yt_dlp/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 (
   8     compat_HTTPError,
   9     compat_parse_qs,
  10     compat_urllib_parse_unquote,
  11     compat_urllib_parse_urlparse,
  12 )
  13 from ..utils import (
  14     dict_get,
  15     ExtractorError,
  16     float_or_none,
  17     int_or_none,
  18     traverse_obj,
  19     try_get,
  20     strip_or_none,
  21     unified_timestamp,
  22     update_url_query,
  23     url_or_none,
  24     xpath_text,
  25 )
  26
  27 from .periscope import (
  28     PeriscopeBaseIE,
  29     PeriscopeIE,
  30 )
  31
  32
  33 class TwitterBaseIE(InfoExtractor):
  34     _API_BASE = 'https://api.twitter.com/1.1/'
  35     _BASE_REGEX = r'https?://(?:(?:www|m(?:obile)?)\.)?twitter\.com/'
  36     _GUEST_TOKEN = None
  37
  38     def _extract_variant_formats(self, variant, video_id):
  39         variant_url = variant.get('url')
  40         if not variant_url:
  41             return [], {}
  42         elif '.m3u8' in variant_url:
  43             return self._extract_m3u8_formats_and_subtitles(
  44                 variant_url, video_id, 'mp4', 'm3u8_native',
  45                 m3u8_id='hls', fatal=False)
  46         else:
  47             tbr = int_or_none(dict_get(variant, ('bitrate', 'bit_rate')), 1000) or None
  48             f = {
  49                 'url': variant_url,
  50                 'format_id': 'http' + ('-%d' % tbr if tbr else ''),
  51                 'tbr': tbr,
  52             }
  53             self._search_dimensions_in_video_url(f, variant_url)
  54             return [f], {}
  55
  56     def _extract_formats_from_vmap_url(self, vmap_url, video_id):
  57         vmap_url = url_or_none(vmap_url)
  58         if not vmap_url:
  59             return [], {}
  60         vmap_data = self._download_xml(vmap_url, video_id)
  61         formats = []
  62         subtitles = {}
  63         urls = []
  64         for video_variant in vmap_data.findall('.//{http://twitter.com/schema/videoVMapV2.xsd}videoVariant'):
  65             video_variant.attrib['url'] = compat_urllib_parse_unquote(
  66                 video_variant.attrib['url'])
  67             urls.append(video_variant.attrib['url'])
  68             fmts, subs = self._extract_variant_formats(
  69                 video_variant.attrib, video_id)
  70             formats.extend(fmts)
  71             subtitles = self._merge_subtitles(subtitles, subs)
  72         video_url = strip_or_none(xpath_text(vmap_data, './/MediaFile'))
  73         if video_url not in urls:
  74             fmts, subs = self._extract_variant_formats({'url': video_url}, video_id)
  75             formats.extend(fmts)
  76             subtitles = self._merge_subtitles(subtitles, subs)
  77         return formats, subtitles
  78
  79     @staticmethod
  80     def _search_dimensions_in_video_url(a_format, video_url):
  81         m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
  82         if m:
  83             a_format.update({
  84                 'width': int(m.group('width')),
  85                 'height': int(m.group('height')),
  86             })
  87
  88     def _call_api(self, path, video_id, query={}):
  89         headers = {
  90             'Authorization': 'Bearer AAAAAAAAAAAAAAAAAAAAAPYXBAAAAAAACLXUNDekMxqa8h%2F40K4moUkGsoc%3DTYfbDKbT3jJPCEVnMYqilB28NHfOPqkca3qaAxGfsyKCs0wRbw',
  91         }
  92         if not self._GUEST_TOKEN:
  93             self._GUEST_TOKEN = self._download_json(
  94                 self._API_BASE + 'guest/activate.json', video_id,
  95                 'Downloading guest token', data=b'',
  96                 headers=headers)['guest_token']
  97         headers['x-guest-token'] = self._GUEST_TOKEN
  98         try:
  99             return self._download_json(
 100                 self._API_BASE + path, video_id, headers=headers, query=query)
 101         except ExtractorError as e:
 102             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 103                 raise ExtractorError(self._parse_json(
 104                     e.cause.read().decode(),
 105                     video_id)['errors'][0]['message'], expected=True)
 106             raise
 107
 108
 109 class TwitterCardIE(InfoExtractor):
 110     IE_NAME = 'twitter:card'
 111     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/(?:cards/tfw/v1|videos(?:/tweet)?)/(?P<id>\d+)'
 112     _TESTS = [
 113         {
