def test_dfxp2srt(self):
dfxp_data = '''<?xml version="1.0" encoding="UTF-8"?>
<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="en" xmlns:tts="http://www.w3.org/ns/ttml#parameter">
<body>
<div xml:lang="en">
<p begin="0" end="1">The following line contains Chinese characters and special symbols</p>
<p begin="1" end="2">第二行<br/>♪♪</p>
<p begin="2" end="3"><span>Third<br/>Line</span></p>
</div>
</body>
</tt>'''
srt_data = '''1
00:00:00,000 --> 00:00:01,000
The following line contains Chinese characters and special symbols
2
00:00:01,000 --> 00:00:02,000
第二行
♪♪
3
00:00:02,000 --> 00:00:03,000
Third
Line
'''
self.assertEqual(dfxp2srt(dfxp_data), srt_data)
def test_dfxp2srt(self):
dfxp_data = '''<?xml version="1.0" encoding="UTF-8"?>
<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="en" xmlns:tts="http://www.w3.org/ns/ttml#parameter">
<body>
<div xml:lang="en">
<p begin="0" end="1">The following line contains Chinese characters and special symbols</p>
<p begin="1" end="2">第二行<br/>♪♪</p>
<p begin="2" dur="1"><span>Third<br/>Line</span></p>
<p begin="3" end="-1">Lines with invalid timestamps are ignored</p>
<p begin="-1" end="-1">Ignore, two</p>
<p begin="3" dur="-1">Ignored, three</p>
</div>
</body>
</tt>'''
srt_data = '''1
00:00:00,000 --> 00:00:01,000
The following line contains Chinese characters and special symbols
2
00:00:01,000 --> 00:00:02,000
第二行
♪♪
3
00:00:02,000 --> 00:00:03,000
Third
Line
'''
self.assertEqual(dfxp2srt(dfxp_data), srt_data)
dfxp_data_no_default_namespace = '''<?xml version="1.0" encoding="UTF-8"?>
<tt xml:lang="en" xmlns:tts="http://www.w3.org/ns/ttml#parameter">
<body>
<div xml:lang="en">
<p begin="0" end="1">The first line</p>
</div>
</body>
</tt>'''
srt_data = '''1
00:00:00,000 --> 00:00:01,000
The first line
'''
self.assertEqual(dfxp2srt(dfxp_data_no_default_namespace), srt_data)
def test_dfxp2srt(self):
dfxp_data = '''<?xml version="1.0" encoding="UTF-8"?>
<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="en" xmlns:tts="http://www.w3.org/ns/ttml#parameter">
<body>
<div xml:lang="en">
<p begin="0" end="1">The following line contains Chinese characters and special symbols</p>
<p begin="1" end="2">第二行<br/>♪♪</p>
<p begin="2" dur="1"><span>Third<br/>Line</span></p>
<p begin="3" end="-1">Lines with invalid timestamps are ignored</p>
<p begin="-1" end="-1">Ignore, two</p>
<p begin="3" dur="-1">Ignored, three</p>
</div>
</body>
</tt>'''
srt_data = '''1
00:00:00,000 --> 00:00:01,000
The following line contains Chinese characters and special symbols
2
00:00:01,000 --> 00:00:02,000
第二行
♪♪
3
00:00:02,000 --> 00:00:03,000
Third
Line
'''
self.assertEqual(dfxp2srt(dfxp_data), srt_data)
dfxp_data_no_default_namespace = '''<?xml version="1.0" encoding="UTF-8"?>
<tt xml:lang="en" xmlns:tts="http://www.w3.org/ns/ttml#parameter">
<body>
<div xml:lang="en">
<p begin="0" end="1">The first line</p>
</div>
</body>
</tt>'''
srt_data = '''1
00:00:00,000 --> 00:00:01,000
The first line
'''
self.assertEqual(dfxp2srt(dfxp_data_no_default_namespace), srt_data)
def run(self, info):
subs = info.get('requested_subtitles')
filename = info['filepath']
new_ext = self.format
new_format = new_ext
if new_format == 'vtt':
new_format = 'webvtt'
if subs is None:
self._downloader.to_screen('[ffmpeg] There aren\'t any subtitles to convert')
return [], info
self._downloader.to_screen('[ffmpeg] Converting subtitles')
sub_filenames = []
for lang, sub in subs.items():
ext = sub['ext']
if ext == new_ext:
self._downloader.to_screen(
'[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 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 open(dfxp_file, 'rb') as f:
srt_data = dfxp2srt(f.read())
with io.open(srt_file, 'wt', encoding='utf-8') as f:
f.write(srt_data)
old_file = srt_file
subs[lang] = {
'ext': 'srt',
'data': srt_data
}
if new_ext == 'srt':
continue
else:
sub_filenames.append(srt_file)
self.run_ffmpeg(old_file, new_file, ['-f', new_format])
with io.open(new_file, 'rt', encoding='utf-8') as f:
subs[lang] = {
'ext': new_ext,
'data': f.read(),
}
return sub_filenames, info
def __init__(self, handler):
self._handler = handler
# See ydl_opts['forcejson']
self._on_info_dict_json = None
self._allow_authentication_request = True
self._skip_authentication = False
