youtube_dlc/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     try_get,
  19     strip_or_none,
  20     unified_timestamp,
  21     update_url_query,
  22     xpath_text,
  23 )
  24
  25 from .periscope import (
  26     PeriscopeBaseIE,
  27     PeriscopeIE,
  28 )
  29
  30
  31 class TwitterBaseIE(InfoExtractor):
  32     _API_BASE = 'https://api.twitter.com/1.1/'
  33     _BASE_REGEX = r'https?://(?:(?:www|m(?:obile)?)\.)?twitter\.com/'
  34     _GUEST_TOKEN = None
  35
  36     def _extract_variant_formats(self, variant, video_id):
  37         variant_url = variant.get('url')
  38         if not variant_url:
  39             return []
  40         elif '.m3u8' in variant_url:
  41             return self._extract_m3u8_formats(
  42                 variant_url, video_id, 'mp4', 'm3u8_native',
  43                 m3u8_id='hls', fatal=False)
  44         else:
  45             tbr = int_or_none(dict_get(variant, ('bitrate', 'bit_rate')), 1000) or None
  46             f = {
  47                 'url': variant_url,
  48                 'format_id': 'http' + ('-%d' % tbr if tbr else ''),
  49                 'tbr': tbr,
  50             }
  51             self._search_dimensions_in_video_url(f, variant_url)
  52             return [f]
  53
  54     def _extract_formats_from_vmap_url(self, vmap_url, video_id):
  55         vmap_data = self._download_xml(vmap_url, video_id)
  56         formats = []
  57         urls = []
  58         for video_variant in vmap_data.findall('.//{http://twitter.com/schema/videoVMapV2.xsd}videoVariant'):
  59             video_variant.attrib['url'] = compat_urllib_parse_unquote(
  60                 video_variant.attrib['url'])
  61             urls.append(video_variant.attrib['url'])
  62             formats.extend(self._extract_variant_formats(
  63                 video_variant.attrib, video_id))
  64         video_url = strip_or_none(xpath_text(vmap_data, './/MediaFile'))
  65         if video_url not in urls:
  66             formats.extend(self._extract_variant_formats({'url': video_url}, video_id))
  67         return formats
  68
  69     @staticmethod
  70     def _search_dimensions_in_video_url(a_format, video_url):
  71         m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
  72         if m:
  73             a_format.update({
  74                 'width': int(m.group('width')),
  75                 'height': int(m.group('height')),
  76             })
  77
  78     def _call_api(self, path, video_id, query={}):
  79         headers = {
  80             'Authorization': 'Bearer AAAAAAAAAAAAAAAAAAAAAPYXBAAAAAAACLXUNDekMxqa8h%2F40K4moUkGsoc%3DTYfbDKbT3jJPCEVnMYqilB28NHfOPqkca3qaAxGfsyKCs0wRbw',
  81         }
  82         if not self._GUEST_TOKEN:
  83             self._GUEST_TOKEN = self._download_json(
  84                 self._API_BASE + 'guest/activate.json', video_id,
  85                 'Downloading guest token', data=b'',
  86                 headers=headers)['guest_token']
  87         headers['x-guest-token'] = self._GUEST_TOKEN
  88         try:
  89             return self._download_json(
  90                 self._API_BASE + path, video_id, headers=headers, query=query)
  91         except ExtractorError as e:
  92             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
  93                 raise ExtractorError(self._parse_json(
  94                     e.cause.read().decode(),
  95                     video_id)['errors'][0]['message'], expected=True)
  96             raise
  97
  98
  99 class TwitterCardIE(InfoExtractor):
 100     IE_NAME = 'twitter:card'
 101     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/(?:cards/tfw/v1|videos(?:/tweet)?)/(?P<id>\d+)'
 102     _TESTS = [
 103         {
 104             'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
 105             # MD5 checksums are different in different places
 106             'info_dict': {
 107                 'id': '560070183650213889',
 108                 'ext': 'mp4',
 109                 'title': "Twitter - You can now shoot, edit and share video on Twitter. Capture life's most moving moments from your perspective.",
 110                 'description': 'md5:18d3e24bb4f6e5007487dd546e53bd96',
 111                 'uploader': 'Twitter',
 112                 'uploader_id': 'Twitter',
 113                 'thumbnail': r're:^https?://.*\.jpg',
 114                 'duration': 30.033,
 115                 'timestamp': 1422366112,
 116                 'upload_date': '20150127',
 117             },
 118         },
 119         {
 120             'url': 'https://twitter.com/i/cards/tfw/v1/623160978427936768',
 121             'md5': '7137eca597f72b9abbe61e5ae0161399',
