import os
import subprocess
import time
+import re
from .common import AudioConversionError, PostProcessor
subtitles_filename,
dfxp2srt,
ISO639Utils,
+ replace_extension,
)
EXT_TO_OUT_FORMATS = {
- "aac": "adts",
- "m4a": "ipod",
- "mka": "matroska",
- "mkv": "matroska",
- "mpg": "mpeg",
- "ogv": "ogg",
- "ts": "mpegts",
- "wma": "asf",
- "wmv": "asf",
+ 'aac': 'adts',
+ 'flac': 'flac',
+ 'm4a': 'ipod',
+ 'mka': 'matroska',
+ 'mkv': 'matroska',
+ 'mpg': 'mpeg',
+ 'ogv': 'ogg',
+ 'ts': 'mpegts',
+ 'wma': 'asf',
+ 'wmv': 'asf',
+}
+ACODECS = {
+ 'mp3': 'libmp3lame',
+ 'aac': 'aac',
+ 'flac': 'flac',
+ 'm4a': 'aac',
+ 'opus': 'libopus',
+ 'vorbis': 'libvorbis',
+ 'wav': None,
}
def _determine_executables(self):
programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
- prefer_ffmpeg = False
+ prefer_ffmpeg = True
+
+ def get_ffmpeg_version(path):
+ ver = get_exe_version(path, args=['-version'])
+ if ver:
+ regexs = [
+ r'(?:\d+:)?([0-9.]+)-[0-9]+ubuntu[0-9.]+$', # Ubuntu, see [1]
+ r'n([0-9.]+)$', # Arch Linux
+ # 1. http://www.ducea.com/2006/06/17/ubuntu-package-version-naming-explanation/
+ ]
+ for regex in regexs:
+ mobj = re.match(regex, ver)
+ if mobj:
+ ver = mobj.group(1)
+ return ver
self.basename = None
self.probe_basename = None
self._paths = None
self._versions = None
if self._downloader:
- prefer_ffmpeg = self._downloader.params.get('prefer_ffmpeg', False)
+ prefer_ffmpeg = self._downloader.params.get('prefer_ffmpeg', True)
location = self._downloader.params.get('ffmpeg_location')
if location is not None:
if not os.path.exists(location):
self._paths = dict(
(p, os.path.join(location, p)) for p in programs)
self._versions = dict(
- (p, get_exe_version(self._paths[p], args=['-version']))
- for p in programs)
+ (p, get_ffmpeg_version(self._paths[p])) for p in programs)
if self._versions is None:
self._versions = dict(
- (p, get_exe_version(p, args=['-version'])) for p in programs)
+ (p, get_ffmpeg_version(p)) for p in programs)
self._paths = dict((p, p) for p in programs)
- if prefer_ffmpeg:
- prefs = ('ffmpeg', 'avconv')
- else:
+ if prefer_ffmpeg is False:
prefs = ('avconv', 'ffmpeg')
+ else:
+ prefs = ('ffmpeg', 'avconv')
for p in prefs:
if self._versions[p]:
self.basename = p
break
- if prefer_ffmpeg:
- prefs = ('ffprobe', 'avprobe')
- else:
+ if prefer_ffmpeg is False:
prefs = ('avprobe', 'ffprobe')
+ else:
+ prefs = ('ffprobe', 'avprobe')
for p in prefs:
if self._versions[p]:
self.probe_basename = p
def probe_executable(self):
return self._paths[self.probe_basename]
+ def get_audio_codec(self, path):
+ if not self.probe_available:
+ raise PostProcessingError('ffprobe or avprobe not found. Please install one.')
+ try:
+ cmd = [
+ encodeFilename(self.probe_executable, True),
+ encodeArgument('-show_streams'),
+ encodeFilename(self._ffmpeg_filename_argument(path), True)]
+ if self._downloader.params.get('verbose', False):
+ self._downloader.to_screen('[debug] %s command line: %s' % (self.basename, shell_quote(cmd)))
+ handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=subprocess.PIPE, stdin=subprocess.PIPE)
+ output = handle.communicate()[0]
+ if handle.wait() != 0:
+ return None
+ except (IOError, OSError):
+ return None
+ audio_codec = None
+ for line in output.decode('ascii', 'ignore').split('\n'):
+ if line.startswith('codec_name='):
+ audio_codec = line.split('=')[1].strip()
+ elif line.strip() == 'codec_type=audio' and audio_codec is not None:
+ return audio_codec
+ return None
+
def run_ffmpeg_multiple_files(self, input_paths, out_path, opts):
self.check_version()
self._preferredquality = preferredquality
self._nopostoverwrites = nopostoverwrites
- def get_audio_codec(self, path):
-
- if not self.probe_available:
- raise PostProcessingError('ffprobe or avprobe not found. Please install one.')