 114             'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
 115             # MD5 checksums are different in different places
 116             'info_dict': {
 117                 'id': '560070183650213889',
 118                 'ext': 'mp4',
 119                 'title': "Twitter - You can now shoot, edit and share video on Twitter. Capture life's most moving moments from your perspective.",
 120                 'description': 'md5:18d3e24bb4f6e5007487dd546e53bd96',
 121                 'uploader': 'Twitter',
 122                 'uploader_id': 'Twitter',
 123                 'thumbnail': r're:^https?://.*\.jpg',
 124                 'duration': 30.033,
 125                 'timestamp': 1422366112,
 126                 'upload_date': '20150127',
 127             },
 128         },
 129         {
 130             'url': 'https://twitter.com/i/cards/tfw/v1/623160978427936768',
 131             'md5': '7137eca597f72b9abbe61e5ae0161399',
 132             'info_dict': {
 133                 'id': '623160978427936768',
 134                 'ext': 'mp4',
 135                 'title': "NASA - Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video.",
 136                 'description': "Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video. https://t.co/BJYgOjSeGA",
 137                 'uploader': 'NASA',
 138                 'uploader_id': 'NASA',
 139                 'timestamp': 1437408129,
 140                 'upload_date': '20150720',
 141             },
 142         },
 143         {
 144             'url': 'https://twitter.com/i/cards/tfw/v1/654001591733886977',
 145             'md5': 'b6d9683dd3f48e340ded81c0e917ad46',
 146             'info_dict': {
 147                 'id': 'dq4Oj5quskI',
 148                 'ext': 'mp4',
 149                 'title': 'Ubuntu 11.10 Overview',
 150                 'description': 'md5:a831e97fa384863d6e26ce48d1c43376',
 151                 'upload_date': '20111013',
 152                 'uploader': 'OMG! UBUNTU!',
 153                 'uploader_id': 'omgubuntu',
 154             },
 155             'add_ie': ['Youtube'],
 156         },
 157         {
 158             'url': 'https://twitter.com/i/cards/tfw/v1/665289828897005568',
 159             'md5': '6dabeaca9e68cbb71c99c322a4b42a11',
 160             'info_dict': {
 161                 'id': 'iBb2x00UVlv',
 162                 'ext': 'mp4',
 163                 'upload_date': '20151113',
 164                 'uploader_id': '1189339351084113920',
 165                 'uploader': 'ArsenalTerje',
 166                 'title': 'Vine by ArsenalTerje',
 167                 'timestamp': 1447451307,
 168             },
 169             'add_ie': ['Vine'],
 170         }, {
 171             'url': 'https://twitter.com/i/videos/tweet/705235433198714880',
 172             'md5': '884812a2adc8aaf6fe52b15ccbfa3b88',
 173             'info_dict': {
 174                 'id': '705235433198714880',
 175                 'ext': 'mp4',
 176                 'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 177                 'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 178                 'uploader': 'Brent Yarina',
 179                 'uploader_id': 'BTNBrentYarina',
 180                 'timestamp': 1456976204,
 181                 'upload_date': '20160303',
 182             },
 183             'skip': 'This content is no longer available.',
 184         }, {
 185             'url': 'https://twitter.com/i/videos/752274308186120192',
 186             'only_matching': True,
 187         },
 188     ]
 189
 190     def _real_extract(self, url):
 191         status_id = self._match_id(url)
 192         return self.url_result(
 193             'https://twitter.com/statuses/' + status_id,
 194             TwitterIE.ie_key(), status_id)
 195
 196
 197 class TwitterIE(TwitterBaseIE):
 198     IE_NAME = 'twitter'
 199     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'(?:(?:i/web|[^/]+)/status|statuses)/(?P<id>\d+)'
 200
 201     _TESTS = [{
 202         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
 203         'info_dict': {
 204             'id': '643211948184596480',
 205             'ext': 'mp4',
 206             'title': 'FREE THE NIPPLE - FTN supporters on Hollywood Blvd today!',
 207             'thumbnail': r're:^https?://.*\.jpg',
 208             'description': 'FTN supporters on Hollywood Blvd today! http://t.co/c7jHH749xJ',
 209             'uploader': 'FREE THE NIPPLE',