 122             'info_dict': {
 123                 'id': '623160978427936768',
 124                 'ext': 'mp4',
 125                 'title': "NASA - Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video.",
 126                 'description': "Fly over Pluto's icy Norgay Mountains and Sputnik Plain in this @NASANewHorizons #PlutoFlyby video. https://t.co/BJYgOjSeGA",
 127                 'uploader': 'NASA',
 128                 'uploader_id': 'NASA',
 129                 'timestamp': 1437408129,
 130                 'upload_date': '20150720',
 131             },
 132         },
 133         {
 134             'url': 'https://twitter.com/i/cards/tfw/v1/654001591733886977',
 135             'md5': 'b6d9683dd3f48e340ded81c0e917ad46',
 136             'info_dict': {
 137                 'id': 'dq4Oj5quskI',
 138                 'ext': 'mp4',
 139                 'title': 'Ubuntu 11.10 Overview',
 140                 'description': 'md5:a831e97fa384863d6e26ce48d1c43376',
 141                 'upload_date': '20111013',
 142                 'uploader': 'OMG! UBUNTU!',
 143                 'uploader_id': 'omgubuntu',
 144             },
 145             'add_ie': ['Youtube'],
 146         },
 147         {
 148             'url': 'https://twitter.com/i/cards/tfw/v1/665289828897005568',
 149             'md5': '6dabeaca9e68cbb71c99c322a4b42a11',
 150             'info_dict': {
 151                 'id': 'iBb2x00UVlv',
 152                 'ext': 'mp4',
 153                 'upload_date': '20151113',
 154                 'uploader_id': '1189339351084113920',
 155                 'uploader': 'ArsenalTerje',
 156                 'title': 'Vine by ArsenalTerje',
 157                 'timestamp': 1447451307,
 158             },
 159             'add_ie': ['Vine'],
 160         }, {
 161             'url': 'https://twitter.com/i/videos/tweet/705235433198714880',
 162             'md5': '884812a2adc8aaf6fe52b15ccbfa3b88',
 163             'info_dict': {
 164                 'id': '705235433198714880',
 165                 'ext': 'mp4',
 166                 'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 167                 'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 168                 'uploader': 'Brent Yarina',
 169                 'uploader_id': 'BTNBrentYarina',
 170                 'timestamp': 1456976204,
 171                 'upload_date': '20160303',
 172             },
 173             'skip': 'This content is no longer available.',
 174         }, {
 175             'url': 'https://twitter.com/i/videos/752274308186120192',
 176             'only_matching': True,
 177         },
 178     ]
 179
 180     def _real_extract(self, url):
 181         status_id = self._match_id(url)
 182         return self.url_result(
 183             'https://twitter.com/statuses/' + status_id,
 184             TwitterIE.ie_key(), status_id)
 185
 186
 187 class TwitterIE(TwitterBaseIE):
 188     IE_NAME = 'twitter'
 189     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'(?:(?:i/web|[^/]+)/status|statuses)/(?P<id>\d+)'
 190
 191     _TESTS = [{
 192         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
 193         'info_dict': {
 194             'id': '643211948184596480',
 195             'ext': 'mp4',
 196             'title': 'FREE THE NIPPLE - FTN supporters on Hollywood Blvd today!',
 197             'thumbnail': r're:^https?://.*\.jpg',
 198             'description': 'FTN supporters on Hollywood Blvd today! http://t.co/c7jHH749xJ',
 199             'uploader': 'FREE THE NIPPLE',
 200             'uploader_id': 'freethenipple',
 201             'duration': 12.922,
 202             'timestamp': 1442188653,
 203             'upload_date': '20150913',
 204             'age_limit': 18,
 205         },
 206     }, {
 207         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
 208         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
 209         'info_dict': {
 210             'id': '657991469417025536',
 211             'ext': 'mp4',
 212             'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
 213             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
 214             'thumbnail': r're:^https?://.*\.png',
 215             'uploader': 'Gifs',
 216             'uploader_id': 'giphz',
 217         },
 218         'expected_warnings': ['height', 'width'],
 219         'skip': 'Account suspended',
 220     }, {
 221         'url': 'https://twitter.com/starwars/status/665052190608723968',
 222         'info_dict': {
 223             'id': '665052190608723968',
 224             'ext': 'mp4',
 225             'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