self._skipped_count = 0
self.ydl_opts = {
'logger': self,
'logtostderr': True,
'no_color': True,
'progress_hooks': [self._on_progress],
'fixup': 'detect_or_warn',
'ignoreerrors': True, # handled via logger error callback
'retries': 10,
'fragment_retries': 10,
'writesubtitles': True,
'allsubtitles': True,
'subtitlesformat': 'vtt/best',
'keepvideo': True,
'postprocessors': [
{'key': 'FFmpegMetadata'},
{'key': 'FFmpegSubtitlesConvertor', 'format': 'vtt'},
{'key': 'FFmpegEmbedSubtitle'},
{'key': 'XAttrMetadata'}]}
url = self._handler.get_url()
download_dir = os.path.abspath(self._handler.get_download_dir())
with tempfile.TemporaryDirectory() as temp_dir:
self.ydl_opts['cookiefile'] = os.path.join(temp_dir, 'cookies')
# Collect info without downloading videos
testplaylist_dir = os.path.join(temp_dir, 'testplaylist')
noplaylist_dir = os.path.join(temp_dir, 'noplaylist')
fullplaylist_dir = os.path.join(temp_dir, 'fullplaylist')
for path in [testplaylist_dir, noplaylist_dir, fullplaylist_dir]:
os.mkdir(path)
self.ydl_opts['writeinfojson'] = True
self.ydl_opts['writethumbnail'] = True
self.ydl_opts['skip_download'] = True
self.ydl_opts['playlistend'] = 2
self.ydl_opts['outtmpl'] = '%(autonumber)s.%(ext)s'
# Test playlist
info_testplaylist, skipped_testplaylist = self._load_playlist(
testplaylist_dir, url)
self.ydl_opts['noplaylist'] = True
if len(info_testplaylist) + skipped_testplaylist > 1:
info_noplaylist, skipped_noplaylist = self._load_playlist(
noplaylist_dir, url)
else:
info_noplaylist = info_testplaylist
skipped_noplaylist = skipped_testplaylist
del self.ydl_opts['noplaylist']
del self.ydl_opts['playlistend']
if (len(info_testplaylist) + skipped_testplaylist >
len(info_noplaylist) + skipped_noplaylist):
self.ydl_opts['noplaylist'] = (
not self._handler.on_playlist_request())
if not self.ydl_opts['noplaylist']:
info_playlist, _ = self._load_playlist(
fullplaylist_dir, url)
else:
info_playlist = info_noplaylist
elif len(info_testplaylist) + skipped_testplaylist > 1:
info_playlist, _ = self._load_playlist(fullplaylist_dir, url)
else:
info_playlist = info_testplaylist
# Download videos
self._allow_authentication_request = False
del self.ydl_opts['writeinfojson']
del self.ydl_opts['writethumbnail']
del self.ydl_opts['skip_download']
# Include id and format_id in outtmpl to prevent youtube-dl
# from continuing wrong file
self.ydl_opts['outtmpl'] = '%(id)s.%(format_id)s.%(ext)s'
# Output info_dict as JSON handled via logger debug callback
self.ydl_opts['forcejson'] = True
mode = self._handler.get_mode()
if mode == 'audio':
resolution = MAX_RESOLUTION
prefer_mpeg = False
self.ydl_opts['format'] = 'bestaudio/best'
self.ydl_opts['postprocessors'].insert(0, {
'key': 'FFmpegExtractAudio',
'preferredcodec': 'mp3',
'preferredquality': '192'})
self.ydl_opts['postprocessors'].insert(1, {
'key': 'EmbedThumbnail',
'already_have_thumbnail': True})
else:
resolution = self._handler.get_resolution()
prefer_mpeg = self._handler.get_prefer_mpeg()
try:
os.makedirs(download_dir, exist_ok=True)
except OSError as e:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
self._handler.on_error(
'ERROR: Failed to create download folder: %s' % e)
sys.exit(1)
for i, (info_path, thumbnail_paths, subtitles) in enumerate(
info_playlist):
with open(info_path) as f:
info = json.load(f)
title = info.get('title') or info.get('id') or 'video'
output_title = self._get_output_title(title)
# Test subtitles
# youtube-dl fails for subtitles that it can't convert or
# are unsupported by ffmpeg
supported_subtitles = []
for sub_path, sub_lang, sub_ext in subtitles:
print('[youtube_dl_slave] Testing subtitle (%r, %r)' %
(sub_lang, sub_ext), file=sys.stderr, flush=True)
if sub_ext in ['dfxp', 'ttml', 'tt']:
# Try to use youtube-dl's internal dfxp2srt converter
with open(sub_path, 'rb') as f:
sub_data = f.read()
try:
sub_data = dfxp2srt(sub_data)
except Exception:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
continue
ff_sub_path = sub_path + '-converted.srt'
with open(ff_sub_path, 'w', encoding='utf-8') as f:
f.write(sub_data)
else:
ff_sub_path = sub_path
# Try to read and convert subtitles with ffmpeg
try:
subprocess.run(
[FFMPEG_EXE, '-i', os.path.abspath(ff_sub_path),
'-f', 'webvtt', '-'],