 210             'uploader_id': 'freethenipple',
 211             'duration': 12.922,
 212             'timestamp': 1442188653,
 213             'upload_date': '20150913',
 214             'age_limit': 18,
 215         },
 216     }, {
 217         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
 218         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
 219         'info_dict': {
 220             'id': '657991469417025536',
 221             'ext': 'mp4',
 222             'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
 223             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
 224             'thumbnail': r're:^https?://.*\.png',
 225             'uploader': 'Gifs',
 226             'uploader_id': 'giphz',
 227         },
 228         'expected_warnings': ['height', 'width'],
 229         'skip': 'Account suspended',
 230     }, {
 231         'url': 'https://twitter.com/starwars/status/665052190608723968',
 232         'info_dict': {
 233             'id': '665052190608723968',
 234             'ext': 'mp4',
 235             'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
 236             'description': 'A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens. https://t.co/OkSqT2fjWJ',
 237             'uploader_id': 'starwars',
 238             'uploader': 'Star Wars',
 239             'timestamp': 1447395772,
 240             'upload_date': '20151113',
 241         },
 242     }, {
 243         'url': 'https://twitter.com/BTNBrentYarina/status/705235433198714880',
 244         'info_dict': {
 245             'id': '705235433198714880',
 246             'ext': 'mp4',
 247             'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 248             'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 249             'uploader_id': 'BTNBrentYarina',
 250             'uploader': 'Brent Yarina',
 251             'timestamp': 1456976204,
 252             'upload_date': '20160303',
 253         },
 254         'params': {
 255             # The same video as https://twitter.com/i/videos/tweet/705235433198714880
 256             # Test case of TwitterCardIE
 257             'skip_download': True,
 258         },
 259     }, {
 260         'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
 261         'info_dict': {
 262             'id': '700207533655363584',
 263             'ext': 'mp4',
 264             'title': 'simon vertugo - BEAT PROD: @suhmeduh #Damndaniel',
 265             'description': 'BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ',
 266             'thumbnail': r're:^https?://.*\.jpg',
 267             'uploader': 'simon vertugo',
 268             'uploader_id': 'simonvertugo',
 269             'duration': 30.0,
 270             'timestamp': 1455777459,
 271             'upload_date': '20160218',
 272         },
 273     }, {
 274         'url': 'https://twitter.com/Filmdrunk/status/713801302971588609',
 275         'md5': '89a15ed345d13b86e9a5a5e051fa308a',
 276         'info_dict': {
 277             'id': 'MIOxnrUteUd',
 278             'ext': 'mp4',
 279             'title': 'Dr.Pepperの飲み方 #japanese #バカ #ドクペ #電動ガン',
 280             'uploader': 'TAKUMA',
 281             'uploader_id': '1004126642786242560',
 282             'timestamp': 1402826626,
 283             'upload_date': '20140615',
 284         },
 285         'add_ie': ['Vine'],
 286     }, {
 287         'url': 'https://twitter.com/captainamerica/status/719944021058060289',
 288         'info_dict': {
 289             'id': '719944021058060289',
 290             'ext': 'mp4',
 291             'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
 292             'description': '@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI',
 293             'uploader_id': 'CaptainAmerica',
 294             'uploader': 'Captain America',
 295             'duration': 3.17,
 296             'timestamp': 1460483005,
 297             'upload_date': '20160412',
 298         },
 299     }, {
 300         'url': 'https://twitter.com/OPP_HSD/status/779210622571536384',
 301         'info_dict': {
 302             'id': '1zqKVVlkqLaKB',
 303             'ext': 'mp4',
 304             'title': 'Sgt Kerry Schmidt - Ontario Provincial Police - Road rage, mischief, assault, rollover and fire in one occurrence',
 305             'upload_date': '20160923',
 306             'uploader_id': '1PmKqpJdOJQoY',