 226             'description': 'A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens. https://t.co/OkSqT2fjWJ',
 227             'uploader_id': 'starwars',
 228             'uploader': 'Star Wars',
 229             'timestamp': 1447395772,
 230             'upload_date': '20151113',
 231         },
 232     }, {
 233         'url': 'https://twitter.com/BTNBrentYarina/status/705235433198714880',
 234         'info_dict': {
 235             'id': '705235433198714880',
 236             'ext': 'mp4',
 237             'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
 238             'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
 239             'uploader_id': 'BTNBrentYarina',
 240             'uploader': 'Brent Yarina',
 241             'timestamp': 1456976204,
 242             'upload_date': '20160303',
 243         },
 244         'params': {
 245             # The same video as https://twitter.com/i/videos/tweet/705235433198714880
 246             # Test case of TwitterCardIE
 247             'skip_download': True,
 248         },
 249     }, {
 250         'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
 251         'info_dict': {
 252             'id': '700207533655363584',
 253             'ext': 'mp4',
 254             'title': 'simon vertugo - BEAT PROD: @suhmeduh #Damndaniel',
 255             'description': 'BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ',
 256             'thumbnail': r're:^https?://.*\.jpg',
 257             'uploader': 'simon vertugo',
 258             'uploader_id': 'simonvertugo',
 259             'duration': 30.0,
 260             'timestamp': 1455777459,
 261             'upload_date': '20160218',
 262         },
 263     }, {
 264         'url': 'https://twitter.com/Filmdrunk/status/713801302971588609',
 265         'md5': '89a15ed345d13b86e9a5a5e051fa308a',
 266         'info_dict': {
 267             'id': 'MIOxnrUteUd',
 268             'ext': 'mp4',
 269             'title': 'Dr.Pepperの飲み方 #japanese #バカ #ドクペ #電動ガン',
 270             'uploader': 'TAKUMA',
 271             'uploader_id': '1004126642786242560',
 272             'timestamp': 1402826626,
 273             'upload_date': '20140615',
 274         },
 275         'add_ie': ['Vine'],
 276     }, {
 277         'url': 'https://twitter.com/captainamerica/status/719944021058060289',
 278         'info_dict': {
 279             'id': '719944021058060289',
 280             'ext': 'mp4',
 281             'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
 282             'description': '@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI',
 283             'uploader_id': 'CaptainAmerica',
 284             'uploader': 'Captain America',
 285             'duration': 3.17,
 286             'timestamp': 1460483005,
 287             'upload_date': '20160412',
 288         },
 289     }, {
 290         'url': 'https://twitter.com/OPP_HSD/status/779210622571536384',
 291         'info_dict': {
 292             'id': '1zqKVVlkqLaKB',
 293             'ext': 'mp4',
 294             'title': 'Sgt Kerry Schmidt - Ontario Provincial Police - Road rage, mischief, assault, rollover and fire in one occurrence',
 295             'upload_date': '20160923',
 296             'uploader_id': '1PmKqpJdOJQoY',
 297             'uploader': 'Sgt Kerry Schmidt - Ontario Provincial Police',
 298             'timestamp': 1474613214,
 299         },
 300         'add_ie': ['Periscope'],
 301     }, {
 302         # has mp4 formats via mobile API
 303         'url': 'https://twitter.com/news_al3alm/status/852138619213144067',
 304         'info_dict': {
 305             'id': '852138619213144067',
 306             'ext': 'mp4',
 307             'title': 'عالم الأخبار - كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة',
 308             'description': 'كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة   https://t.co/xg6OhpyKfN',
 309             'uploader': 'عالم الأخبار',
 310             'uploader_id': 'news_al3alm',
 311             'duration': 277.4,
 312             'timestamp': 1492000653,
 313             'upload_date': '20170412',
 314         },
 315         'skip': 'Account suspended',
 316     }, {
 317         'url': 'https://twitter.com/i/web/status/910031516746514432',
 318         'info_dict': {
 319             'id': '910031516746514432',
 320             'ext': 'mp4',
 321             '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.',
 322             'thumbnail': r're:^https?://.*\.jpg',
 323             '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',
 324             'uploader': 'Préfet de Guadeloupe',
 325             'uploader_id': 'Prefet971',
 326             'duration': 47.48,
 327             'timestamp': 1505803395,
