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     format_field,
  17     float_or_none,
  18     int_or_none,
  19     traverse_obj,
  20     try_get,
  21     strip_or_none,
  22     unified_timestamp,
  23     update_url_query,
  24     url_or_none,
  25     xpath_text,
  26 )
  27
  28 from .periscope import (
  29     PeriscopeBaseIE,
  30     PeriscopeIE,
  31 )
  32
  33
  34 class TwitterBaseIE(InfoExtractor):
  35     _API_BASE = 'https://api.twitter.com/1.1/'
  36     _BASE_REGEX = r'https?://(?:(?:www|m(?:obile)?)\.)?twitter\.com/'
  37     _GUEST_TOKEN = None
  38
  39     def _extract_variant_formats(self, variant, video_id):
  40         variant_url = variant.get('url')
  41         if not variant_url:
  42             return [], {}
  43         elif '.m3u8' in variant_url:
  44             return self._extract_m3u8_formats_and_subtitles(
  45                 variant_url, video_id, 'mp4', 'm3u8_native',
  46                 m3u8_id='hls', fatal=False)
  47         else:
  48             tbr = int_or_none(dict_get(variant, ('bitrate', 'bit_rate')), 1000) or None
  49             f = {
  50                 'url': variant_url,
  51                 'format_id': 'http' + ('-%d' % tbr if tbr else ''),
  52                 'tbr': tbr,
  53             }
  54             self._search_dimensions_in_video_url(f, variant_url)
  55             return [f], {}
  56
  57     def _extract_formats_from_vmap_url(self, vmap_url, video_id):
  58         vmap_url = url_or_none(vmap_url)
  59         if not vmap_url:
  60             return [], {}
  61         vmap_data = self._download_xml(vmap_url, video_id)
  62         formats = []
  63         subtitles = {}
  64         urls = []
  65         for video_variant in vmap_data.findall('.//{http://twitter.com/schema/videoVMapV2.xsd}videoVariant'):
  66             video_variant.attrib['url'] = compat_urllib_parse_unquote(
  67                 video_variant.attrib['url'])
  68             urls.append(video_variant.attrib['url'])
  69             fmts, subs = self._extract_variant_formats(
  70                 video_variant.attrib, video_id)
  71             formats.extend(fmts)
  72             subtitles = self._merge_subtitles(subtitles, subs)
  73         video_url = strip_or_none(xpath_text(vmap_data, './/MediaFile'))
  74         if video_url not in urls:
  75             fmts, subs = self._extract_variant_formats({'url': video_url}, video_id)
  76             formats.extend(fmts)
  77             subtitles = self._merge_subtitles(subtitles, subs)
  78         return formats, subtitles
  79
  80     @staticmethod
  81     def _search_dimensions_in_video_url(a_format, video_url):
  82         m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
  83         if m:
  84             a_format.update({
  85                 'width': int(m.group('width')),
  86                 'height': int(m.group('height')),
  87             })
  88
  89     def _call_api(self, path, video_id, query={}):
  90         headers = {
  91             'Authorization': 'Bearer AAAAAAAAAAAAAAAAAAAAAPYXBAAAAAAACLXUNDekMxqa8h%2F40K4moUkGsoc%3DTYfbDKbT3jJPCEVnMYqilB28NHfOPqkca3qaAxGfsyKCs0wRbw',
  92         }
  93         token = self._get_cookies(self._API_BASE).get('ct0')
  94         if token:
  95             headers['x-csrf-token'] = token.value
  96         if not self._GUEST_TOKEN:
  97             self._GUEST_TOKEN = self._download_json(
  98                 self._API_BASE + 'guest/activate.json', video_id,
  99                 'Downloading guest token', data=b'',
 100                 headers=headers)['guest_token']
 101         headers['x-guest-token'] = self._GUEST_TOKEN
 102         try:
 103             return self._download_json(
 104                 self._API_BASE + path, video_id, headers=headers, query=query)
 105         except ExtractorError as e:
 106             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 107                 raise ExtractorError(self._parse_json(
 108                     e.cause.read().decode(),
 109                     video_id)['errors'][0]['message'], expected=True)
 110             raise
 111
 112
 113 class TwitterCardIE(InfoExtractor):
 114     IE_NAME = 'twitter:card'
 115     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/(?:cards/tfw/v1|videos(?:/tweet)?)/(?P<id>\d+)'
 116     _TESTS = [
 117         {
 118             'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