 307             'uploader': 'Sgt Kerry Schmidt - Ontario Provincial Police',
 308             'timestamp': 1474613214,
 309         },
 310         'add_ie': ['Periscope'],
 311     }, {
 312         # has mp4 formats via mobile API
 313         'url': 'https://twitter.com/news_al3alm/status/852138619213144067',
 314         'info_dict': {
 315             'id': '852138619213144067',
 316             'ext': 'mp4',
 317             'title': 'عالم الأخبار - كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة',
 318             'description': 'كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة   https://t.co/xg6OhpyKfN',
 319             'uploader': 'عالم الأخبار',
 320             'uploader_id': 'news_al3alm',
 321             'duration': 277.4,
 322             'timestamp': 1492000653,
 323             'upload_date': '20170412',
 324         },
 325         'skip': 'Account suspended',
 326     }, {
 327         'url': 'https://twitter.com/i/web/status/910031516746514432',
 328         'info_dict': {
 329             'id': '910031516746514432',
 330             'ext': 'mp4',
 331             '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.',
 332             'thumbnail': r're:^https?://.*\.jpg',
 333             'description': '[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',
 334             'uploader': 'Préfet de Guadeloupe',
 335             'uploader_id': 'Prefet971',
 336             'duration': 47.48,
 337             'timestamp': 1505803395,
 338             'upload_date': '20170919',
 339         },
 340         'params': {
 341             'skip_download': True,  # requires ffmpeg
 342         },
 343     }, {
 344         # card via api.twitter.com/1.1/videos/tweet/config
 345         'url': 'https://twitter.com/LisPower1/status/1001551623938805763',
 346         'info_dict': {
 347             'id': '1001551623938805763',
 348             'ext': 'mp4',
 349             'title': 're:.*?Shep is on a roll today.*?',
 350             'thumbnail': r're:^https?://.*\.jpg',
 351             'description': 'md5:37b9f2ff31720cef23b2bd42ee8a0f09',
 352             'uploader': 'Lis Power',
 353             'uploader_id': 'LisPower1',
 354             'duration': 111.278,
 355             'timestamp': 1527623489,
 356             'upload_date': '20180529',
 357         },
 358         'params': {
 359             'skip_download': True,  # requires ffmpeg
 360         },
 361     }, {
 362         'url': 'https://twitter.com/foobar/status/1087791357756956680',
 363         'info_dict': {
 364             'id': '1087791357756956680',
 365             'ext': 'mp4',
 366             'title': 'Twitter - A new is coming.  Some of you got an opt-in to try it now. Check out the emoji button, quick keyboard shortcuts, upgraded trends, advanced search, and more. Let us know your thoughts!',
 367             'thumbnail': r're:^https?://.*\.jpg',
 368             'description': 'md5:6dfd341a3310fb97d80d2bf7145df976',
 369             'uploader': 'Twitter',
 370             'uploader_id': 'Twitter',
 371             'duration': 61.567,
 372             'timestamp': 1548184644,
 373             'upload_date': '20190122',
 374         },
 375     }, {
 376         # not available in Periscope
 377         'url': 'https://twitter.com/ViviEducation/status/1136534865145286656',
 378         'info_dict': {
 379             'id': '1vOGwqejwoWxB',
 380             'ext': 'mp4',
 381             'title': 'Vivi - Vivi founder @lior_rauchy announcing our new student feedback tool live at @EduTECH_AU #EduTECH2019',
 382             'uploader': 'Vivi',
 383             'uploader_id': '1eVjYOLGkGrQL',
 384         },
 385         'add_ie': ['TwitterBroadcast'],
 386     }, {
 387         # unified card
 388         'url': 'https://twitter.com/BrooklynNets/status/1349794411333394432?s=20',
 389         'info_dict': {
 390             'id': '1349794411333394432',
 391             'ext': 'mp4',
 392             'title': 'md5:d1c4941658e4caaa6cb579260d85dcba',
 393             'thumbnail': r're:^https?://.*\.jpg',
 394             'description': 'md5:71ead15ec44cee55071547d6447c6a3e',
 395             'uploader': 'Brooklyn Nets',
 396             'uploader_id': 'BrooklynNets',
 397             'duration': 324.484,
 398             'timestamp': 1610651040,
 399             'upload_date': '20210114',
 400         },
 401         'params': {
 402             'skip_download': True,
 403         },
 404     }, {
