Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/twitter.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..compat import compat_urlparse
   8 from ..utils import (
   9     determine_ext,
  10     dict_get,
  11     ExtractorError,
  12     float_or_none,
  13     int_or_none,
  14     remove_end,
  15     try_get,
  16     xpath_text,
  17 )
  18
  19 from .periscope import PeriscopeIE
  20
  21
  22 class TwitterBaseIE(InfoExtractor):
  23     def _extract_formats_from_vmap_url(self, vmap_url, video_id):
  24         vmap_data = self._download_xml(vmap_url, video_id)
  25         video_url = xpath_text(vmap_data, './/MediaFile').strip()
  26         if determine_ext(video_url) == 'm3u8':
  27             return self._extract_m3u8_formats(
  28                 video_url, video_id, ext='mp4', m3u8_id='hls',
  29                 entry_protocol='m3u8_native')
  30         return [{
  31             'url': video_url,
  32         }]
  33
  34     @staticmethod
  35     def _search_dimensions_in_video_url(a_format, video_url):
  36         m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
  37         if m:
  38             a_format.update({
  39                 'width': int(m.group('width')),
  40                 'height': int(m.group('height')),
  41             })
  42
  43
  44 class TwitterCardIE(TwitterBaseIE):
  45     IE_NAME = 'twitter:card'
  46     _VALID_URL = r'https?://(?:www\.)?twitter\.com/i/(?P<path>cards/tfw/v1|videos(?:/tweet)?)/(?P<id>\d+)'
  47     _TESTS = [
  48         {
  49             'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
  50             # MD5 checksums are different in different places
  51             'info_dict': {
  52                 'id': '560070183650213889',
  53                 'ext': 'mp4',
  54                 'title': 'Twitter web player',
  55                 'thumbnail': r're:^https?://.*\.jpg$',
  56                 'duration': 30.033,
  57             },
  58         },
  59         {
  60             'url': 'https://twitter.com/i/cards/tfw/v1/623160978427936768',
  61             'md5': '7ee2a553b63d1bccba97fbed97d9e1c8',
  62             'info_dict': {
  63                 'id': '623160978427936768',
  64                 'ext': 'mp4',
  65                 'title': 'Twitter web player',
  66                 'thumbnail': r're:^https?://.*(?:\bformat=|\.)jpg',
  67             },
  68         },
  69         {
  70             'url': 'https://twitter.com/i/cards/tfw/v1/654001591733886977',
  71             'md5': 'b6d9683dd3f48e340ded81c0e917ad46',
  72             'info_dict': {
  73                 'id': 'dq4Oj5quskI',
  74                 'ext': 'mp4',
  75                 'title': 'Ubuntu 11.10 Overview',
  76                 'description': 'md5:a831e97fa384863d6e26ce48d1c43376',
  77                 'upload_date': '20111013',
  78                 'uploader': 'OMG! Ubuntu!',
  79                 'uploader_id': 'omgubuntu',
  80             },
  81             'add_ie': ['Youtube'],
  82         },
  83         {
  84             'url': 'https://twitter.com/i/cards/tfw/v1/665289828897005568',
  85             'md5': '6dabeaca9e68cbb71c99c322a4b42a11',
  86             'info_dict': {
  87                 'id': 'iBb2x00UVlv',
  88                 'ext': 'mp4',
  89                 'upload_date': '20151113',
  90                 'uploader_id': '1189339351084113920',
  91                 'uploader': 'ArsenalTerje',
  92                 'title': 'Vine by ArsenalTerje',
  93                 'timestamp': 1447451307,
  94             },
  95             'add_ie': ['Vine'],
  96         }, {
  97             'url': 'https://twitter.com/i/videos/tweet/705235433198714880',
  98             'md5': '884812a2adc8aaf6fe52b15ccbfa3b88',
  99             'info_dict': {
 100                 'id': '705235433198714880',
 101                 'ext': 'mp4',
 102                 'title': 'Twitter web player',
 103                 'thumbnail': r're:^https?://.*',
 104             },
 105         }, {
 106             'url': 'https://twitter.com/i/videos/752274308186120192',
 107             'only_matching': True,
 108         },
 109     ]
 110
 111     def _parse_media_info(self, media_info, video_id):
 112         formats = []
 113         for media_variant in media_info.get('variants', []):
 114             media_url = media_variant['url']
 115             if media_url.endswith('.m3u8'):