check=True, stdin=subprocess.DEVNULL,
stdout=subprocess.DEVNULL)
except FileNotFoundError:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
self._handler.on_error(
'ERROR: %r not found' % FFMPEG_EXE)
sys.exit(1)
except subprocess.CalledProcessError:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
continue
supported_subtitles.append((sub_lang, sub_ext))
# Choose supported subtitles
new_info_subtitles = {}
for sub_lang, subs in (info.get('subtitles') or {}).items():
new_subs = []
for sub in subs or []:
if (sub_lang, sub.get('ext')) in supported_subtitles:
new_subs.append(sub)
if new_subs:
new_info_subtitles[sub_lang] = new_subs
info['subtitles'] = new_info_subtitles
thumbnail_path = thumbnail_paths[0] if thumbnail_paths else ''
if thumbnail_path:
# Convert thumbnail to JPEG and limit resolution
print('[youtube_dl_slave] Converting thumbnail',
file=sys.stderr, flush=True)
new_thumbnail_path = thumbnail_path + '-converted.jpg'
try:
subprocess.run(
[FFMPEG_EXE, '-i', os.path.abspath(thumbnail_path),
'-vf', ('scale=\'min({0},iw):min({0},ih):'
'force_original_aspect_ratio=decrease\''
).format(MAX_THUMBNAIL_RESOLUTION),
os.path.abspath(new_thumbnail_path)],
check=True, stdin=subprocess.DEVNULL,
stdout=subprocess.DEVNULL)
except FileNotFoundError:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
self._handler.on_error(
'ERROR: %r not found' % FFMPEG_EXE)
sys.exit(1)
except subprocess.CalledProcessError:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
new_thumbnail_path = ''
# No longer needed
os.remove(thumbnail_path)
thumbnail_path = new_thumbnail_path
self._handler.on_progress_start(i, len(info_playlist), title,
thumbnail_path)
for thumbnail in info.get('thumbnails') or []:
thumbnail['filename'] = thumbnail_path
# Remove description, because long comments cause problems when
# displayed in Nautilus and other applications.
with contextlib.suppress(KeyError):
del info['description']
sort_formats(info.get('formats') or [], resolution,
prefer_mpeg)
with open(info_path, 'w') as f:
json.dump(info, f)
# Check if we already got the file
existing_filename = self._find_existing_download(
download_dir, output_title, mode)
if existing_filename is not None:
self._handler.on_progress_end(existing_filename)
continue
# Download into separate directory because youtube-dl generates
# many temporary files
temp_download_dir = os.path.join(
download_dir, output_title + '.part')
# Lock download directory to prevent other processes from
# writing to the same files
temp_download_dir_cm = contextlib.ExitStack()
try:
temp_download_dir_cm.enter_context(
self._create_and_lock_dir(temp_download_dir))
except OSError as e:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
self._handler.on_error(
'ERROR: Failed to lock download folder: %s' % e)
sys.exit(1)
with temp_download_dir_cm:
# Check if the file got downloaded in the meantime
existing_filename = self._find_existing_download(
download_dir, output_title, mode)
if existing_filename is not None:
filename = existing_filename
else:
info_dict = None
# See ydl_opts['forcejson']
def on_info_dict_json(info_dict_):
nonlocal info_dict
info_dict = info_dict_
self._on_info_dict_json = on_info_dict_json
self._load_video(temp_download_dir, info_path)
if self._on_info_dict_json:
raise RuntimeError('info_dict not received')
# Find the temporary filename
temp_filename_root, temp_filename_ext = (
os.path.splitext(info_dict['_filename']))
if mode == 'audio':
temp_filename_ext = '.mp3'
else:
# youtube-dl changes extension for incompatible
# formats to .mkv
for ext in [temp_filename_ext, '.mkv']:
if os.path.exists(temp_filename_root + ext):
temp_filename_ext = ext
break
temp_filename = temp_filename_root + temp_filename_ext
filename = output_title + temp_filename_ext
# Move finished download from download to target dir
try:
os.replace(
os.path.join(temp_download_dir, temp_filename),
os.path.join(download_dir, filename))
except OSError as e:
traceback.print_exc(file=sys.stderr)
sys.stderr.flush()
self._handler.on_error((
'ERROR: Falied to move finished download to '
'download folder: %s') % e)
sys.exit(1)
# Delete download directory
with contextlib.suppress(OSError):
shutil.rmtree(temp_download_dir)
self._handler.on_progress_end(filename)