 405         # Twitch Clip Embed
 406         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 407         'only_matching': True,
 408     }, {
 409         # promo_video_website card
 410         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 411         'only_matching': True,
 412     }, {
 413         # promo_video_convo card
 414         'url': 'https://twitter.com/poco_dandy/status/1047395834013384704',
 415         'only_matching': True,
 416     }, {
 417         # appplayer card
 418         'url': 'https://twitter.com/poco_dandy/status/1150646424461176832',
 419         'only_matching': True,
 420     }, {
 421         # video_direct_message card
 422         'url': 'https://twitter.com/qarev001/status/1348948114569269251',
 423         'only_matching': True,
 424     }, {
 425         # poll2choice_video card
 426         'url': 'https://twitter.com/CAF_Online/status/1349365911120195585',
 427         'only_matching': True,
 428     }, {
 429         # poll3choice_video card
 430         'url': 'https://twitter.com/SamsungMobileSA/status/1348609186725289984',
 431         'only_matching': True,
 432     }, {
 433         # poll4choice_video card
 434         'url': 'https://twitter.com/SouthamptonFC/status/1347577658079641604',
 435         'only_matching': True,
 436     }]
 437
 438     def _real_extract(self, url):
 439         twid = self._match_id(url)
 440         status = self._call_api(
 441             'statuses/show/%s.json' % twid, twid, {
 442                 'cards_platform': 'Web-12',
 443                 'include_cards': 1,
 444                 'include_reply_count': 1,
 445                 'include_user_entities': 0,
 446                 'tweet_mode': 'extended',
 447             })
 448
 449         title = description = status['full_text'].replace('\n', ' ')
 450         # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
 451         title = re.sub(r'\s+(https?://[^ ]+)', '', title)
 452         user = status.get('user') or {}
 453         uploader = user.get('name')
 454         if uploader:
 455             title = '%s - %s' % (uploader, title)
 456         uploader_id = user.get('screen_name')
 457
 458         tags = []
 459         for hashtag in (try_get(status, lambda x: x['entities']['hashtags'], list) or []):
 460             hashtag_text = hashtag.get('text')
 461             if not hashtag_text:
 462                 continue
 463             tags.append(hashtag_text)
 464
 465         info = {
 466             'id': twid,
 467             'title': title,
 468             'description': description,
 469             'uploader': uploader,
 470             'timestamp': unified_timestamp(status.get('created_at')),
 471             'uploader_id': uploader_id,
 472             'uploader_url': 'https://twitter.com/' + uploader_id if uploader_id else None,
 473             'like_count': int_or_none(status.get('favorite_count')),
 474             'repost_count': int_or_none(status.get('retweet_count')),
 475             'comment_count': int_or_none(status.get('reply_count')),
 476             'age_limit': 18 if status.get('possibly_sensitive') else 0,
 477             'tags': tags,
 478         }
 479
 480         def extract_from_video_info(media):
 481             video_info = media.get('video_info') or {}
 482
 483             formats = []
 484             subtitles = {}
 485             for variant in video_info.get('variants', []):
 486                 fmts, subs = self._extract_variant_formats(variant, twid)
 487                 subtitles = self._merge_subtitles(subtitles, subs)
 488                 formats.extend(fmts)
 489             self._sort_formats(formats, ('res', 'br', 'size', 'proto'))  # The codec of http formats are unknown
 490
 491             thumbnails = []
 492             media_url = media.get('media_url_https') or media.get('media_url')
 493             if media_url:
 494                 def add_thumbnail(name, size):
 495                     thumbnails.append({
 496                         'id': name,
 497                         'url': update_url_query(media_url, {'name': name}),
 498                         'width': int_or_none(size.get('w') or size.get('width')),
 499                         'height': int_or_none(size.get('h') or size.get('height')),
 500                     })
 501                 for name, size in media.get('sizes', {}).items():
 502                     add_thumbnail(name, size)