 119             # MD5 checksums are different in different places
 120             'info_dict': {
 121                 'id': '560070183650213889',
 122                 'ext': 'mp4',
 123                 'title': "Twitter - You can now shoot, edit and share video on Twitter. Capture life's most moving moments from your perspective.",
 124                 'description': 'md5:18d3e24bb4f6e5007487dd546e53bd96',
 125                 'uploader': 'Twitter',
 126                 'uploader_id': 'Twitter',
 127                 'thumbnail': r're:^https?://.*\.jpg',
 128                 'duration': 30.033,
 129                 'timestamp': 1422366112,
 130                 'upload_date': '20150127',
 131             },
 132         },
 133         {
 134             'url': 'https://twitter.com/i/cards/tfw/v1/623160978427936768',
 135             'md5': '7137eca597f72b9abbe61e5ae0161399',
 136             'info_dict': {
 137                 'id': '623160978427936768',
 138                 'ext': 'mp4',
 139                 'title': "NASA - Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video.",
 140                 'description': "Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video. https://t.co/BJYgOjSeGA",
 141                 'uploader': 'NASA',
 142                 'uploader_id': 'NASA',
 143                 'timestamp': 1437408129,
 144                 'upload_date': '20150720',
 145             },
 146         },
 147         {
 148             'url': 'https://twitter.com/i/cards/tfw/v1/654001591733886977',
 149             'md5': 'b6d9683dd3f48e340ded81c0e917ad46',
 150             'info_dict': {
 151                 'id': 'dq4Oj5quskI',
 152                 'ext': 'mp4',
 153                 'title': 'Ubuntu 11.10 Overview',
 154                 'description': 'md5:a831e97fa384863d6e26ce48d1c43376',
 155                 'upload_date': '20111013',
 156                 'uploader': 'OMG! UBUNTU!',
 157                 'uploader_id': 'omgubuntu',
 158             },
 159             'add_ie': ['Youtube'],
 160         },
 161         {
 162             'url': 'https://twitter.com/i/cards/tfw/v1/665289828897005568',
 163             'md5': '6dabeaca9e68cbb71c99c322a4b42a11',
 164             'info_dict': {
 165                 'id': 'iBb2x00UVlv',
 166                 'ext': 'mp4',
 167                 'upload_date': '20151113',
 168                 'uploader_id': '1189339351084113920',
 169                 'uploader': 'ArsenalTerje',
 170                 'title': 'Vine by ArsenalTerje',
 171                 'timestamp': 1447451307,
 172             },
 173             'add_ie': ['Vine'],
 174         }, {
 175             'url': 'https://twitter.com/i/videos/tweet/705235433198714880',
 176             'md5': '884812a2adc8aaf6fe52b15ccbfa3b88',
 177             'info_dict': {
 178                 'id': '705235433198714880',
 179                 'ext': 'mp4',
 180                 'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 181                 'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 182                 'uploader': 'Brent Yarina',
 183                 'uploader_id': 'BTNBrentYarina',
 184                 'timestamp': 1456976204,
 185                 'upload_date': '20160303',
 186             },
 187             'skip': 'This content is no longer available.',
 188         }, {
 189             'url': 'https://twitter.com/i/videos/752274308186120192',
 190             'only_matching': True,
 191         },
 192     ]
 193
 194     def _real_extract(self, url):
 195         status_id = self._match_id(url)
 196         return self.url_result(
 197             'https://twitter.com/statuses/' + status_id,
 198             TwitterIE.ie_key(), status_id)
 199
 200
 201 class TwitterIE(TwitterBaseIE):
 202     IE_NAME = 'twitter'
 203     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'(?:(?:i/web|[^/]+)/status|statuses)/(?P<id>\d+)'
 204
 205     _TESTS = [{
 206         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
 207         'info_dict': {
 208             'id': '643211948184596480',
 209             'ext': 'mp4',
 210             'title': 'FREE THE NIPPLE - FTN supporters on Hollywood Blvd today!',
 211             'thumbnail': r're:^https?://.*\.jpg',
 212             'description': 'FTN supporters on Hollywood Blvd today! http://t.co/c7jHH749xJ',