 116                 formats.extend(self._extract_m3u8_formats(media_url, video_id, ext='mp4', m3u8_id='hls'))
 117             elif media_url.endswith('.mpd'):
 118                 formats.extend(self._extract_mpd_formats(media_url, video_id, mpd_id='dash'))
 119             else:
 120                 tbr = int_or_none(dict_get(media_variant, ('bitRate', 'bitrate')), scale=1000)
 121                 a_format = {
 122                     'url': media_url,
 123                     'format_id': 'http-%d' % tbr if tbr else 'http',
 124                     'tbr': tbr,
 125                 }
 126                 # Reported bitRate may be zero
 127                 if not a_format['tbr']:
 128                     del a_format['tbr']
 129
 130                 self._search_dimensions_in_video_url(a_format, media_url)
 131
 132                 formats.append(a_format)
 133         return formats
 134
 135     def _extract_mobile_formats(self, username, video_id):
 136         webpage = self._download_webpage(
 137             'https://mobile.twitter.com/%s/status/%s' % (username, video_id),
 138             video_id, 'Downloading mobile webpage',
 139             headers={
 140                 # A recent mobile UA is necessary for `gt` cookie
 141                 'User-Agent': 'Mozilla/5.0 (Android 6.0.1; Mobile; rv:54.0) Gecko/54.0 Firefox/54.0',
 142             })
 143         main_script_url = self._html_search_regex(
 144             r'<script[^>]+src="([^"]+main\.[^"]+)"', webpage, 'main script URL')
 145         main_script = self._download_webpage(
 146             main_script_url, video_id, 'Downloading main script')
 147         bearer_token = self._search_regex(
 148             r'BEARER_TOKEN\s*:\s*"([^"]+)"',
 149             main_script, 'bearer token')
 150         # https://developer.twitter.com/en/docs/tweets/post-and-engage/api-reference/get-statuses-show-id
 151         api_data = self._download_json(
 152             'https://api.twitter.com/1.1/statuses/show/%s.json' % video_id,
 153             video_id, 'Downloading API data',
 154             headers={
 155                 'Authorization': 'Bearer ' + bearer_token,
 156             })
 157         media_info = try_get(api_data, lambda o: o['extended_entities']['media'][0]['video_info']) or {}
 158         return self._parse_media_info(media_info, video_id)
 159
 160     def _real_extract(self, url):
 161         path, video_id = re.search(self._VALID_URL, url).groups()
 162
 163         config = None
 164         formats = []
 165         duration = None
 166
 167         urls = [url]
 168         if path.startswith('cards/'):
 169             urls.append('https://twitter.com/i/videos/' + video_id)
 170
 171         for u in urls:
 172             webpage = self._download_webpage(u, video_id)
 173
 174             iframe_url = self._html_search_regex(
 175                 r'<iframe[^>]+src="((?:https?:)?//(?:www\.youtube\.com/embed/[^"]+|(?:www\.)?vine\.co/v/\w+/card))"',
 176                 webpage, 'video iframe', default=None)
 177             if iframe_url:
 178                 return self.url_result(iframe_url)
 179
 180             config = self._parse_json(self._html_search_regex(
 181                 r'data-(?:player-)?config="([^"]+)"', webpage,
 182                 'data player config', default='{}'),
 183                 video_id)
 184
 185             if config.get('source_type') == 'vine':
 186                 return self.url_result(config['player_url'], 'Vine')
 187
 188             periscope_url = PeriscopeIE._extract_url(webpage)
 189             if periscope_url:
 190                 return self.url_result(periscope_url, PeriscopeIE.ie_key())
 191
 192             video_url = config.get('video_url') or config.get('playlist', [{}])[0].get('source')
 193
 194             if video_url:
 195                 if determine_ext(video_url) == 'm3u8':
 196                     formats.extend(self._extract_m3u8_formats(video_url, video_id, ext='mp4', m3u8_id='hls'))
 197                 else:
 198                     f = {
 199                         'url': video_url,
 200                     }
 201
 202                     self._search_dimensions_in_video_url(f, video_url)
 203
 204                     formats.append(f)
 205
 206             vmap_url = config.get('vmapUrl') or config.get('vmap_url')
 207             if vmap_url:
 208                 formats.extend(
 209                     self._extract_formats_from_vmap_url(vmap_url, video_id))