 328             'upload_date': '20170919',
 329         },
 330         'params': {
 331             'skip_download': True,  # requires ffmpeg
 332         },
 333     }, {
 334         # card via api.twitter.com/1.1/videos/tweet/config
 335         'url': 'https://twitter.com/LisPower1/status/1001551623938805763',
 336         'info_dict': {
 337             'id': '1001551623938805763',
 338             'ext': 'mp4',
 339             'title': 're:.*?Shep is on a roll today.*?',
 340             'thumbnail': r're:^https?://.*\.jpg',
 341             'description': 'md5:37b9f2ff31720cef23b2bd42ee8a0f09',
 342             'uploader': 'Lis Power',
 343             'uploader_id': 'LisPower1',
 344             'duration': 111.278,
 345             'timestamp': 1527623489,
 346             'upload_date': '20180529',
 347         },
 348         'params': {
 349             'skip_download': True,  # requires ffmpeg
 350         },
 351     }, {
 352         'url': 'https://twitter.com/foobar/status/1087791357756956680',
 353         'info_dict': {
 354             'id': '1087791357756956680',
 355             'ext': 'mp4',
 356             '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!',
 357             'thumbnail': r're:^https?://.*\.jpg',
 358             'description': 'md5:6dfd341a3310fb97d80d2bf7145df976',
 359             'uploader': 'Twitter',
 360             'uploader_id': 'Twitter',
 361             'duration': 61.567,
 362             'timestamp': 1548184644,
 363             'upload_date': '20190122',
 364         },
 365     }, {
 366         # not available in Periscope
 367         'url': 'https://twitter.com/ViviEducation/status/1136534865145286656',
 368         'info_dict': {
 369             'id': '1vOGwqejwoWxB',
 370             'ext': 'mp4',
 371             'title': 'Vivi - Vivi founder @lior_rauchy announcing our new student feedback tool live at @EduTECH_AU #EduTECH2019',
 372             'uploader': 'Vivi',
 373             'uploader_id': '1eVjYOLGkGrQL',
 374         },
 375         'add_ie': ['TwitterBroadcast'],
 376     }, {
 377         # Twitch Clip Embed
 378         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 379         'only_matching': True,
 380     }, {
 381         # promo_video_website card
 382         'url': 'https://twitter.com/GunB1g/status/1163218564784017422',
 383         'only_matching': True,
 384     }, {
 385         # promo_video_convo card
 386         'url': 'https://twitter.com/poco_dandy/status/1047395834013384704',
 387         'only_matching': True,
 388     }, {
 389         # appplayer card
 390         'url': 'https://twitter.com/poco_dandy/status/1150646424461176832',
 391         'only_matching': True,
 392     }]
 393
 394     def _real_extract(self, url):
 395         twid = self._match_id(url)
 396         status = self._call_api(
 397             'statuses/show/%s.json' % twid, twid, {
 398                 'cards_platform': 'Web-12',
 399                 'include_cards': 1,
 400                 'include_reply_count': 1,
 401                 'include_user_entities': 0,
 402                 'tweet_mode': 'extended',
 403             })
 404
 405         title = description = status['full_text'].replace('\n', ' ')
 406         # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
 407         title = re.sub(r'\s+(https?://[^ ]+)', '', title)
 408         user = status.get('user') or {}
 409         uploader = user.get('name')
 410         if uploader:
 411             title = '%s - %s' % (uploader, title)
 412         uploader_id = user.get('screen_name')
 413
 414         tags = []
 415         for hashtag in (try_get(status, lambda x: x['entities']['hashtags'], list) or []):
 416             hashtag_text = hashtag.get('text')
 417             if not hashtag_text:
 418                 continue
 419             tags.append(hashtag_text)
 420
 421         info = {
 422             'id': twid,
 423             'title': title,
 424             'description': description,
 425             'uploader': uploader,
 426             'timestamp': unified_timestamp(status.get('created_at')),
 427             'uploader_id': uploader_id,
 428             'uploader_url': 'https://twitter.com/' + uploader_id if uploader_id else None,
 429             'like_count': int_or_none(status.get('favorite_count')),
 430             'repost_count': int_or_none(status.get('retweet_count')),
 431             'comment_count': int_or_none(status.get('reply_count')),
 432             'age_limit': 18 if status.get('possibly_sensitive') else 0,
 433             'tags': tags,
 434         }
 435
 436         media = try_get(status, lambda x: x['extended_entities']['media'][0])
 437         if media and media.get('type') != 'photo':
