) | ord(data[pos])) & 0xffffffff
pos += 1
if startcode & 0xFFFFFF00 != 0x100:
# No startcode found yet
continue
if startcode >= 0x120 and startcode <= 0x12F:
# We have the VOL startcode. Pull 64 bits of it and treat
# as a bitstream
v = struct.unpack(">Q", data[pos : pos+8])[0]
offset = 10
if bits(v, 9, 1):
# is_ol_id, skip over vo_ver_id and vo_priority
offset += 7
ar_info = bits(v, offset, 4)
if ar_info == 15:
# Extended aspect
num = bits(v, offset + 4, 8)
den = bits(v, offset + 12, 8)
else:
# A standard pixel aspect
num, den = PIXEL_ASPECT.get(ar_info, (0, 0))
# num/den indicates pixel aspect; convert to video aspect,
# so we need frame width and height.
if 0 not in [num, den]:
width, height = self.video[-1].width, self.video[-1].height
self.video[-1].aspect = num / float(den) * width / height
done = True
break
startcode = 0xff
i += 8 + len(data)
if done:
# We have the aspect, no need to continue parsing the movi
# list, so break out of the loop.
break
if i < size:
# Seek past whatever might be remaining of the movi list.
file.seek(size-i,1)
def _parseLIST(self,t):
retval = {}
i = 0
size = len(t)
while i < size-8:
# skip zero
if ord(t[i]) == 0: i += 1
key = t[i:i+4]
sz = 0
if key == 'LIST':
sz = struct.unpack('<I',t[i+4:i+8])[0]
i+=8
key = "LIST:"+t[i:i+4]
value = self._parseLIST(t[i:i+sz])
if key == 'strl':
for k in value.keys():
retval[k] = value[k]
else:
retval[key] = value
i+=sz
elif key == 'avih':
sz = struct.unpack('<I',t[i+4:i+8])[0]
i += 8
value = self._parseAVIH(t[i:i+sz])
i += sz
retval[key] = value
elif key == 'strl':
i += 4
(value, sz) = self._parseSTRL(t[i:])
key = value['strh']['fccType']
i += sz
retval[key] = value
elif key == 'odml':
i += 4
(value, sz) = self._parseODML(t[i:])
i += sz
elif key == 'vprp':
i += 4
(value, sz) = self._parseVPRP(t[i:])
retval[key] = value
i += sz
elif key == 'JUNK':
sz = struct.unpack('<I',t[i+4:i+8])[0]
i += sz + 8
else:
sz = struct.unpack('<I',t[i+4:i+8])[0]
i+=8
# in most cases this is some info stuff
if not key in AVIINFO.keys() and key != 'IDIT':
log.debug("Unknown Key: %s, len: %d" % (key,sz))
value = t[i:i+sz]
if key == 'ISFT':
# product information
if value.find('\0') > 0:
# works for Casio S500 camera videos
value = value[:value.find('\0')]
value = value.replace('\0', '').lstrip().rstrip()
value = value.replace('\0', '').lstrip().rstrip()
if value:
retval[key] = value
if key in ['IDIT', 'ICRD']:
# Timestamp the video was created. Spec says it
# should be a format like "Wed Jan 02 02:03:55 1990"
# Casio S500 uses "2005/12/24/ 14:11", but I've
# also seen "December 24, 2005"
specs = ('%a %b %d %H:%M:%S %Y', '%Y/%m/%d/ %H:%M', '%B %d, %Y')
for tmspec in specs:
try:
tm = time.strptime(value, tmspec)
# save timestamp as int
self.timestamp = int(time.mktime(tm))
break
except ValueError:
pass
else:
log.debug('no support for time format %s', value)
i+=sz
return retval
def _parseRIFFChunk(self,file):
h = file.read(8)
if len(h) < 8:
return False
name = h[:4]
size = struct.unpack('<I',h[4:8])[0]
if name == 'LIST':
pos = file.tell() - 8
key = file.read(4)
if key == 'movi' and self.video and not self.video[-1].aspect and \
self.video[-1].width and self.video[-1].height and \
self.video[-1].format in ['DIVX', 'XVID', 'FMP4']: # any others?
