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

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..compat import compat_urlparse
   8 from ..utils import (
   9     determine_ext,
  10     dict_get,
  11     ExtractorError,
  12     float_or_none,
  13     int_or_none,
  14     remove_end,
  15     try_get,
  16     xpath_text,
  17 )
  18
  19 from .periscope import PeriscopeIE
  20
  21
  22 class TwitterBaseIE(InfoExtractor):
  23     def _extract_formats_from_vmap_url(self, vmap_url, video_id):
  24         vmap_data = self._download_xml(vmap_url, video_id)
  25         video_url = xpath_text(vmap_data, './/MediaFile').strip()
  26         if determine_ext(video_url) == 'm3u8':
  27             return self._extract_m3u8_formats(
  28                 video_url, video_id, ext='mp4', m3u8_id='hls',
  29                 entry_protocol='m3u8_native')
  30         return [{
  31             'url': video_url,
  32         }]
  33
  34     @staticmethod
  35     def _search_dimensions_in_video_url(a_format, video_url):
  36         m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
  37         if m:
  38             a_format.update({
  39                 'width': int(m.group('width')),
  40                 'height': int(m.group('height')),
  41             })
  42
  43
  44 class TwitterCardIE(TwitterBaseIE):
  45     IE_NAME = 'twitter:card'
  46     _VALID_URL = r'https?://(?:www\.)?twitter\.com/i/(?:cards/tfw/v1|videos(?:/tweet)?)/(?P<id>\d+)'
  47     _TESTS = [
  48         {
  49             'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
  50             # MD5 checksums are different in different places
  51             'info_dict': {
  52                 'id': '560070183650213889',
  53                 'ext': 'mp4',
  54                 'title': 'Twitter Card',
  55                 'thumbnail': r're:^https?://.*\.jpg$',
  56                 'duration': 30.033,
  57             },
  58             'skip': 'Video gone',
  59         },
  60         {
  61             'url': 'https://twitter.com/i/cards/tfw/v1/623160978427936768',
  62             'md5': '7ee2a553b63d1bccba97fbed97d9e1c8',
  63             'info_dict': {
  64                 'id': '623160978427936768',
  65                 'ext': 'mp4',
  66                 'title': 'Twitter Card',
  67                 'thumbnail': r're:^https?://.*\.jpg',
  68                 'duration': 80.155,
  69             },
  70             'skip': 'Video gone',
  71         },
  72         {
  73             'url': 'https://twitter.com/i/cards/tfw/v1/654001591733886977',
  74             'md5': 'b6d9683dd3f48e340ded81c0e917ad46',
  75             'info_dict': {
  76                 'id': 'dq4Oj5quskI',
  77                 'ext': 'mp4',
  78                 'title': 'Ubuntu 11.10 Overview',
  79                 'description': 'md5:a831e97fa384863d6e26ce48d1c43376',
  80                 'upload_date': '20111013',
  81                 'uploader': 'OMG! Ubuntu!',
  82                 'uploader_id': 'omgubuntu',
  83             },
  84             'add_ie': ['Youtube'],
  85         },
  86         {
  87             'url': 'https://twitter.com/i/cards/tfw/v1/665289828897005568',
  88             'md5': '6dabeaca9e68cbb71c99c322a4b42a11',
  89             'info_dict': {
  90                 'id': 'iBb2x00UVlv',
  91                 'ext': 'mp4',
  92                 'upload_date': '20151113',
  93                 'uploader_id': '1189339351084113920',
  94                 'uploader': 'ArsenalTerje',
  95                 'title': 'Vine by ArsenalTerje',
  96                 'timestamp': 1447451307,
  97             },
  98             'add_ie': ['Vine'],
  99         }, {
 100             'url': 'https://twitter.com/i/videos/tweet/705235433198714880',
 101             'md5': '884812a2adc8aaf6fe52b15ccbfa3b88',
 102             'info_dict': {
 103                 'id': '705235433198714880',
 104                 'ext': 'mp4',
 105                 'title': 'Twitter web player',
 106                 'thumbnail': r're:^https?://.*',
 107             },
 108         }, {
 109             'url': 'https://twitter.com/i/videos/752274308186120192',
 110             'only_matching': True,
 111         },
 112     ]
 113
 114     def _parse_media_info(self, media_info, video_id):
 115         formats = []
 116         for media_variant in media_info.get('variants', []):
 117             media_url = media_variant['url']
 118             if media_url.endswith('.m3u8'):