 503                 add_thumbnail('orig', media.get('original_info') or {})
 504
 505             info.update({
 506                 'formats': formats,
 507                 'subtitles': subtitles,
 508                 'thumbnails': thumbnails,
 509                 'duration': float_or_none(video_info.get('duration_millis'), 1000),
 510             })
 511
 512         media = traverse_obj(status, ((None, 'quoted_status'), 'extended_entities', 'media', 0), get_all=False)
 513         if media and media.get('type') != 'photo':
 514             extract_from_video_info(media)
 515         else:
 516             card = status.get('card')
 517             if card:
 518                 binding_values = card['binding_values']
 519
 520                 def get_binding_value(k):
 521                     o = binding_values.get(k) or {}
 522                     return try_get(o, lambda x: x[x['type'].lower() + '_value'])
 523
 524                 card_name = card['name'].split(':')[-1]
 525                 if card_name == 'player':
 526                     info.update({
 527                         '_type': 'url',
 528                         'url': get_binding_value('player_url'),
 529                     })
 530                 elif card_name == 'periscope_broadcast':
 531                     info.update({
 532                         '_type': 'url',
 533                         'url': get_binding_value('url') or get_binding_value('player_url'),
 534                         'ie_key': PeriscopeIE.ie_key(),
 535                     })
 536                 elif card_name == 'broadcast':
 537                     info.update({
 538                         '_type': 'url',
 539                         'url': get_binding_value('broadcast_url'),
 540                         'ie_key': TwitterBroadcastIE.ie_key(),
 541                     })
 542                 elif card_name == 'summary':
 543                     info.update({
 544                         '_type': 'url',
 545                         'url': get_binding_value('card_url'),
 546                     })
 547                 elif card_name == 'unified_card':
 548                     media_entities = self._parse_json(get_binding_value('unified_card'), twid)['media_entities']
 549                     extract_from_video_info(next(iter(media_entities.values())))
 550                 # amplify, promo_video_website, promo_video_convo, appplayer,
 551                 # video_direct_message, poll2choice_video, poll3choice_video,
 552                 # poll4choice_video, ...
 553                 else:
 554                     is_amplify = card_name == 'amplify'
 555                     vmap_url = get_binding_value('amplify_url_vmap') if is_amplify else get_binding_value('player_stream_url')
 556                     content_id = get_binding_value('%s_content_id' % (card_name if is_amplify else 'player'))
 557                     formats, subtitles = self._extract_formats_from_vmap_url(vmap_url, content_id or twid)
 558                     self._sort_formats(formats)
 559
 560                     thumbnails = []
 561                     for suffix in ('_small', '', '_large', '_x_large', '_original'):
 562                         image = get_binding_value('player_image' + suffix) or {}
 563                         image_url = image.get('url')
 564                         if not image_url or '/player-placeholder' in image_url:
 565                             continue
 566                         thumbnails.append({
 567                             'id': suffix[1:] if suffix else 'medium',
 568                             'url': image_url,
 569                             'width': int_or_none(image.get('width')),
 570                             'height': int_or_none(image.get('height')),
 571                         })
 572
 573                     info.update({
 574                         'formats': formats,
 575                         'subtitles': subtitles,
 576                         'thumbnails': thumbnails,
 577                         'duration': int_or_none(get_binding_value(
 578                             'content_duration_seconds')),
 579                     })
 580             else:
 581                 expanded_url = try_get(status, lambda x: x['entities']['urls'][0]['expanded_url'])
 582                 if not expanded_url:
 583                     raise ExtractorError("There's no video in this tweet.")