- try:
- cmd = [
- encodeFilename(self.probe_executable, True),
- encodeArgument('-show_streams'),
- encodeFilename(self._ffmpeg_filename_argument(path), True)]
- if self._downloader.params.get('verbose', False):
- self._downloader.to_screen('[debug] %s command line: %s' % (self.basename, shell_quote(cmd)))
- handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=subprocess.PIPE, stdin=subprocess.PIPE)
- output = handle.communicate()[0]
- if handle.wait() != 0:
- return None
- except (IOError, OSError):
- return None
- audio_codec = None
- for line in output.decode('ascii', 'ignore').split('\n'):
- if line.startswith('codec_name='):
- audio_codec = line.split('=')[1].strip()
- elif line.strip() == 'codec_type=audio' and audio_codec is not None:
- return audio_codec
- return None
-
def run_ffmpeg(self, path, out_path, codec, more_opts):
if codec is None:
acodec_opts = []
acodec = 'copy'
extension = 'm4a'
more_opts = ['-bsf:a', 'aac_adtstoasc']
- elif filecodec in ['aac', 'mp3', 'vorbis', 'opus']:
+ elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']:
# Lossless if possible
acodec = 'copy'
extension = filecodec
else:
more_opts += ['-b:a', self._preferredquality + 'k']
else:
- # We convert the audio (lossy)
- acodec = {'mp3': 'libmp3lame', 'aac': 'aac', 'm4a': 'aac', 'opus': 'opus', 'vorbis': 'libvorbis', 'wav': None}[self._preferredcodec]
+ # We convert the audio (lossy if codec is lossy)
+ acodec = ACODECS[self._preferredcodec]
extension = self._preferredcodec
more_opts = []
if self._preferredquality is not None:
prefix, sep, ext = path.rpartition('.') # not os.path.splitext, since the latter does not work on unicode in all setups
new_path = prefix + sep + extension
+ information['filepath'] = new_path
+ information['ext'] = extension
+
# If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly.
if (new_path == path or
(self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))):
new_path, time.time(), information['filetime'],
errnote='Cannot update utime of audio file')
- information['filepath'] = new_path
- information['ext'] = extension
-
return [path], information
opts += ['-c:s', 'mov_text']
for (i, lang) in enumerate(sub_langs):
opts.extend(['-map', '%d:0' % (i + 1)])
- lang_code = ISO639Utils.short2long(lang)
- if lang_code is not None:
- opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
+ lang_code = ISO639Utils.short2long(lang) or lang
+ opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
temp_filename = prepend_extension(filename, 'temp')
self._downloader.to_screen('[ffmpeg] Embedding subtitles in \'%s\'' % filename)
filename = info['filepath']
temp_filename = prepend_extension(filename, 'temp')
+ in_filenames = [filename]
+ options = []
if info['ext'] == 'm4a':
- options = ['-vn', '-acodec', 'copy']
+ options.extend(['-vn', '-acodec', 'copy'])
else:
- options = ['-c', 'copy']
+ options.extend(['-c', 'copy'])
for (name, value) in metadata.items():
options.extend(['-metadata', '%s=%s' % (name, value)])
+ chapters = info.get('chapters', [])
+ if chapters:
+ metadata_filename = replace_extension(filename, 'meta')
+ with io.open(metadata_filename, 'wt', encoding='utf-8') as f:
+ def ffmpeg_escape(text):
+ return re.sub(r'(=|;|#|\\|\n)', r'\\\1', text)
+
+ metadata_file_content = ';FFMETADATA1\n'
+ for chapter in chapters:
+ metadata_file_content += '[CHAPTER]\nTIMEBASE=1/1000\n'
+ metadata_file_content += 'START=%d\n' % (chapter['start_time'] * 1000)
+ metadata_file_content += 'END=%d\n' % (chapter['end_time'] * 1000)
+ chapter_title = chapter.get('title')
+ if chapter_title:
+ metadata_file_content += 'title=%s\n' % ffmpeg_escape(chapter_title)
+ f.write(metadata_file_content)
+ in_filenames.append(metadata_filename)
+ options.extend(['-map_metadata', '1'])
+
self._downloader.to_screen('[ffmpeg] Adding metadata to \'%s\'' % filename)
- self.run_ffmpeg(filename, temp_filename, options)
+ self.run_ffmpeg_multiple_files(in_filenames, temp_filename, options)
+ if chapters:
+ os.remove(metadata_filename)
os.remove(encodeFilename(filename))
os.rename(encodeFilename(temp_filename), encodeFilename(filename))
return [], info
class FFmpegFixupM3u8PP(FFmpegPostProcessor):
def run(self, info):
filename = info['filepath']
- temp_filename = prepend_extension(filename, 'temp')
+ if self.get_audio_codec(filename) == 'aac':
+ temp_filename = prepend_extension(filename, 'temp')
- options = ['-c', 'copy', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']
- self._downloader.to_screen('[ffmpeg] Fixing malformated aac bitstream in "%s"' % filename)
- self.run_ffmpeg(filename, temp_filename, options)
-
- os.remove(encodeFilename(filename))
- os.rename(encodeFilename(temp_filename), encodeFilename(filename))
+ options = ['-c', 'copy', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']
+ self._downloader.to_screen('[ffmpeg] Fixing malformed AAC bitstream in "%s"' % filename)
+ self.run_ffmpeg(filename, temp_filename, options)
+ os.remove(encodeFilename(filename))
+ os.rename(encodeFilename(temp_filename), encodeFilename(filename))
return [], info
ext = sub['ext']
if ext == new_ext:
self._downloader.to_screen(
- '[ffmpeg] Subtitle file for %s is already in the requested'
- 'format' % new_ext)
+ '[ffmpeg] Subtitle file for %s is already in the requested format' % new_ext)
continue
old_file = subtitles_filename(filename, lang, ext)
sub_filenames.append(old_file)
new_file = subtitles_filename(filename, lang, new_ext)
- if ext == 'dfxp' or ext == 'ttml' or ext == 'tt':
+ if ext in ('dfxp', 'ttml', 'tt'):
self._downloader.report_warning(
'You have requested to convert dfxp (TTML) subtitles into another format, '
'which results in style information loss')
dfxp_file = old_file
srt_file = subtitles_filename(filename, lang, 'srt')
- with io.open(dfxp_file, 'rt', encoding='utf-8') as f:
+ with open(dfxp_file, 'rb') as f:
srt_data = dfxp2srt(f.read())
with io.open(srt_file, 'wt', encoding='utf-8') as f:
Rapsys Git / youtubedl / blobdiff