 210
 211             media_info = None
 212
 213             for entity in config.get('status', {}).get('entities', []):
 214                 if 'mediaInfo' in entity:
 215                     media_info = entity['mediaInfo']
 216
 217             if media_info:
 218                 formats.extend(self._parse_media_info(media_info, video_id))
 219                 duration = float_or_none(media_info.get('duration', {}).get('nanos'), scale=1e9)
 220
 221             username = config.get('user', {}).get('screen_name')
 222             if username:
 223                 formats.extend(self._extract_mobile_formats(username, video_id))
 224
 225             if formats:
 226                 break
 227
 228         self._remove_duplicate_formats(formats)
 229         self._sort_formats(formats)
 230
 231         title = self._search_regex(r'<title>([^<]+)</title>', webpage, 'title')
 232         thumbnail = config.get('posterImageUrl') or config.get('image_src')
 233         duration = float_or_none(config.get('duration'), scale=1000) or duration
 234
 235         return {
 236             'id': video_id,
 237             'title': title,
 238             'thumbnail': thumbnail,
 239             'duration': duration,
 240             'formats': formats,
 241         }
 242
 243
 244 class TwitterIE(InfoExtractor):
 245     IE_NAME = 'twitter'
 246     _VALID_URL = r'https?://(?:www\.|m\.|mobile\.)?twitter\.com/(?:i/web|(?P<user_id>[^/]+))/status/(?P<id>\d+)'
 247     _TEMPLATE_URL = 'https://twitter.com/%s/status/%s'
 248     _TEMPLATE_STATUSES_URL = 'https://twitter.com/statuses/%s'
 249
 250     _TESTS = [{
 251         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
 252         'info_dict': {
 253             'id': '643211948184596480',
 254             'ext': 'mp4',
 255             'title': 'FREE THE NIPPLE - FTN supporters on Hollywood Blvd today!',
 256             'thumbnail': r're:^https?://.*\.jpg',
 257             'description': 'FREE THE NIPPLE on Twitter: "FTN supporters on Hollywood Blvd today! http://t.co/c7jHH749xJ"',
 258             'uploader': 'FREE THE NIPPLE',
 259             'uploader_id': 'freethenipple',
 260             'duration': 12.922,
 261         },
 262     }, {
 263         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
 264         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
 265         'info_dict': {
 266             'id': '657991469417025536',
 267             'ext': 'mp4',
 268             'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
 269             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
 270             'thumbnail': r're:^https?://.*\.png',
 271             'uploader': 'Gifs',
 272             'uploader_id': 'giphz',
 273         },
 274         'expected_warnings': ['height', 'width'],
 275         'skip': 'Account suspended',
 276     }, {
 277         'url': 'https://twitter.com/starwars/status/665052190608723968',
 278         'info_dict': {
 279             'id': '665052190608723968',
 280             'ext': 'mp4',
 281             'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
 282             'description': 'Star Wars on Twitter: "A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens."',
 283             'uploader_id': 'starwars',
 284             'uploader': 'Star Wars',
 285         },
 286     }, {
 287         'url': 'https://twitter.com/BTNBrentYarina/status/705235433198714880',
 288         'info_dict': {
 289             'id': '705235433198714880',
 290             'ext': 'mp4',
 291             'title': 'Brent Yarina - Khalil Iverson\'s missed highlight dunk. And made highlight dunk. In one highlight.',
 292             'description': 'Brent Yarina on Twitter: "Khalil Iverson\'s missed highlight dunk. And made highlight dunk. In one highlight."',
 293             'uploader_id': 'BTNBrentYarina',
 294             'uploader': 'Brent Yarina',
 295         },
 296         'params': {
 297             # The same video as https://twitter.com/i/videos/tweet/705235433198714880
 298             # Test case of TwitterCardIE
 299             'skip_download': True,
 300         },
 301     }, {
 302         'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
 303         'info_dict': {
 304             'id': '700207533655363584',
 305             'ext': 'mp4',
 306             'title': 'JG - BEAT PROD: @suhmeduh #Damndaniel',
 307             'description': 'JG on Twitter: "BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ"',