# If we don't have the aspect (i.e. it isn't in odml vprp
# header), but we do know the video's dimensions, and
# we're dealing with an mpeg4 stream, try to get the aspect
# from the VOL header in the mpeg4 stream.
self._parseLISTmovi(size-4, file)
return True
elif size > 80000:
log.debug('RIFF LIST "%s" too long to parse: %s bytes' % (key, size))
t = file.seek(size-4,1)
return True
elif size < 5:
log.debug('RIFF LIST "%s" too short: %s bytes' % (key, size))
return True
t = file.read(size-4)
log.debug('parse RIFF LIST "%s": %d bytes' % (key, size))
value = self._parseLIST(t)
self.header[key] = value
if key == 'INFO':
self.infoStart = pos
self._appendtable('AVIINFO', value)
elif key == 'MID ':
self._appendtable('AVIMID', value)
elif key == 'hdrl':
# no need to add this info to a table
pass
else:
log.debug('Skipping table info %s' % key)
elif name == 'JUNK':
self.junkStart = file.tell() - 8
self.junkSize = size
file.seek(size, 1)
elif name == 'idx1':
self.has_idx = True
log.debug('idx1: %s bytes' % size)
# no need to parse this
t = file.seek(size,1)
elif name == 'RIFF':
log.debug("New RIFF chunk, extended avi [%i]" % size)
type = file.read(4)
if type != 'AVIX':
log.debug("Second RIFF chunk is %s, not AVIX, skipping", type)
file.seek(size-4, 1)
# that's it, no new informations should be in AVIX
return False
elif name == 'fmt ' and size <= 50:
# This is a wav file.
data = file.read(size)
fmt = struct.unpack("<HHLLHH", data[:16])
self._set('codec', hex(fmt[0]))
self._set('samplerate', fmt[2])
# fmt[3] is average bytes per second, so we must divide it
# by 125 to get kbits per second
self._set('bitrate', fmt[3] / 125)
# ugly hack: remember original rate in bytes per second
# so that the length can be calculated in next elif block
self._set('byterate', fmt[3])
# Set a dummy fourcc so codec will be resolved in finalize.
self._set('fourcc', 'dummy')
elif name == 'data':
# XXX: this is naive and may not be right. For example if the
# stream is something that supports VBR like mp3, the value
# will be off. The only way to properly deal with this issue
# is to decode part of the stream based on its codec, but
# kaa.metadata doesn't have this capability (yet?)
# ugly hack: use original rate in bytes per second
self._set('length', size / float(self.byterate))
file.seek(size, 1)
elif not name.strip(string.printable + string.whitespace):
# check if name is something usefull at all, maybe it is no
# avi or broken
t = file.seek(size,1)
log.debug("Skipping %s [%i]" % (name,size))
else:
# bad avi
log.debug("Bad or broken avi")
return False
return True
Parser = Riff
libs/enzyme/strutils.py
+80
View File
@@ -0,0 +1,80 @@
# -*- coding: utf-8 -*-
# enzyme - Video metadata parser
# Copyright (C) 2011 Antoine Bertin <diaoulael@gmail.com>
# Copyright (C) 2006-2009 Dirk Meyer <dischi@freevo.org>
# Copyright (C) 2006-2009 Jason Tackaberry
#
# This file is part of enzyme.
#
# enzyme is free software; you can redistribute it and/or modify it under
# the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# enzyme is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
__all__ = ['ENCODING', 'str_to_unicode', 'unicode_to_str']
import locale
# find the correct encoding
try:
ENCODING = locale.getdefaultlocale()[1]
''.encode(ENCODING)
except (UnicodeError, TypeError):
ENCODING = 'latin-1'
def str_to_unicode(s, encoding=None):
"""
Attempts to convert a string of unknown character set to a unicode
string. First it tries to decode the string based on the locale's
preferred encoding, and if that fails, fall back to UTF-8 and then
latin-1. If all fails, it will force encoding to the preferred
charset, replacing unknown characters. If the given object is no
string, this function will return the given object.
"""
if not type(s) == str:
return s
if not encoding:
encoding = ENCODING
for c in [encoding, "utf-8", "latin-1"]:
try:
return s.decode(c)
except UnicodeDecodeError:
pass
return s.decode(encoding, "replace")
def unicode_to_str(s, encoding=None):
"""
Attempts to convert a unicode string of unknown character set to a
string. First it tries to encode the string based on the locale's
preferred encoding, and if that fails, fall back to UTF-8 and then
latin-1. If all fails, it will force encoding to the preferred
charset, replacing unknown characters. If the given object is no
unicode string, this function will return the given object.
"""
if not type(s) == unicode:
return s
if not encoding:
encoding = ENCODING
for c in [encoding, "utf-8", "latin-1"]:
try:
return s.encode(c)
except UnicodeDecodeError:
pass
return s.encode(encoding, "replace")