yt_dlp/extractor/twitcasting.py

   1 import itertools
   2 import re
   3
   4 from .common import InfoExtractor
   5 from ..dependencies import websockets
   6 from ..utils import (
   7     clean_html,
   8     ExtractorError,
   9     float_or_none,
  10     get_element_by_class,
  11     get_element_by_id,
  12     parse_duration,
  13     qualities,
  14     str_to_int,
  15     traverse_obj,
  16     try_get,
  17     unified_timestamp,
  18     urlencode_postdata,
  19     urljoin,
  20 )
  21
  22
  23 class TwitCastingIE(InfoExtractor):
  24     _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<uploader_id>[^/]+)/(?:movie|twplayer)/(?P<id>\d+)'
  25     _M3U8_HEADERS = {
  26         'Origin': 'https://twitcasting.tv',
  27         'Referer': 'https://twitcasting.tv/',
  28     }
  29     _TESTS = [{
  30         'url': 'https://twitcasting.tv/ivetesangalo/movie/2357609',
  31         'md5': '745243cad58c4681dc752490f7540d7f',
  32         'info_dict': {
  33             'id': '2357609',
  34             'ext': 'mp4',
  35             'title': 'Live #2357609',
  36             'uploader_id': 'ivetesangalo',
  37             'description': 'Twitter Oficial da cantora brasileira Ivete Sangalo.',
  38             'thumbnail': r're:^https?://.*\.jpg$',
  39             'upload_date': '20110822',
  40             'timestamp': 1314010824,
  41             'duration': 32,
  42             'view_count': int,
  43         },
  44         'params': {
  45             'skip_download': True,
  46         },
  47     }, {
  48         'url': 'https://twitcasting.tv/mttbernardini/movie/3689740',
  49         'info_dict': {
  50             'id': '3689740',
  51             'ext': 'mp4',
  52             'title': 'Live playing something #3689740',
  53             'uploader_id': 'mttbernardini',
  54             'description': 'Salve, io sono Matto (ma con la e). Questa è la mia presentazione, in quanto sono letteralmente matto (nel senso di strano), con qualcosa in più.',
  55             'thumbnail': r're:^https?://.*\.jpg$',
  56             'upload_date': '20120212',
  57             'timestamp': 1329028024,
  58             'duration': 681,
  59             'view_count': int,
  60         },
  61         'params': {
  62             'skip_download': True,
  63             'videopassword': 'abc',
  64         },
  65     }, {
  66         'note': 'archive is split in 2 parts',
  67         'url': 'https://twitcasting.tv/loft_heaven/movie/685979292',
  68         'info_dict': {
  69             'id': '685979292',
  70             'ext': 'mp4',
  71             'title': '南波一海のhear_here “ナタリー望月哲さんに聞く編集と「渋谷系狂騒曲」”',
  72             'duration': 6964.599334,
  73         },
  74         'playlist_mincount': 2,
  75     }]
  76
  77     def _real_extract(self, url):
  78         uploader_id, video_id = self._match_valid_url(url).groups()
  79
  80         video_password = self.get_param('videopassword')
  81         request_data = None
  82         if video_password:
  83             request_data = urlencode_postdata({
  84                 'password': video_password,
  85             }, encoding='utf-8')
  86         webpage, urlh = self._download_webpage_handle(
  87             url, video_id, data=request_data,
  88             headers={'Origin': 'https://twitcasting.tv'})
  89         if urlh.geturl() != url and request_data:
  90             webpage = self._download_webpage(
  91                 urlh.geturl(), video_id, data=request_data,
  92                 headers={'Origin': 'https://twitcasting.tv'},
  93                 note='Retrying authentication')
  94         # has to check here as the first request can contain password input form even if the password is correct
  95         if re.search(r'<form\s+method="POST">\s*<input\s+[^>]+?name="password"', webpage):
  96             raise ExtractorError('This video is protected by a password, use the --video-password option', expected=True)
  97
  98         title = (clean_html(get_element_by_id('movietitle', webpage))
  99                  or self._html_search_meta(['og:title', 'twitter:title'], webpage, fatal=True))
 100
 101         video_js_data = try_get(
 102             webpage,
 103             lambda x: self._parse_json(self._search_regex(
 104                 r'data-movie-playlist=\'([^\']+?)\'',
 105                 x, 'movie playlist', default=None), video_id)['2'], list)
 106
 107         thumbnail = traverse_obj(video_js_data, (0, 'thumbnailUrl')) or self._og_search_thumbnail(webpage)