 308             'thumbnail': r're:^https?://.*\.jpg',
 309             'uploader': 'JG',
 310             'uploader_id': 'jaydingeer',
 311             'duration': 30.0,
 312         },
 313     }, {
 314         'url': 'https://twitter.com/Filmdrunk/status/713801302971588609',
 315         'md5': '89a15ed345d13b86e9a5a5e051fa308a',
 316         'info_dict': {
 317             'id': 'MIOxnrUteUd',
 318             'ext': 'mp4',
 319             'title': 'Vince Mancini - Vine of the day',
 320             'description': 'Vince Mancini on Twitter: "Vine of the day https://t.co/xmTvRdqxWf"',
 321             'uploader': 'Vince Mancini',
 322             'uploader_id': 'Filmdrunk',
 323             'timestamp': 1402826626,
 324             'upload_date': '20140615',
 325         },
 326         'add_ie': ['Vine'],
 327     }, {
 328         'url': 'https://twitter.com/captainamerica/status/719944021058060289',
 329         'info_dict': {
 330             'id': '719944021058060289',
 331             'ext': 'mp4',
 332             'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
 333             'description': 'Captain America on Twitter: "@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI"',
 334             'uploader_id': 'captainamerica',
 335             'uploader': 'Captain America',
 336             'duration': 3.17,
 337         },
 338     }, {
 339         'url': 'https://twitter.com/OPP_HSD/status/779210622571536384',
 340         'info_dict': {
 341             'id': '1zqKVVlkqLaKB',
 342             'ext': 'mp4',
 343             'title': 'Sgt Kerry Schmidt - LIVE on #Periscope: Road rage, mischief, assault, rollover and fire in one occurrence',
 344             'description': 'Sgt Kerry Schmidt on Twitter: "LIVE on #Periscope: Road rage, mischief, assault, rollover and fire in one occurrence  https://t.co/EKrVgIXF3s"',
 345             'upload_date': '20160923',
 346             'uploader_id': 'OPP_HSD',
 347             'uploader': 'Sgt Kerry Schmidt',
 348             'timestamp': 1474613214,
 349         },
 350         'add_ie': ['Periscope'],
 351     }, {
 352         # has mp4 formats via mobile API
 353         'url': 'https://twitter.com/news_al3alm/status/852138619213144067',
 354         'info_dict': {
 355             'id': '852138619213144067',
 356             'ext': 'mp4',
 357             'title': 'عالم الأخبار - كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة',
 358             'description': 'عالم الأخبار on Twitter: "كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة   https://t.co/xg6OhpyKfN"',
 359             'uploader': 'عالم الأخبار',
 360             'uploader_id': 'news_al3alm',
 361             'duration': 277.4,
 362         },
 363     }, {
 364         'url': 'https://twitter.com/i/web/status/910031516746514432',
 365         'info_dict': {
 366             'id': '910031516746514432',
 367             'ext': 'mp4',
 368             'title': 'Préfet de Guadeloupe - [Direct] #Maria Le centre se trouve actuellement au sud de Basse-Terre. Restez confinés. Réfugiez-vous dans la pièce la + sûre.',
 369             'thumbnail': r're:^https?://.*\.jpg',
 370             'description': 'Préfet de Guadeloupe on Twitter: "[Direct] #Maria Le centre se trouve actuellement au sud de Basse-Terre. Restez confinés. Réfugiez-vous dans la pièce la + sûre. https://t.co/mwx01Rs4lo"',
 371             'uploader': 'Préfet de Guadeloupe',
 372             'uploader_id': 'Prefet971',
 373             'duration': 47.48,
 374         },
 375         'params': {
 376             'skip_download': True,  # requires ffmpeg
 377         },
 378     }]
 379
 380     def _real_extract(self, url):
 381         mobj = re.match(self._VALID_URL, url)
 382         user_id = mobj.group('user_id')
 383         twid = mobj.group('id')
 384
 385         webpage, urlh = self._download_webpage_handle(
 386             self._TEMPLATE_STATUSES_URL % twid, twid)
 387
 388         if 'twitter.com/account/suspended' in urlh.geturl():
 389             raise ExtractorError('Account suspended by Twitter.', expected=True)
 390
 391         if user_id is None:
 392             mobj = re.match(self._VALID_URL, urlh.geturl())
 393             user_id = mobj.group('user_id')
 394
 395         username = remove_end(self._og_search_title(webpage), ' on Twitter')
 396
 397         title = description = self._og_search_description(webpage).strip('').replace('\n', ' ').strip('“”')