 119                 formats.extend(self._extract_m3u8_formats(media_url, video_id, ext='mp4', m3u8_id='hls'))
 120             elif media_url.endswith('.mpd'):
 121                 formats.extend(self._extract_mpd_formats(media_url, video_id, mpd_id='dash'))
 122             else:
 123                 vbr = int_or_none(dict_get(media_variant, ('bitRate', 'bitrate')), scale=1000)
 124                 a_format = {
 125                     'url': media_url,
 126                     'format_id': 'http-%d' % vbr if vbr else 'http',
 127                     'vbr': vbr,
 128                 }
 129                 # Reported bitRate may be zero
 130                 if not a_format['vbr']:
 131                     del a_format['vbr']
 132
 133                 self._search_dimensions_in_video_url(a_format, media_url)
 134
 135                 formats.append(a_format)
 136         return formats
 137
 138     def _extract_mobile_formats(self, username, video_id):
 139         webpage = self._download_webpage(
 140             'https://mobile.twitter.com/%s/status/%s' % (username, video_id),
 141             video_id, 'Downloading mobile webpage',
 142             headers={
 143                 # A recent mobile UA is necessary for `gt` cookie
 144                 'User-Agent': 'Mozilla/5.0 (Android 6.0.1; Mobile; rv:54.0) Gecko/54.0 Firefox/54.0',
 145             })
 146         main_script_url = self._html_search_regex(
 147             r'<script[^>]+src="([^"]+main\.[^"]+)"', webpage, 'main script URL')
 148         main_script = self._download_webpage(
 149             main_script_url, video_id, 'Downloading main script')
 150         bearer_token = self._search_regex(
 151             r'BEARER_TOKEN\s*:\s*"([^"]+)"',
 152             main_script, 'bearer token')
 153         guest_token = self._search_regex(
 154             r'document\.cookie\s*=\s*decodeURIComponent\("gt=(\d+)',
 155             webpage, 'guest token')
 156         api_data = self._download_json(
 157             'https://api.twitter.com/2/timeline/conversation/%s.json' % video_id,
 158             video_id, 'Downloading mobile API data',
 159             headers={
 160                 'Authorization': 'Bearer ' + bearer_token,
 161                 'x-guest-token': guest_token,
 162             })
 163         media_info = try_get(api_data, lambda o: o['globalObjects']['tweets'][video_id]
 164                                                   ['extended_entities']['media'][0]['video_info']) or {}
 165         return self._parse_media_info(media_info, video_id)
 166
 167     def _real_extract(self, url):
 168         video_id = self._match_id(url)
 169
 170         config = None
 171         formats = []
 172         duration = None
 173
 174         webpage = self._download_webpage(url, video_id)
 175
 176         iframe_url = self._html_search_regex(
 177             r'<iframe[^>]+src="((?:https?:)?//(?:www\.youtube\.com/embed/[^"]+|(?:www\.)?vine\.co/v/\w+/card))"',
 178             webpage, 'video iframe', default=None)
 179         if iframe_url:
 180             return self.url_result(iframe_url)
 181
 182         config = self._parse_json(self._html_search_regex(
 183             r'data-(?:player-)?config="([^"]+)"', webpage,
 184             'data player config', default='{}'),
 185             video_id)
 186
 187         if config.get('source_type') == 'vine':
 188             return self.url_result(config['player_url'], 'Vine')
 189
 190         periscope_url = PeriscopeIE._extract_url(webpage)
 191         if periscope_url:
 192             return self.url_result(periscope_url, PeriscopeIE.ie_key())
 193
 194         video_url = config.get('video_url') or config.get('playlist', [{}])[0].get('source')
 195
 196         if video_url:
 197             if determine_ext(video_url) == 'm3u8':
 198                 formats.extend(self._extract_m3u8_formats(video_url, video_id, ext='mp4', m3u8_id='hls'))
 199             else:
 200                 f = {
 201                     'url': video_url,
 202                 }
 203
 204                 self._search_dimensions_in_video_url(f, video_url)
 205
 206                 formats.append(f)
 207
 208         vmap_url = config.get('vmapUrl') or config.get('vmap_url')
 209         if vmap_url:
 210             formats.extend(
 211                 self._extract_formats_from_vmap_url(vmap_url, video_id))
 212
 213         media_info = None
 214
 215         for entity in config.get('status', {}).get('entities', []):
 216             if 'mediaInfo' in entity:
 217                 media_info = entity['mediaInfo']
 218
 219         if media_info:
 220             formats.extend(self._parse_media_info(media_info, video_id))
 221             duration = float_or_none(media_info.get('duration', {}).get('nanos'), scale=1e9)
 222
 223         username = config.get('user', {}).get('screen_name')
 224         if username:
 225             formats.extend(self._extract_mobile_formats(username, video_id))
 226
 227         self._remove_duplicate_formats(formats)
 228         self._sort_formats(formats)
 229
 230         title = self._search_regex(r'<title>([^<]+)</title>', webpage, 'title')
 231         thumbnail = config.get('posterImageUrl') or config.get('image_src')
 232         duration = float_or_none(config.get('duration'), scale=1000) or duration
 233
 234         return {
 235             'id': video_id,
 236             'title': title,
 237             'thumbnail': thumbnail,
 238             'duration': duration,
 239             'formats': formats,
 240         }
 241
 242
 243 class TwitterIE(InfoExtractor):
 244     IE_NAME = 'twitter'
 245     _VALID_URL = r'https?://(?:www\.|m\.|mobile\.)?twitter\.com/(?:i/web|(?P<user_id>[^/]+))/status/(?P<id>\d+)'
 246     _TEMPLATE_URL = 'https://twitter.com/%s/status/%s'
 247     _TEMPLATE_STATUSES_URL = 'https://twitter.com/statuses/%s'
 248
 249     _TESTS = [{
 250         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
 251         'info_dict': {
 252             'id': '643211948184596480',
 253             'ext': 'mp4',
 254             'title': 'FREE THE NIPPLE - FTN supporters on Hollywood Blvd today!',
 255             'thumbnail': r're:^https?://.*\.jpg',
 256             'description': 'FREE THE NIPPLE on Twitter: "FTN supporters on Hollywood Blvd today! http://t.co/c7jHH749xJ"',
 257             'uploader': 'FREE THE NIPPLE',
 258             'uploader_id': 'freethenipple',
 259             'duration': 12.922,
 260         },
 261         'params': {
 262             'skip_download': True,  # requires ffmpeg
 263         },
 264     }, {
 265         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
 266         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
 267         'info_dict': {
 268             'id': '657991469417025536',
 269             'ext': 'mp4',
 270             'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
 271             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
 272             'thumbnail': r're:^https?://.*\.png',
 273             'uploader': 'Gifs',
 274             'uploader_id': 'giphz',
 275         },
 276         'expected_warnings': ['height', 'width'],
 277         'skip': 'Account suspended',
 278     }, {
 279         'url': 'https://twitter.com/starwars/status/665052190608723968',
 280         'md5': '39b7199856dee6cd4432e72c74bc69d4',
 281         'info_dict': {
 282             'id': '665052190608723968',
 283             'ext': 'mp4',
 284             'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
 285             'description': 'Star Wars on Twitter: "A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens."',
 286             'uploader_id': 'starwars',
 287             'uploader': 'Star Wars',
 288         },
 289     }, {
 290         'url': 'https://twitter.com/BTNBrentYarina/status/705235433198714880',
 291         'info_dict': {
 292             'id': '705235433198714880',
 293             'ext': 'mp4',
 294             'title': 'Brent Yarina - Khalil Iverson\'s missed highlight dunk. And made highlight dunk. In one highlight.',
 295             'description': 'Brent Yarina on Twitter: "Khalil Iverson\'s missed highlight dunk. And made highlight dunk. In one highlight."',
 296             'uploader_id': 'BTNBrentYarina',
 297             'uploader': 'Brent Yarina',
 298         },
 299         'params': {
 300             # The same video as https://twitter.com/i/videos/tweet/705235433198714880
 301             # Test case of TwitterCardIE
 302             'skip_download': True,
 303         },
 304     }, {
 305         'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
 306         'md5': '',
 307         'info_dict': {
 308             'id': '700207533655363584',
 309             'ext': 'mp4',
 310             'title': 'あかさ - BEAT PROD: @suhmeduh #Damndaniel',
 311             'description': 'あかさ on Twitter: "BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ"',
 312             'thumbnail': r're:^https?://.*\.jpg',