 213             'uploader': 'FREE THE NIPPLE',
 214             'uploader_id': 'freethenipple',
 215             'duration': 12.922,
 216             'timestamp': 1442188653,
 217             'upload_date': '20150913',
 218             'age_limit': 18,
 219         },
 220     }, {
 221         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
 222         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
 223         'info_dict': {
 224             'id': '657991469417025536',
 225             'ext': 'mp4',
 226             'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
 227             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
 228             'thumbnail': r're:^https?://.*\.png',
 229             'uploader': 'Gifs',
 230             'uploader_id': 'giphz',
 231         },
 232         'expected_warnings': ['height', 'width'],
 233         'skip': 'Account suspended',
 234     }, {
 235         'url': 'https://twitter.com/starwars/status/665052190608723968',
 236         'info_dict': {
 237             'id': '665052190608723968',
 238             'ext': 'mp4',
 239             'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
 240             'description': 'A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens. https://t.co/OkSqT2fjWJ',
 241             'uploader_id': 'starwars',
 242             'uploader': 'Star Wars',
 243             'timestamp': 1447395772,
 244             'upload_date': '20151113',
 245         },
 246     }, {
 247         'url': 'https://twitter.com/BTNBrentYarina/status/705235433198714880',
 248         'info_dict': {
 249             'id': '705235433198714880',
 250             'ext': 'mp4',
 251             'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 252             'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 253             'uploader_id': 'BTNBrentYarina',
 254             'uploader': 'Brent Yarina',
 255             'timestamp': 1456976204,
 256             'upload_date': '20160303',
 257         },
 258         'params': {
 259             # The same video as https://twitter.com/i/videos/tweet/705235433198714880
 260             # Test case of TwitterCardIE
 261             'skip_download': True,
 262         },
 263     }, {
 264         'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
 265         'info_dict': {
 266             'id': '700207533655363584',
 267             'ext': 'mp4',
 268             'title': 'simon vertugo - BEAT PROD: @suhmeduh #Damndaniel',
 269             'description': 'BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ',
 270             'thumbnail': r're:^https?://.*\.jpg',
 271             'uploader': 'simon vertugo',
 272             'uploader_id': 'simonvertugo',
 273             'duration': 30.0,
 274             'timestamp': 1455777459,
 275             'upload_date': '20160218',
 276         },
 277     }, {
 278         'url': 'https://twitter.com/Filmdrunk/status/713801302971588609',
 279         'md5': '89a15ed345d13b86e9a5a5e051fa308a',
 280         'info_dict': {
 281             'id': 'MIOxnrUteUd',
 282             'ext': 'mp4',
 283             'title': 'Dr.Pepperの飲み方 #japanese #バカ #ドクペ #電動ガン',
 284             'uploader': 'TAKUMA',
 285             'uploader_id': '1004126642786242560',
 286             'timestamp': 1402826626,
 287             'upload_date': '20140615',
 288         },
 289         'add_ie': ['Vine'],
 290     }, {
 291         'url': 'https://twitter.com/captainamerica/status/719944021058060289',
 292         'info_dict': {
 293             'id': '719944021058060289',
 294             'ext': 'mp4',
 295             'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
 296             'description': '@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI',
 297             'uploader_id': 'CaptainAmerica',
 298             'uploader': 'Captain America',
 299             'duration': 3.17,
 300             'timestamp': 1460483005,
 301             'upload_date': '20160412',
 302         },
 303     }, {
 304         'url': 'https://twitter.com/OPP_HSD/status/779210622571536384',
 305         'info_dict': {
 306             'id': '1zqKVVlkqLaKB',
 307             'ext': 'mp4',
 308             'title': 'Sgt Kerry Schmidt - Ontario Provincial Police - Road rage, mischief, assault, rollover and fire in one occurrence',
 309             'upload_date': '20160923',
