youtube_dl/extractor/facebook.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import socket
   6
   7 from .common import InfoExtractor
   8 from ..compat import (
   9     compat_etree_fromstring,
  10     compat_http_client,
  11     compat_urllib_error,
  12     compat_urllib_parse_unquote,
  13     compat_urllib_parse_unquote_plus,
  14 )
  15 from ..utils import (
  16     clean_html,
  17     error_to_compat_str,
  18     ExtractorError,
  19     get_element_by_id,
  20     int_or_none,
  21     js_to_json,
  22     limit_length,
  23     parse_count,
  24     sanitized_Request,
  25     try_get,
  26     urlencode_postdata,
  27 )
  28
  29
  30 class FacebookIE(InfoExtractor):
  31     _VALID_URL = r'''(?x)
  32                 (?:
  33                     https?://
  34                         (?:[\w-]+\.)?(?:facebook\.com|facebookcorewwwi\.onion)/
  35                         (?:[^#]*?\#!/)?
  36                         (?:
  37                             (?:
  38                                 video/video\.php|
  39                                 photo\.php|
  40                                 video\.php|
  41                                 video/embed|
  42                                 story\.php
  43                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  44                             [^/]+/videos/(?:[^/]+/)?|
  45                             [^/]+/posts/|
  46                             groups/[^/]+/permalink/
  47                         )|
  48                     facebook:
  49                 )
  50                 (?P<id>[0-9]+)
  51                 '''
  52     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  53     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  54     _NETRC_MACHINE = 'facebook'
  55     IE_NAME = 'facebook'
  56
  57     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
  58
  59     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  60     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
  61
  62     _TESTS = [{
  63         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  64         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  65         'info_dict': {
  66             'id': '637842556329505',
  67             'ext': 'mp4',
  68             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  69             'uploader': 'Tennis on Facebook',
  70             'upload_date': '20140908',
  71             'timestamp': 1410199200,
  72         },
  73         'skip': 'Requires logging in',
  74     }, {
  75         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  76         'info_dict': {
  77             'id': '274175099429670',
  78             'ext': 'mp4',
  79             'title': 're:^Asif Nawab Butt posted a video',
  80             'uploader': 'Asif Nawab Butt',
  81             'upload_date': '20140506',
  82             'timestamp': 1399398998,
  83             'thumbnail': r're:^https?://.*',
  84         },
  85         'expected_warnings': [
  86             'title'
  87         ]
  88     }, {
  89         'note': 'Video with DASH manifest',
  90         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  91         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
  92         'info_dict': {
  93             'id': '957955867617029',
  94             'ext': 'mp4',
  95             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
  96             'uploader': 'Demy de Zeeuw',
  97             'upload_date': '20160110',
  98             'timestamp': 1452431627,
  99         },
 100         'skip': 'Requires logging in',
 101     }, {
 102         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 103         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 104         'info_dict': {
 105             'id': '544765982287235',
 106             'ext': 'mp4',
 107             'title': '"What are you doing running in the snow?"',
 108             'uploader': 'FailArmy',
 109         },
 110         'skip': 'Video gone',
 111     }, {
 112         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 113         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 114         'info_dict': {
 115             'id': '1035862816472149',
 116             'ext': 'mp4',
 117             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 118             'uploader': 'S. Saint',
 119         },
 120         'skip': 'Video gone',
 121     }, {
 122         'note': 'swf params escaped',
 123         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 124         'md5': '97ba073838964d12c70566e0085c2b91',
 125         'info_dict': {
 126             'id': '10153664894881749',
 127             'ext': 'mp4',
 128             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 129             'thumbnail': r're:^https?://.*',
 130             'timestamp': 1456259628,
 131             'upload_date': '20160223',
 132             'uploader': 'Barack Obama',
 133         },
 134     }, {
 135         # have 1080P, but only up to 720p in swf params
 136         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 137         'md5': '9571fae53d4165bbbadb17a94651dcdc',
 138         'info_dict': {
 139             'id': '10155529876156509',
 140             'ext': 'mp4',
 141             'title': 'She survived the holocaust — and years later, she’s getting her citizenship s...',
 142             'timestamp': 1477818095,
 143             'upload_date': '20161030',
 144             'uploader': 'CNN',
 145             'thumbnail': r're:^https?://.*',
 146             'view_count': int,
 147         },
 148     }, {
 149         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 150         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 151         'info_dict': {
 152             'id': '1417995061575415',
 153             'ext': 'mp4',
 154             'title': 'md5:1db063d6a8c13faa8da727817339c857',
 155             'timestamp': 1486648217,
 156             'upload_date': '20170209',
 157             'uploader': 'Yaroslav Korpan',
 158         },
 159         'params': {
 160             'skip_download': True,
 161         },
 162     }, {
 163         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 164         'info_dict': {
 165             'id': '1072691702860471',
 166             'ext': 'mp4',
 167             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 168             'timestamp': 1477305000,
 169             'upload_date': '20161024',
 170             'uploader': 'La Guía Del Varón',
 171             'thumbnail': r're:^https?://.*',
 172         },
 173         'params': {
 174             'skip_download': True,
 175         },
 176     }, {
 177         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 178         'info_dict': {
 179             'id': '1396382447100162',
 180             'ext': 'mp4',
 181             'title': 'md5:19a428bbde91364e3de815383b54a235',
 182             'timestamp': 1486035494,
 183             'upload_date': '20170202',
 184             'uploader': 'Elisabeth Ahtn',
 185         },
 186         'params': {
 187             'skip_download': True,
 188         },
 189     }, {
 190         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 191         'only_matching': True,
 192     }, {
 193         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 194         'only_matching': True,
 195     }, {