 438             video_info = media.get('video_info') or {}
 439
 440             formats = []
 441             for variant in video_info.get('variants', []):
 442                 formats.extend(self._extract_variant_formats(variant, twid))
 443             self._sort_formats(formats)
 444
 445             thumbnails = []
 446             media_url = media.get('media_url_https') or media.get('media_url')
 447             if media_url:
 448                 def add_thumbnail(name, size):
 449                     thumbnails.append({
 450                         'id': name,
 451                         'url': update_url_query(media_url, {'name': name}),
 452                         'width': int_or_none(size.get('w') or size.get('width')),
 453                         'height': int_or_none(size.get('h') or size.get('height')),
 454                     })
 455                 for name, size in media.get('sizes', {}).items():
 456                     add_thumbnail(name, size)
 457                 add_thumbnail('orig', media.get('original_info') or {})
 458
 459             info.update({
 460                 'formats': formats,
 461                 'thumbnails': thumbnails,
 462                 'duration': float_or_none(video_info.get('duration_millis'), 1000),
 463             })
 464         else:
 465             card = status.get('card')
 466             if card:
 467                 binding_values = card['binding_values']
 468
 469                 def get_binding_value(k):
 470                     o = binding_values.get(k) or {}
 471                     return try_get(o, lambda x: x[x['type'].lower() + '_value'])
 472
 473                 card_name = card['name'].split(':')[-1]
 474                 if card_name == 'player':
 475                     info.update({
 476                         '_type': 'url',
 477                         'url': get_binding_value('player_url'),
 478                     })
 479                 elif card_name == 'periscope_broadcast':
 480                     info.update({
 481                         '_type': 'url',
 482                         'url': get_binding_value('url') or get_binding_value('player_url'),
 483                         'ie_key': PeriscopeIE.ie_key(),
 484                     })
 485                 elif card_name == 'broadcast':
 486                     info.update({
 487                         '_type': 'url',
 488                         'url': get_binding_value('broadcast_url'),
 489                         'ie_key': TwitterBroadcastIE.ie_key(),
 490                     })
 491                 elif card_name == 'summary':
 492                     info.update({
 493                         '_type': 'url',
 494                         'url': get_binding_value('card_url'),
 495                     })
 496                 # amplify, promo_video_website, promo_video_convo, appplayer, ...
 497                 else:
 498                     is_amplify = card_name == 'amplify'
 499                     vmap_url = get_binding_value('amplify_url_vmap') if is_amplify else get_binding_value('player_stream_url')
 500                     content_id = get_binding_value('%s_content_id' % (card_name if is_amplify else 'player'))
 501                     formats = self._extract_formats_from_vmap_url(vmap_url, content_id or twid)
 502                     self._sort_formats(formats)
 503
 504                     thumbnails = []
 505                     for suffix in ('_small', '', '_large', '_x_large', '_original'):
 506                         image = get_binding_value('player_image' + suffix) or {}
 507                         image_url = image.get('url')
 508                         if not image_url or '/player-placeholder' in image_url:
 509                             continue
 510                         thumbnails.append({
 511                             'id': suffix[1:] if suffix else 'medium',
 512                             'url': image_url,
 513                             'width': int_or_none(image.get('width')),
 514                             'height': int_or_none(image.get('height')),
 515                         })
 516
 517                     info.update({
 518                         'formats': formats,
 519                         'thumbnails': thumbnails,
 520                         'duration': int_or_none(get_binding_value(
 521                             'content_duration_seconds')),
 522                     })
 523             else:
 524                 expanded_url = try_get(status, lambda x: x['entities']['urls'][0]['expanded_url'])
 525                 if not expanded_url:
 526                     raise ExtractorError("There's no video in this tweet.")