 313             'uploader': 'あかさ',
 314             'uploader_id': 'jaydingeer',
 315             'duration': 30.0,
 316         },
 317         'params': {
 318             'skip_download': True,  # requires ffmpeg
 319         },
 320     }, {
 321         'url': 'https://twitter.com/Filmdrunk/status/713801302971588609',
 322         'md5': '89a15ed345d13b86e9a5a5e051fa308a',
 323         'info_dict': {
 324             'id': 'MIOxnrUteUd',
 325             'ext': 'mp4',
 326             'title': 'Vince Mancini - Vine of the day',
 327             'description': 'Vince Mancini on Twitter: "Vine of the day https://t.co/xmTvRdqxWf"',
 328             'uploader': 'Vince Mancini',
 329             'uploader_id': 'Filmdrunk',
 330             'timestamp': 1402826626,
 331             'upload_date': '20140615',
 332         },
 333         'add_ie': ['Vine'],
 334     }, {
 335         'url': 'https://twitter.com/captainamerica/status/719944021058060289',
 336         'info_dict': {
 337             'id': '719944021058060289',
 338             'ext': 'mp4',
 339             'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
 340             'description': 'Captain America on Twitter: "@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI"',
 341             'uploader_id': 'captainamerica',
 342             'uploader': 'Captain America',
 343             'duration': 3.17,
 344         },
 345         'params': {
 346             'skip_download': True,  # requires ffmpeg
 347         },
 348     }, {
 349         'url': 'https://twitter.com/OPP_HSD/status/779210622571536384',
 350         'info_dict': {
 351             'id': '1zqKVVlkqLaKB',
 352             'ext': 'mp4',
 353             'title': 'Sgt Kerry Schmidt - LIVE on #Periscope: Road rage, mischief, assault, rollover and fire in one occurrence',
 354             'description': 'Sgt Kerry Schmidt on Twitter: "LIVE on #Periscope: Road rage, mischief, assault, rollover and fire in one occurrence  https://t.co/EKrVgIXF3s"',
 355             'upload_date': '20160923',
 356             'uploader_id': 'OPP_HSD',
 357             'uploader': 'Sgt Kerry Schmidt',
 358             'timestamp': 1474613214,
 359         },
 360         'add_ie': ['Periscope'],
 361     }, {
 362         # has mp4 formats via mobile API
 363         'url': 'https://twitter.com/news_al3alm/status/852138619213144067',
 364         'info_dict': {
 365             'id': '852138619213144067',
 366             'ext': 'mp4',
 367             'title': 'عالم الأخبار - كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة',
 368             'description': 'عالم الأخبار on Twitter: "كلمة تاريخية بجلسة الجناسي التاريخية.. النائب خالد مؤنس العتيبي للمعارضين : اتقوا الله .. الظلم ظلمات يوم القيامة   https://t.co/xg6OhpyKfN"',
 369             'uploader': 'عالم الأخبار',
 370             'uploader_id': 'news_al3alm',
 371             'duration': 277.4,
 372         },
 373         'params': {
 374             'format': 'best[format_id^=http-]',
 375         },
 376     }, {
 377         'url': 'https://twitter.com/i/web/status/910031516746514432',
 378         'info_dict': {
 379             'id': '910031516746514432',
 380             'ext': 'mp4',
 381             '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.',
 382             'thumbnail': r're:^https?://.*\.jpg',
 383             'description': 'Préfet de Guadeloupe on Twitter: "[Direct] #Maria Le centre se trouve actuellement au sud de Basse-Terre. Restez confinés. Réfugiez-vous dans la pièce la + sûre. https://t.co/mwx01Rs4lo"',
 384             'uploader': 'Préfet de Guadeloupe',
 385             'uploader_id': 'Prefet971',
 386             'duration': 47.48,
 387         },
 388         'params': {
 389             'skip_download': True,  # requires ffmpeg
 390         },
 391     }]
 392
 393     def _real_extract(self, url):
 394         mobj = re.match(self._VALID_URL, url)
 395         user_id = mobj.group('user_id')
 396         twid = mobj.group('id')
 397
 398         webpage, urlh = self._download_webpage_handle(
 399             self._TEMPLATE_STATUSES_URL % twid, twid)
 400
 401         if 'twitter.com/account/suspended' in urlh.geturl():
 402             raise ExtractorError('Account suspended by Twitter.', expected=True)
 403
 404         if user_id is None:
 405             mobj = re.match(self._VALID_URL, urlh.geturl())
 406             user_id = mobj.group('user_id')
 407
 408         username = remove_end(self._og_search_title(webpage), ' on Twitter')