 310             'uploader_id': '1PmKqpJdOJQoY',
 311             'uploader': 'Sgt Kerry Schmidt - Ontario Provincial Police',
 312             'timestamp': 1474613214,
 313         },
 314         'add_ie': ['Periscope'],
 315     }, {
 316         # has mp4 formats via mobile API
 317         'url': 'https://twitter.com/news_al3alm/status/852138619213144067',
 318         'info_dict': {
 319             'id': '852138619213144067',
 320             'ext': 'mp4',
 321             'title': 'عالم الأخبار - كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة',
 322             'description': 'كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة   https://t.co/xg6OhpyKfN',
 323             'uploader': 'عالم الأخبار',
 324             'uploader_id': 'news_al3alm',
 325             'duration': 277.4,
 326             'timestamp': 1492000653,
 327             'upload_date': '20170412',
 328         },
 329         'skip': 'Account suspended',
 330     }, {
 331         'url': 'https://twitter.com/i/web/status/910031516746514432',
 332         'info_dict': {
 333             'id': '910031516746514432',
 334             'ext': 'mp4',
 335             '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.',
 336             'thumbnail': r're:^https?://.*\.jpg',
 337             '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',
 338             'uploader': 'Préfet de Guadeloupe',
 339             'uploader_id': 'Prefet971',
 340             'duration': 47.48,
 341             'timestamp': 1505803395,
 342             'upload_date': '20170919',
 343         },
 344         'params': {
 345             'skip_download': True,  # requires ffmpeg
 346         },
 347     }, {
 348         # card via api.twitter.com/1.1/videos/tweet/config
 349         'url': 'https://twitter.com/LisPower1/status/1001551623938805763',
 350         'info_dict': {
 351             'id': '1001551623938805763',
 352             'ext': 'mp4',
 353             'title': 're:.*?Shep is on a roll today.*?',
 354             'thumbnail': r're:^https?://.*\.jpg',
 355             'description': 'md5:37b9f2ff31720cef23b2bd42ee8a0f09',
 356             'uploader': 'Lis Power',
 357             'uploader_id': 'LisPower1',
 358             'duration': 111.278,
 359             'timestamp': 1527623489,
 360             'upload_date': '20180529',
 361         },
 362         'params': {
 363             'skip_download': True,  # requires ffmpeg
 364         },
 365     }, {
 366         'url': 'https://twitter.com/foobar/status/1087791357756956680',
 367         'info_dict': {
 368             'id': '1087791357756956680',
 369             'ext': 'mp4',
 370             '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!',
 371             'thumbnail': r're:^https?://.*\.jpg',
 372             'description': 'md5:6dfd341a3310fb97d80d2bf7145df976',
 373             'uploader': 'Twitter',
 374             'uploader_id': 'Twitter',
 375             'duration': 61.567,
 376             'timestamp': 1548184644,
 377             'upload_date': '20190122',
 378         },
 379     }, {
 380         # not available in Periscope
 381         'url': 'https://twitter.com/ViviEducation/status/1136534865145286656',
 382         'info_dict': {
 383             'id': '1vOGwqejwoWxB',
 384             'ext': 'mp4',
 385             'title': 'Vivi - Vivi founder @lior_rauchy announcing our new student feedback tool live at @EduTECH_AU #EduTECH2019',
 386             'uploader': 'Vivi',
 387             'uploader_id': '1eVjYOLGkGrQL',
 388         },
 389         'add_ie': ['TwitterBroadcast'],
 390     }, {
 391         # unified card
 392         'url': 'https://twitter.com/BrooklynNets/status/1349794411333394432?s=20',
 393         'info_dict': {
 394             'id': '1349794411333394432',
 395             'ext': 'mp4',
 396             'title': 'md5:d1c4941658e4caaa6cb579260d85dcba',
 397             'thumbnail': r're:^https?://.*\.jpg',
 398             'description': 'md5:71ead15ec44cee55071547d6447c6a3e',
 399             'uploader': 'Brooklyn Nets',
 400             'uploader_id': 'BrooklynNets',
 401             'duration': 324.484,
 402             'timestamp': 1610651040,
 403             'upload_date': '20210114',
 404         },
 405         'params': {
 406             'skip_download': True,
 407         },
 408     }, {
 409         # Twitch Clip Embed