 584                 info.update({
 585                     '_type': 'url',
 586                     'url': expanded_url,
 587                 })
 588         return info
 589
 590
 591 class TwitterAmplifyIE(TwitterBaseIE):
 592     IE_NAME = 'twitter:amplify'
 593     _VALID_URL = r'https?://amp\.twimg\.com/v/(?P<id>[0-9a-f\-]{36})'
 594
 595     _TEST = {
 596         'url': 'https://amp.twimg.com/v/0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 597         'md5': '7df102d0b9fd7066b86f3159f8e81bf6',
 598         'info_dict': {
 599             'id': '0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 600             'ext': 'mp4',
 601             'title': 'Twitter Video',
 602             'thumbnail': 're:^https?://.*',
 603         },
 604     }
 605
 606     def _real_extract(self, url):
 607         video_id = self._match_id(url)
 608         webpage = self._download_webpage(url, video_id)
 609
 610         vmap_url = self._html_search_meta(
 611             'twitter:amplify:vmap', webpage, 'vmap url')
 612         formats = self._extract_formats_from_vmap_url(vmap_url, video_id)
 613
 614         thumbnails = []
 615         thumbnail = self._html_search_meta(
 616             'twitter:image:src', webpage, 'thumbnail', fatal=False)
 617
 618         def _find_dimension(target):
 619             w = int_or_none(self._html_search_meta(
 620                 'twitter:%s:width' % target, webpage, fatal=False))
 621             h = int_or_none(self._html_search_meta(
 622                 'twitter:%s:height' % target, webpage, fatal=False))
 623             return w, h
 624
 625         if thumbnail:
 626             thumbnail_w, thumbnail_h = _find_dimension('image')
 627             thumbnails.append({
 628                 'url': thumbnail,
 629                 'width': thumbnail_w,
 630                 'height': thumbnail_h,
 631             })
 632
 633         video_w, video_h = _find_dimension('player')
 634         formats[0].update({
 635             'width': video_w,
 636             'height': video_h,
 637         })
 638
 639         return {
 640             'id': video_id,
 641             'title': 'Twitter Video',
 642             'formats': formats,
 643             'thumbnails': thumbnails,
 644         }
 645
 646
 647 class TwitterBroadcastIE(TwitterBaseIE, PeriscopeBaseIE):
 648     IE_NAME = 'twitter:broadcast'
 649     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/broadcasts/(?P<id>[0-9a-zA-Z]{13})'
 650
 651     _TEST = {
 652         # untitled Periscope video
 653         'url': 'https://twitter.com/i/broadcasts/1yNGaQLWpejGj',
 654         'info_dict': {
 655             'id': '1yNGaQLWpejGj',
 656             'ext': 'mp4',
 657             'title': 'Andrea May Sahouri - Periscope Broadcast',
 658             'uploader': 'Andrea May Sahouri',
 659             'uploader_id': '1PXEdBZWpGwKe',
 660         },
 661     }
 662
 663     def _real_extract(self, url):
 664         broadcast_id = self._match_id(url)
 665         broadcast = self._call_api(
 666             'broadcasts/show.json', broadcast_id,
 667             {'ids': broadcast_id})['broadcasts'][broadcast_id]
 668         info = self._parse_broadcast_data(broadcast, broadcast_id)
 669         media_key = broadcast['media_key']
 670         source = self._call_api(
 671             'live_video_stream/status/' + media_key, media_key)['source']
 672         m3u8_url = source.get('noRedirectPlaybackUrl') or source['location']
 673         if '/live_video_stream/geoblocked/' in m3u8_url:
 674             self.raise_geo_restricted()
 675         m3u8_id = compat_parse_qs(compat_urllib_parse_urlparse(
 676             m3u8_url).query).get('type', [None])[0]
 677         state, width, height = self._extract_common_format_info(broadcast)
 678         info['formats'] = self._extract_pscp_m3u8_formats(
 679             m3u8_url, broadcast_id, m3u8_id, state, width, height)
 680         return info
 681
 682
 683 class TwitterShortenerIE(TwitterBaseIE):
 684     IE_NAME = 'twitter:shortener'
 685     _VALID_URL = r'https?://t.co/(?P<id>[^?]+)|tco:(?P<eid>[^?]+)'
 686     _BASE_URL = 'https://t.co/'
 687
 688     def _real_extract(self, url):
 689         mobj = self._match_valid_url(url)
 690         eid, id = mobj.group('eid', 'id')
 691         if eid:
 692             id = eid
 693             url = self._BASE_URL + id
 694         new_url = self._request_webpage(url, id, headers={'User-Agent': 'curl'}).geturl()
 695         __UNSAFE_LINK = "https://twitter.com/safety/unsafe_link_warning?unsafe_link="
 696         if new_url.startswith(__UNSAFE_LINK):
 697             new_url = new_url.replace(__UNSAFE_LINK, "")
 698         return self.url_result(new_url)