 409
 410         title = description = self._og_search_description(webpage).strip('').replace('\n', ' ').strip('“”')
 411
 412         # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
 413         title = re.sub(r'\s+(https?://[^ ]+)', '', title)
 414
 415         info = {
 416             'uploader_id': user_id,
 417             'uploader': username,
 418             'webpage_url': url,
 419             'description': '%s on Twitter: "%s"' % (username, description),
 420             'title': username + ' - ' + title,
 421         }
 422
 423         mobj = re.search(r'''(?x)
 424             <video[^>]+class="animated-gif"(?P<more_info>[^>]+)>\s*
 425                 <source[^>]+video-src="(?P<url>[^"]+)"
 426         ''', webpage)
 427
 428         if mobj:
 429             more_info = mobj.group('more_info')
 430             height = int_or_none(self._search_regex(
 431                 r'data-height="(\d+)"', more_info, 'height', fatal=False))
 432             width = int_or_none(self._search_regex(
 433                 r'data-width="(\d+)"', more_info, 'width', fatal=False))
 434             thumbnail = self._search_regex(
 435                 r'poster="([^"]+)"', more_info, 'poster', fatal=False)
 436             info.update({
 437                 'id': twid,
 438                 'url': mobj.group('url'),
 439                 'height': height,
 440                 'width': width,
 441                 'thumbnail': thumbnail,
 442             })
 443             return info
 444
 445         twitter_card_url = None
 446         if 'class="PlayableMedia' in webpage:
 447             twitter_card_url = '%s//twitter.com/i/videos/tweet/%s' % (self.http_scheme(), twid)
 448         else:
 449             twitter_card_iframe_url = self._search_regex(
 450                 r'data-full-card-iframe-url=([\'"])(?P<url>(?:(?!\1).)+)\1',
 451                 webpage, 'Twitter card iframe URL', default=None, group='url')
 452             if twitter_card_iframe_url:
 453                 twitter_card_url = compat_urlparse.urljoin(url, twitter_card_iframe_url)
 454
 455         if twitter_card_url:
 456             info.update({
 457                 '_type': 'url_transparent',
 458                 'ie_key': 'TwitterCard',
 459                 'url': twitter_card_url,
 460             })
 461             return info
 462
 463         raise ExtractorError('There\'s no video in this tweet.')
 464
 465
 466 class TwitterAmplifyIE(TwitterBaseIE):
 467     IE_NAME = 'twitter:amplify'
 468     _VALID_URL = r'https?://amp\.twimg\.com/v/(?P<id>[0-9a-f\-]{36})'
 469
 470     _TEST = {
 471         'url': 'https://amp.twimg.com/v/0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 472         'md5': '7df102d0b9fd7066b86f3159f8e81bf6',
 473         'info_dict': {
 474             'id': '0ba0c3c7-0af3-4c0a-bed5-7efd1ffa2951',
 475             'ext': 'mp4',
 476             'title': 'Twitter Video',
 477             'thumbnail': 're:^https?://.*',
 478         },
 479     }
 480
 481     def _real_extract(self, url):
 482         video_id = self._match_id(url)
 483         webpage = self._download_webpage(url, video_id)
 484
 485         vmap_url = self._html_search_meta(
 486             'twitter:amplify:vmap', webpage, 'vmap url')
 487         formats = self._extract_formats_from_vmap_url(vmap_url, video_id)
 488
 489         thumbnails = []
 490         thumbnail = self._html_search_meta(
 491             'twitter:image:src', webpage, 'thumbnail', fatal=False)
 492
 493         def _find_dimension(target):
 494             w = int_or_none(self._html_search_meta(
 495                 'twitter:%s:width' % target, webpage, fatal=False))
 496             h = int_or_none(self._html_search_meta(
 497                 'twitter:%s:height' % target, webpage, fatal=False))
 498             return w, h
 499
 500         if thumbnail:
 501             thumbnail_w, thumbnail_h = _find_dimension('image')
 502             thumbnails.append({
 503                 'url': thumbnail,
 504                 'width': thumbnail_w,
 505                 'height': thumbnail_h,
 506             })
 507
 508         video_w, video_h = _find_dimension('player')
 509         formats[0].update({
 510             'width': video_w,
 511             'height': video_h,
 512         })
 513
 514         return {
 515             'id': video_id,
 516             'title': 'Twitter Video',
 517             'formats': formats,
 518             'thumbnails': thumbnails,
 519         }