 410         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 411         'only_matching': True,
 412     }, {
 413         # promo_video_website card
 414         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 415         'only_matching': True,
 416     }, {
 417         # promo_video_convo card
 418         'url': 'https://twitter.com/poco_dandy/status/1047395834013384704',
 419         'only_matching': True,
 420     }, {
 421         # appplayer card
 422         'url': 'https://twitter.com/poco_dandy/status/1150646424461176832',
 423         'only_matching': True,
 424     }, {
 425         # video_direct_message card
 426         'url': 'https://twitter.com/qarev001/status/1348948114569269251',
 427         'only_matching': True,
 428     }, {
 429         # poll2choice_video card
 430         'url': 'https://twitter.com/CAF_Online/status/1349365911120195585',
 431         'only_matching': True,
 432     }, {
 433         # poll3choice_video card
 434         'url': 'https://twitter.com/SamsungMobileSA/status/1348609186725289984',
 435         'only_matching': True,
 436     }, {
 437         # poll4choice_video card
 438         'url': 'https://twitter.com/SouthamptonFC/status/1347577658079641604',
 439         'only_matching': True,
 440     }]
 441
 442     def _real_extract(self, url):
 443         twid = self._match_id(url)
 444         status = self._call_api(
 445             'statuses/show/%s.json' % twid, twid, {
 446                 'cards_platform': 'Web-12',
 447                 'include_cards': 1,
 448                 'include_reply_count': 1,
 449                 'include_user_entities': 0,
 450                 'tweet_mode': 'extended',
 451             })
 452
 453         title = description = status['full_text'].replace('\n', ' ')
 454         # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
 455         title = re.sub(r'\s+(https?://[^ ]+)', '', title)
 456         user = status.get('user') or {}
 457         uploader = user.get('name')
 458         if uploader:
 459             title = '%s - %s' % (uploader, title)
 460         uploader_id = user.get('screen_name')
 461
 462         tags = []
 463         for hashtag in (try_get(status, lambda x: x['entities']['hashtags'], list) or []):
 464             hashtag_text = hashtag.get('text')
 465             if not hashtag_text:
 466                 continue
 467             tags.append(hashtag_text)
 468
 469         info = {
 470             'id': twid,
 471             'title': title,
 472             'description': description,
 473             'uploader': uploader,
 474             'timestamp': unified_timestamp(status.get('created_at')),
 475             'uploader_id': uploader_id,
 476             'uploader_url': format_field(uploader_id, template='https://twitter.com/%s'),
 477             'like_count': int_or_none(status.get('favorite_count')),
 478             'repost_count': int_or_none(status.get('retweet_count')),
 479             'comment_count': int_or_none(status.get('reply_count')),
 480             'age_limit': 18 if status.get('possibly_sensitive') else 0,
 481             'tags': tags,
 482         }
 483
 484         def extract_from_video_info(media):
 485             video_info = media.get('video_info') or {}
 486
 487             formats = []
 488             subtitles = {}
 489             for variant in video_info.get('variants', []):
 490                 fmts, subs = self._extract_variant_formats(variant, twid)
 491                 subtitles = self._merge_subtitles(subtitles, subs)
 492                 formats.extend(fmts)
 493             self._sort_formats(formats, ('res', 'br', 'size', 'proto'))  # The codec of http formats are unknown
 494
 495             thumbnails = []
 496             media_url = media.get('media_url_https') or media.get('media_url')
 497             if media_url:
 498                 def add_thumbnail(name, size):
 499                     thumbnails.append({
 500                         'id': name,
 501                         'url': update_url_query(media_url, {'name': name}),
 502                         'width': int_or_none(size.get('w') or size.get('width')),
 503                         'height': int_or_none(size.get('h') or size.get('height')),
 504                     })
 505                 for name, size in media.get('sizes', {}).items():
 506                     add_thumbnail(name, size)