 398
 399         # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
 400         title = re.sub(r'\s+(https?://[^ ]+)', '', title)
 401
 402         info = {
 403             'uploader_id': user_id,
 404             'uploader': username,
 405             'webpage_url': url,
 406             'description': '%s on Twitter: "%s"' % (username, description),
 407             'title': username + ' - ' + title,
 408         }
 409
 410         mobj = re.search(r'''(?x)
 411             <video[^>]+class="animated-gif"(?P<more_info>[^>]+)>\s*
 412                 <source[^>]+video-src="(?P<url>[^"]+)"
 413         ''', webpage)
 414
 415         if mobj:
 416             more_info = mobj.group('more_info')
 417             height = int_or_none(self._search_regex(
 418                 r'data-height="(\d+)"', more_info, 'height', fatal=False))
 419             width = int_or_none(self._search_regex(
 420                 r'data-width="(\d+)"', more_info, 'width', fatal=False))
 421             thumbnail = self._search_regex(
 422                 r'poster="([^"]+)"', more_info, 'poster', fatal=False)
 423             info.update({
 424                 'id': twid,
 425                 'url': mobj.group('url'),
 426                 'height': height,
 427                 'width': width,
 428                 'thumbnail': thumbnail,
 429             })
 430             return info
 431
 432         twitter_card_url = None
 433         if 'class="PlayableMedia' in webpage:
 434             twitter_card_url = '%s//twitter.com/i/videos/tweet/%s' % (self.http_scheme(), twid)
 435         else:
 436             twitter_card_iframe_url = self._search_regex(
 437                 r'data-full-card-iframe-url=([\'"])(?P<url>(?:(?!\1).)+)\1',
 438                 webpage, 'Twitter card iframe URL', default=None, group='url')
 439             if twitter_card_iframe_url:
 440                 twitter_card_url = compat_urlparse.urljoin(url, twitter_card_iframe_url)
 441
 442         if twitter_card_url:
 443             info.update({
 444                 '_type': 'url_transparent',
 445                 'ie_key': 'TwitterCard',
 446                 'url': twitter_card_url,
 447             })
 448             return info
 449
 450         raise ExtractorError('There\'s no video in this tweet.')
 451
 452
 453 class TwitterAmplifyIE(TwitterBaseIE):
 454     IE_NAME = 'twitter:amplify'
 455     _VALID_URL = r'https?://amp\.twimg\.com/v/(?P<id>[0-9a-f\-]{36})'
 456
 457     _TEST = {
 458         'url': 'https://amp.twimg.com/v/0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 459         'md5': '7df102d0b9fd7066b86f3159f8e81bf6',
 460         'info_dict': {
 461             'id': '0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 462             'ext': 'mp4',
 463             'title': 'Twitter Video',
 464             'thumbnail': 're:^https?://.*',
 465         },
 466     }
 467
 468     def _real_extract(self, url):
 469         video_id = self._match_id(url)
 470         webpage = self._download_webpage(url, video_id)
 471
 472         vmap_url = self._html_search_meta(
 473             'twitter:amplify:vmap', webpage, 'vmap url')
 474         formats = self._extract_formats_from_vmap_url(vmap_url, video_id)
 475
 476         thumbnails = []
 477         thumbnail = self._html_search_meta(
 478             'twitter:image:src', webpage, 'thumbnail', fatal=False)
 479
 480         def _find_dimension(target):
 481             w = int_or_none(self._html_search_meta(
 482                 'twitter:%s:width' % target, webpage, fatal=False))
 483             h = int_or_none(self._html_search_meta(
 484                 'twitter:%s:height' % target, webpage, fatal=False))
 485             return w, h
 486
 487         if thumbnail:
 488             thumbnail_w, thumbnail_h = _find_dimension('image')
 489             thumbnails.append({
 490                 'url': thumbnail,
 491                 'width': thumbnail_w,
 492                 'height': thumbnail_h,
 493             })
 494
 495         video_w, video_h = _find_dimension('player')
 496         formats[0].update({
 497             'width': video_w,
 498             'height': video_h,
 499         })
 500
 501         return {
 502             'id': video_id,
 503             'title': 'Twitter Video',
 504             'formats': formats,
 505             'thumbnails': thumbnails,
 506         }