 108         description = clean_html(get_element_by_id(
 109             'authorcomment', webpage)) or self._html_search_meta(
 110             ['description', 'og:description', 'twitter:description'], webpage)
 111         duration = (try_get(video_js_data, lambda x: sum(float_or_none(y.get('duration')) for y in x) / 1000)
 112                     or parse_duration(clean_html(get_element_by_class('tw-player-duration-time', webpage))))
 113         view_count = str_to_int(self._search_regex(
 114             (r'Total\s*:\s*([\d,]+)\s*Views', r'総視聴者\s*:\s*([\d,]+)\s*</'), webpage, 'views', None))
 115         timestamp = unified_timestamp(self._search_regex(
 116             r'data-toggle="true"[^>]+datetime="([^"]+)"',
 117             webpage, 'datetime', None))
 118
 119         stream_server_data = self._download_json(
 120             'https://twitcasting.tv/streamserver.php?target=%s&mode=client' % uploader_id, video_id,
 121             'Downloading live info', fatal=False)
 122
 123         is_live = 'data-status="online"' in webpage
 124         if not traverse_obj(stream_server_data, 'llfmp4') and is_live:
 125             self.raise_login_required(method='cookies')
 126
 127         base_dict = {
 128             'title': title,
 129             'description': description,
 130             'thumbnail': thumbnail,
 131             'timestamp': timestamp,
 132             'uploader_id': uploader_id,
 133             'duration': duration,
 134             'view_count': view_count,
 135             'is_live': is_live,
 136         }
 137
 138         def find_dmu(x):
 139             data_movie_url = self._search_regex(
 140                 r'data-movie-url=(["\'])(?P<url>(?:(?!\1).)+)\1',
 141                 x, 'm3u8 url', group='url', default=None)
 142             if data_movie_url:
 143                 return [data_movie_url]
 144
 145         m3u8_urls = (try_get(webpage, find_dmu, list)
 146                      or traverse_obj(video_js_data, (..., 'source', 'url'))
 147                      or ([f'https://twitcasting.tv/{uploader_id}/metastream.m3u8'] if is_live else None))
 148         if not m3u8_urls:
 149             raise ExtractorError('Failed to get m3u8 playlist')
 150
 151         if is_live:
 152             m3u8_url = m3u8_urls[0]
 153             formats = self._extract_m3u8_formats(
 154                 m3u8_url, video_id, ext='mp4', m3u8_id='hls',
 155                 live=True, headers=self._M3U8_HEADERS)
 156
 157             if traverse_obj(stream_server_data, ('hls', 'source')):
 158                 formats.extend(self._extract_m3u8_formats(
 159                     m3u8_url, video_id, ext='mp4', m3u8_id='source',
 160                     live=True, query={'mode': 'source'},
 161                     note='Downloading source quality m3u8',
 162                     headers=self._M3U8_HEADERS, fatal=False))
 163
 164             if websockets:
 165                 qq = qualities(['base', 'mobilesource', 'main'])
 166                 streams = traverse_obj(stream_server_data, ('llfmp4', 'streams')) or {}
 167                 for mode, ws_url in streams.items():
 168                     formats.append({
 169                         'url': ws_url,
 170                         'format_id': 'ws-%s' % mode,
 171                         'ext': 'mp4',
 172                         'quality': qq(mode),
 173                         'source_preference': -10,
 174                         # TwitCasting simply sends moof atom directly over WS
 175                         'protocol': 'websocket_frag',
 176                     })
 177
 178             self._sort_formats(formats, ('source',))
 179
 180             infodict = {
 181                 'formats': formats
 182             }
 183         elif len(m3u8_urls) == 1:
 184             formats = self._extract_m3u8_formats(
 185                 m3u8_urls[0], video_id, 'mp4', headers=self._M3U8_HEADERS)
 186             self._sort_formats(formats)
 187             infodict = {
 188                 # No problem here since there's only one manifest
 189                 'formats': formats,
 190                 'http_headers': self._M3U8_HEADERS,
 191             }
 192         else:
 193             infodict = {
 194                 '_type': 'multi_video',
 195                 'entries': [{
 196                     'id': f'{video_id}-{num}',
 197                     'url': m3u8_url,
 198                     'ext': 'mp4',
 199                     # Requesting the manifests here will cause download to fail.
