- for a in dash_doc.findall('.//' + _add_ns('AdaptationSet')):
- mime_type = a.attrib.get('mimeType')
- for r in a.findall(_add_ns('Representation')):
- mime_type = r.attrib.get('mimeType') or mime_type
- url_el = r.find(_add_ns('BaseURL'))
- if mime_type == 'text/vtt':
- # TODO implement WebVTT downloading
- pass
- elif mime_type.startswith('audio/') or mime_type.startswith('video/'):
- segment_list = r.find(_add_ns('SegmentList'))
- format_id = r.attrib['id']
- video_url = url_el.text if url_el is not None else None
- filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
- f = {
- 'format_id': format_id,
- 'url': video_url,
- 'width': int_or_none(r.attrib.get('width')),
- 'height': int_or_none(r.attrib.get('height')),
- 'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
- 'asr': int_or_none(r.attrib.get('audioSamplingRate')),
- 'filesize': filesize,
- 'fps': int_or_none(r.attrib.get('frameRate')),
- }
- if segment_list is not None:
- initialization_url = segment_list.find(_add_ns('Initialization')).attrib['sourceURL']
- f.update({
- 'initialization_url': initialization_url,
- 'segment_urls': [segment.attrib.get('media') for segment in segment_list.findall(_add_ns('SegmentURL'))],
- 'protocol': 'http_dash_segments',
- })
- if not f.get('url'):
- f['url'] = initialization_url
- try:
- existing_format = next(
- fo for fo in formats
- if fo['format_id'] == format_id)
- except StopIteration:
- full_info = formats_dict.get(format_id, {}).copy()
- full_info.update(f)
- codecs = r.attrib.get('codecs')
- if codecs:
- if mime_type.startswith('video/'):
- vcodec, acodec = codecs, 'none'
- else: # mime_type.startswith('audio/')
- vcodec, acodec = 'none', codecs
-
- full_info.update({
- 'vcodec': vcodec,
- 'acodec': acodec,
+ for period in mpd_doc.findall(_add_ns('Period')):
+ period_duration = parse_duration(period.get('duration')) or mpd_duration
+ period_ms_info = extract_multisegment_info(period, {
+ 'start_number': 1,
+ 'timescale': 1,
+ })
+ for adaptation_set in period.findall(_add_ns('AdaptationSet')):
+ if is_drm_protected(adaptation_set):
+ continue
+ adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info)
+ for representation in adaptation_set.findall(_add_ns('Representation')):
+ if is_drm_protected(representation):
+ continue
+ representation_attrib = adaptation_set.attrib.copy()
+ representation_attrib.update(representation.attrib)
+ mime_type = representation_attrib.get('mimeType')
+ content_type = mime_type.split('/')[0] if mime_type else representation_attrib.get('contentType')
+ if content_type == 'text':
+ # TODO implement WebVTT downloading
+ pass
+ elif content_type == 'video' or content_type == 'audio':
+ base_url = ''
+ for element in (representation, adaptation_set, period, mpd_doc):
+ base_url_e = element.find(_add_ns('BaseURL'))
+ if base_url_e is not None:
+ base_url = base_url_e.text + base_url
+ if re.match(r'^https?://', base_url):
+ break
+ if mpd_base_url and not re.match(r'^https?://', base_url):
+ if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
+ mpd_base_url += '/'
+ base_url = mpd_base_url + base_url
+ representation_id = representation_attrib.get('id')
+ lang = representation_attrib.get('lang')
+ url_el = representation.find(_add_ns('BaseURL'))
+ filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
+ f = {
+ 'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
+ 'url': base_url,
+ 'ext': codec2ext(representation_attrib.get('codecs')),
+ 'width': int_or_none(representation_attrib.get('width')),
+ 'height': int_or_none(representation_attrib.get('height')),
+ 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
+ 'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
+ 'fps': int_or_none(representation_attrib.get('frameRate')),
+ 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
+ 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
+ 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
+ 'format_note': 'DASH %s' % content_type,
+ 'filesize': filesize,
+ }
+ representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
+ if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
+ if 'total_number' not in representation_ms_info and 'segment_duration':
+ segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
+ representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
+ media_template = representation_ms_info['media_template']
+ media_template = media_template.replace('$RepresentationID$', representation_id)
+ media_template = re.sub(r'\$(Number|Bandwidth)(?:%(0\d+)d)?\$', r'%(\1)\2d', media_template)
+ media_template.replace('$$', '$')
+ representation_ms_info['segment_urls'] = [media_template % {'Number': segment_number, 'Bandwidth': representation_attrib.get('bandwidth')} for segment_number in range(representation_ms_info['start_number'], representation_ms_info['total_number'] + representation_ms_info['start_number'])]
+ if 'segment_urls' in representation_ms_info:
+ f.update({
+ 'segment_urls': representation_ms_info['segment_urls'],
+ 'protocol': 'http_dash_segments',