 507                 add_thumbnail('orig', media.get('original_info') or {})
 508
 509             info.update({
 510                 'formats': formats,
 511                 'subtitles': subtitles,
 512                 'thumbnails': thumbnails,
 513                 'duration': float_or_none(video_info.get('duration_millis'), 1000),
 514             })
 515
 516         media = traverse_obj(status, ((None, 'quoted_status'), 'extended_entities', 'media', 0), get_all=False)
 517         if media and media.get('type') != 'photo':
 518             extract_from_video_info(media)
 519         else:
 520             card = status.get('card')
 521             if card:
 522                 binding_values = card['binding_values']
 523
 524                 def get_binding_value(k):
 525                     o = binding_values.get(k) or {}
 526                     return try_get(o, lambda x: x[x['type'].lower() + '_value'])
 527
 528                 card_name = card['name'].split(':')[-1]
 529                 if card_name == 'player':
 530                     info.update({
 531                         '_type': 'url',
 532                         'url': get_binding_value('player_url'),
 533                     })
 534                 elif card_name == 'periscope_broadcast':
 535                     info.update({
 536                         '_type': 'url',
 537                         'url': get_binding_value('url') or get_binding_value('player_url'),
 538                         'ie_key': PeriscopeIE.ie_key(),
 539                     })
 540                 elif card_name == 'broadcast':
 541                     info.update({
 542                         '_type': 'url',
 543                         'url': get_binding_value('broadcast_url'),
 544                         'ie_key': TwitterBroadcastIE.ie_key(),
 545                     })
 546                 elif card_name == 'summary':
 547                     info.update({
 548                         '_type': 'url',
 549                         'url': get_binding_value('card_url'),
 550                     })
 551                 elif card_name == 'unified_card':
 552                     media_entities = self._parse_json(get_binding_value('unified_card'), twid)['media_entities']
 553                     extract_from_video_info(next(iter(media_entities.values())))
 554                 # amplify, promo_video_website, promo_video_convo, appplayer,
 555                 # video_direct_message, poll2choice_video, poll3choice_video,
 556                 # poll4choice_video, ...
 557                 else:
 558                     is_amplify = card_name == 'amplify'
 559                     vmap_url = get_binding_value('amplify_url_vmap') if is_amplify else get_binding_value('player_stream_url')
 560                     content_id = get_binding_value('%s_content_id' % (card_name if is_amplify else 'player'))
 561                     formats, subtitles = self._extract_formats_from_vmap_url(vmap_url, content_id or twid)
 562                     self._sort_formats(formats)
 563
 564                     thumbnails = []
 565                     for suffix in ('_small', '', '_large', '_x_large', '_original'):
 566                         image = get_binding_value('player_image' + suffix) or {}
 567                         image_url = image.get('url')
 568                         if not image_url or '/player-placeholder' in image_url:
 569                             continue
 570                         thumbnails.append({
 571                             'id': suffix[1:] if suffix else 'medium',
 572                             'url': image_url,
 573                             'width': int_or_none(image.get('width')),
 574                             'height': int_or_none(image.get('height')),
 575                         })
 576
 577                     info.update({
 578                         'formats': formats,
 579                         'subtitles': subtitles,
 580                         'thumbnails': thumbnails,
 581                         'duration': int_or_none(get_binding_value(
 582                             'content_duration_seconds')),
 583                     })
 584             else:
 585                 expanded_url = try_get(status, lambda x: x['entities']['urls'][0]['expanded_url'])
 586                 if not expanded_url:
 587                     raise ExtractorError("There's no video in this tweet.")