 527                 info.update({
 528                     '_type': 'url',
 529                     'url': expanded_url,
 530                 })
 531         return info
 532
 533
 534 class TwitterAmplifyIE(TwitterBaseIE):
 535     IE_NAME = 'twitter:amplify'
 536     _VALID_URL = r'https?://amp\.twimg\.com/v/(?P<id>[0-9a-f\-]{36})'
 537
 538     _TEST = {
 539         'url': 'https://amp.twimg.com/v/0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 540         'md5': '7df102d0b9fd7066b86f3159f8e81bf6',
 541         'info_dict': {
 542             'id': '0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 543             'ext': 'mp4',
 544             'title': 'Twitter Video',
 545             'thumbnail': 're:^https?://.*',
 546         },
 547     }
 548
 549     def _real_extract(self, url):
 550         video_id = self._match_id(url)
 551         webpage = self._download_webpage(url, video_id)
 552
 553         vmap_url = self._html_search_meta(
 554             'twitter:amplify:vmap', webpage, 'vmap url')
 555         formats = self._extract_formats_from_vmap_url(vmap_url, video_id)
 556
 557         thumbnails = []
 558         thumbnail = self._html_search_meta(
 559             'twitter:image:src', webpage, 'thumbnail', fatal=False)
 560
 561         def _find_dimension(target):
 562             w = int_or_none(self._html_search_meta(
 563                 'twitter:%s:width' % target, webpage, fatal=False))
 564             h = int_or_none(self._html_search_meta(
 565                 'twitter:%s:height' % target, webpage, fatal=False))
 566             return w, h
 567
 568         if thumbnail:
 569             thumbnail_w, thumbnail_h = _find_dimension('image')
 570             thumbnails.append({
 571                 'url': thumbnail,
 572                 'width': thumbnail_w,
 573                 'height': thumbnail_h,
 574             })
 575
 576         video_w, video_h = _find_dimension('player')
 577         formats[0].update({
 578             'width': video_w,
 579             'height': video_h,
 580         })
 581
 582         return {
 583             'id': video_id,
 584             'title': 'Twitter Video',
 585             'formats': formats,
 586             'thumbnails': thumbnails,
 587         }
 588
 589
 590 class TwitterBroadcastIE(TwitterBaseIE, PeriscopeBaseIE):
 591     IE_NAME = 'twitter:broadcast'
 592     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'i/broadcasts/(?P<id>[0-9a-zA-Z]{13})'
 593
 594     _TEST = {
 595         # untitled Periscope video
 596         'url': 'https://twitter.com/i/broadcasts/1yNGaQLWpejGj',
 597         'info_dict': {
 598             'id': '1yNGaQLWpejGj',
 599             'ext': 'mp4',
 600             'title': 'Andrea May Sahouri - Periscope Broadcast',
 601             'uploader': 'Andrea May Sahouri',
 602             'uploader_id': '1PXEdBZWpGwKe',
 603         },
 604     }
 605
 606     def _real_extract(self, url):
 607         broadcast_id = self._match_id(url)
 608         broadcast = self._call_api(
 609             'broadcasts/show.json', broadcast_id,
 610             {'ids': broadcast_id})['broadcasts'][broadcast_id]
 611         info = self._parse_broadcast_data(broadcast, broadcast_id)
 612         media_key = broadcast['media_key']
 613         source = self._call_api(
 614             'live_video_stream/status/' + media_key, media_key)['source']
 615         m3u8_url = source.get('noRedirectPlaybackUrl') or source['location']
 616         if '/live_video_stream/geoblocked/' in m3u8_url:
 617             self.raise_geo_restricted()
 618         m3u8_id = compat_parse_qs(compat_urllib_parse_urlparse(
 619             m3u8_url).query).get('type', [None])[0]
 620         state, width, height = self._extract_common_format_info(broadcast)
 621         info['formats'] = self._extract_pscp_m3u8_formats(
 622             m3u8_url, broadcast_id, m3u8_id, state, width, height)
 623         return info
 624
 625
 626 class TwitterShortenerIE(TwitterBaseIE):
 627     IE_NAME = 'twitter:shortener'
 628     _VALID_URL = r'https?://t.co/(?P<id>[^?]+)|tco:(?P<eid>[^?]+)'
 629     _BASE_URL = 'https://t.co/'
 630
 631     def _real_extract(self, url):
 632         mobj = re.match(self._VALID_URL, url)
 633         eid, id = mobj.group('eid', 'id')
 634         if eid:
 635             id = eid
 636             url = self._BASE_URL + id
 637         new_url = self._request_webpage(url, id, headers={'User-Agent': 'curl'}).geturl()
 638         __UNSAFE_LINK = "https://twitter.com/safety/unsafe_link_warning?unsafe_link="
 639         if new_url.startswith(__UNSAFE_LINK):
 640             new_url = new_url.replace(__UNSAFE_LINK, "")
 641         return self.url_result(new_url)