 200                     # So use ffmpeg instead. See: https://github.com/yt-dlp/yt-dlp/issues/382
 201                     'protocol': 'm3u8',
 202                     'http_headers': self._M3U8_HEADERS,
 203                     **base_dict,
 204                 } for (num, m3u8_url) in enumerate(m3u8_urls)],
 205             }
 206
 207         return {
 208             'id': video_id,
 209             **base_dict,
 210             **infodict,
 211         }
 212
 213
 214 class TwitCastingLiveIE(InfoExtractor):
 215     _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/?(?:[#?]|$)'
 216     _TESTS = [{
 217         'url': 'https://twitcasting.tv/ivetesangalo',
 218         'only_matching': True,
 219     }]
 220
 221     def _real_extract(self, url):
 222         uploader_id = self._match_id(url)
 223         self.to_screen(
 224             'Downloading live video of user {0}. '
 225             'Pass "https://twitcasting.tv/{0}/show" to download the history'.format(uploader_id))
 226
 227         webpage = self._download_webpage(url, uploader_id)
 228         current_live = self._search_regex(
 229             (r'data-type="movie" data-id="(\d+)">',
 230              r'tw-sound-flag-open-link" data-id="(\d+)" style=',),
 231             webpage, 'current live ID', default=None)
 232         if not current_live:
 233             # fetch unfiltered /show to find running livestreams; we can't get ID of the password-protected livestream above
 234             webpage = self._download_webpage(
 235                 f'https://twitcasting.tv/{uploader_id}/show/', uploader_id,
 236                 note='Downloading live history')
 237             is_live = self._search_regex(r'(?s)(<span\s*class="tw-movie-thumbnail-badge"\s*data-status="live">\s*LIVE)', webpage, 'is live?', default=None)
 238             if is_live:
 239                 # get the first live; running live is always at the first
 240                 current_live = self._search_regex(
 241                     r'(?s)<a\s+class="tw-movie-thumbnail"\s*href="/[^/]+/movie/(?P<video_id>\d+)"\s*>.+?</a>',
 242                     webpage, 'current live ID 2', default=None, group='video_id')
 243         if not current_live:
 244             raise ExtractorError('The user is not currently live')
 245         return self.url_result('https://twitcasting.tv/%s/movie/%s' % (uploader_id, current_live))
 246
 247
 248 class TwitCastingUserIE(InfoExtractor):
 249     _VALID_URL = r'https?://(?:[^/]+\.)?twitcasting\.tv/(?P<id>[^/]+)/show/?(?:[#?]|$)'
 250     _TESTS = [{
 251         'url': 'https://twitcasting.tv/noriyukicas/show',
 252         'only_matching': True,
 253     }]
 254
 255     def _entries(self, uploader_id):
 256         base_url = next_url = 'https://twitcasting.tv/%s/show' % uploader_id
 257         for page_num in itertools.count(1):
 258             webpage = self._download_webpage(
 259                 next_url, uploader_id, query={'filter': 'watchable'}, note='Downloading page %d' % page_num)
 260             matches = re.finditer(
 261                 r'''(?isx)<a\s+class="tw-movie-thumbnail"\s*href="(?P<url>/[^/]+/movie/\d+)"\s*>.+?</a>''',
 262                 webpage)
 263             for mobj in matches:
 264                 yield self.url_result(urljoin(base_url, mobj.group('url')))
 265
 266             next_url = self._search_regex(
 267                 r'<a href="(/%s/show/%d-\d+)[?"]' % (re.escape(uploader_id), page_num),
 268                 webpage, 'next url', default=None)
 269             next_url = urljoin(base_url, next_url)
 270             if not next_url:
 271                 return
 272
 273     def _real_extract(self, url):
 274         uploader_id = self._match_id(url)
 275         return self.playlist_result(
 276             self._entries(uploader_id), uploader_id, '%s - Live History' % uploader_id)