 588                 info.update({
 589                     '_type': 'url',
 590                     'url': expanded_url,
 591                 })
 592         return info
 593
 594
 595 class TwitterAmplifyIE(TwitterBaseIE):
 596     IE_NAME = 'twitter:amplify'
 597     _VALID_URL = r'https?://amp\.twimg\.com/v/(?P<id>[0-9a-f\-]{36})'
 598
 599     _TEST = {
 600         'url': 'https://amp.twimg.com/v/0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 601         'md5': '7df102d0b9fd7066b86f3159f8e81bf6',
 602         'info_dict': {
 603             'id': '0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 604             'ext': 'mp4',
 605             'title': 'Twitter Video',
 606             'thumbnail': 're:^https?://.*',
 607         },
 608     }
 609
 610     def _real_extract(self, url):
 611         video_id = self._match_id(url)
 612         webpage = self._download_webpage(url, video_id)
 613
 614         vmap_url = self._html_search_meta(
 615             'twitter:amplify:vmap', webpage, 'vmap url')
 616         formats = self._extract_formats_from_vmap_url(vmap_url, video_id)
 617
 618         thumbnails = []
 619         thumbnail = self._html_search_meta(
 620             'twitter:image:src', webpage, 'thumbnail', fatal=False)
 621
 622         def _find_dimension(target):
 623             w = int_or_none(self._html_search_meta(
 624                 'twitter:%s:width' % target, webpage, fatal=False))
 625             h = int_or_none(self._html_search_meta(
 626                 'twitter:%s:height' % target, webpage, fatal=False))
 627             return w, h
 628
 629         if thumbnail:
 630             thumbnail_w, thumbnail_h = _find_dimension('image')
 631             thumbnails.append({
 632                 'url': thumbnail,
 633                 'width': thumbnail_w,
 634                 'height': thumbnail_h,
 635             })
 636
 637         video_w, video_h = _find_dimension('player')
 638         formats[0].update({
 639             'width': video_w,
 640             'height': video_h,
 641         })
 642
 643         return {
 644             'id': video_id,
 645             'title': 'Twitter Video',
 646             'formats': formats,
 647             'thumbnails': thumbnails,
 648         }
 649
 650
 651 class TwitterBroadcastIE(TwitterBaseIE, PeriscopeBaseIE):
 652     IE_NAME = 'twitter:broadcast'
 653     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/broadcasts/(?P<id>[0-9a-zA-Z]{13})'
 654
 655     _TEST = {
 656         # untitled Periscope video
 657         'url': 'https://twitter.com/i/broadcasts/1yNGaQLWpejGj',
 658         'info_dict': {
 659             'id': '1yNGaQLWpejGj',
 660             'ext': 'mp4',
 661             'title': 'Andrea May Sahouri - Periscope Broadcast',
 662             'uploader': 'Andrea May Sahouri',
 663             'uploader_id': '1PXEdBZWpGwKe',
 664         },
 665     }
 666
 667     def _real_extract(self, url):
 668         broadcast_id = self._match_id(url)
 669         broadcast = self._call_api(
 670             'broadcasts/show.json', broadcast_id,
 671             {'ids': broadcast_id})['broadcasts'][broadcast_id]
 672         info = self._parse_broadcast_data(broadcast, broadcast_id)
 673         media_key = broadcast['media_key']
 674         source = self._call_api(
 675             'live_video_stream/status/' + media_key, media_key)['source']
 676         m3u8_url = source.get('noRedirectPlaybackUrl') or source['location']
 677         if '/live_video_stream/geoblocked/' in m3u8_url:
 678             self.raise_geo_restricted()
 679         m3u8_id = compat_parse_qs(compat_urllib_parse_urlparse(
 680             m3u8_url).query).get('type', [None])[0]
 681         state, width, height = self._extract_common_format_info(broadcast)
 682         info['formats'] = self._extract_pscp_m3u8_formats(
 683             m3u8_url, broadcast_id, m3u8_id, state, width, height)
 684         return info
 685
 686
 687 class TwitterShortenerIE(TwitterBaseIE):
 688     IE_NAME = 'twitter:shortener'
 689     _VALID_URL = r'https?://t.co/(?P<id>[^?]+)|tco:(?P<eid>[^?]+)'
 690     _BASE_URL = 'https://t.co/'
 691
 692     def _real_extract(self, url):
 693         mobj = self._match_valid_url(url)
 694         eid, id = mobj.group('eid', 'id')
 695         if eid:
 696             id = eid
 697             url = self._BASE_URL + id
 698         new_url = self._request_webpage(url, id, headers={'User-Agent': 'curl'}).geturl()
 699         __UNSAFE_LINK = "https://twitter.com/safety/unsafe_link_warning?unsafe_link="
 700         if new_url.startswith(__UNSAFE_LINK):
 701             new_url = new_url.replace(__UNSAFE_LINK, "")
 702         return self.url_result(new_url)