 196         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 197         'only_matching': True,
 198     }, {
 199         'url': 'facebook:544765982287235',
 200         'only_matching': True,
 201     }, {
 202         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 203         'only_matching': True,
 204     }, {
 205         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 206         'only_matching': True,
 207     }, {
 208         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
 209         'only_matching': True,
 210     }, {
 211         # no title
 212         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 213         'only_matching': True,
 214     }, {
 215         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 216         'info_dict': {
 217             'id': '359649331226507',
 218             'ext': 'mp4',
 219             'title': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 220             'uploader': 'ESL One Dota 2',
 221         },
 222         'params': {
 223             'skip_download': True,
 224         },
 225     }]
 226
 227     @staticmethod
 228     def _extract_urls(webpage):
 229         urls = []
 230         for mobj in re.finditer(
 231                 r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
 232                 webpage):
 233             urls.append(mobj.group('url'))
 234         # Facebook API embed
 235         # see https://developers.facebook.com/docs/plugins/embedded-video-player
 236         for mobj in re.finditer(r'''(?x)<div[^>]+
 237                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
 238                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage):
 239             urls.append(mobj.group('url'))
 240         return urls
 241
 242     def _login(self):
 243         useremail, password = self._get_login_info()
 244         if useremail is None:
 245             return
 246
 247         login_page_req = sanitized_Request(self._LOGIN_URL)
 248         self._set_cookie('facebook.com', 'locale', 'en_US')
 249         login_page = self._download_webpage(login_page_req, None,
 250                                             note='Downloading login page',
 251                                             errnote='Unable to download login page')
 252         lsd = self._search_regex(
 253             r'<input type="hidden" name="lsd" value="([^"]*)"',
 254             login_page, 'lsd')
 255         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 256
 257         login_form = {
 258             'email': useremail,
 259             'pass': password,
 260             'lsd': lsd,
 261             'lgnrnd': lgnrnd,
 262             'next': 'http://facebook.com/home.php',
 263             'default_persistent': '0',
 264             'legacy_return': '1',
 265             'timezone': '-60',
 266             'trynum': '1',
 267         }
 268         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 269         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 270         try:
 271             login_results = self._download_webpage(request, None,
 272                                                    note='Logging in', errnote='unable to fetch login page')
 273             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 274                 error = self._html_search_regex(
 275                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 276                     login_results, 'login error', default=None, group='error')
 277                 if error:
 278                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 279                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 280                 return
 281
 282             fb_dtsg = self._search_regex(
 283                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 284             h = self._search_regex(
 285                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 286
 287             if not fb_dtsg or not h:
 288                 return
 289
 290             check_form = {
 291                 'fb_dtsg': fb_dtsg,
 292                 'h': h,
 293                 'name_action_selected': 'dont_save',
 294             }
 295             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 296             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 297             check_response = self._download_webpage(check_req, None,
 298                                                     note='Confirming login')
 299             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 300                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 301         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 302             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
 303             return
 304
 305     def _real_initialize(self):
 306         self._login()
 307
 308     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
 309         req = sanitized_Request(url)
 310         req.add_header('User-Agent', self._CHROME_USER_AGENT)
 311         webpage = self._download_webpage(req, video_id)
 312
 313         video_data = None
 314
 315         def extract_video_data(instances):
 316             for item in instances:
 317                 if item[1][0] == 'VideoConfig':
 318                     video_item = item[2][0]
 319                     if video_item.get('video_id'):
 320                         return video_item['videoData']
 321
 322         server_js_data = self._parse_json(self._search_regex(
 323             r'handleServerJS\(({.+})(?:\);|,")', webpage,
 324             'server js data', default='{}'), video_id, fatal=False)
 325
 326         if server_js_data:
 327             video_data = extract_video_data(server_js_data.get('instances', []))
 328
 329         def extract_from_jsmods_instances(js_data):
 330             if js_data:
 331                 return extract_video_data(try_get(
 332                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 333
 334         if not video_data:
 335             server_js_data = self._parse_json(
 336                 self._search_regex(
 337                     r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_\d+)',
 338                     webpage, 'js data', default='{}'),
 339                 video_id, transform_source=js_to_json, fatal=False)
 340             video_data = extract_from_jsmods_instances(server_js_data)
 341
 342         if not video_data:
 343             if not fatal_if_no_video:
 344                 return webpage, False
 345             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 346             if m_msg is not None:
 347                 raise ExtractorError(
 348                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 349                     expected=True)
 350             elif '>You must log in to continue' in webpage:
 351                 self.raise_login_required()
 352
 353             # Video info not in first request, do a secondary request using
 354             # tahoe player specific URL
 355             tahoe_data = self._download_webpage(
 356                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 357                 data=urlencode_postdata({
 358                     '__a': 1,
 359                     '__pc': self._search_regex(
 360                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 361                         'pkg cohort', default='PHASED:DEFAULT'),
 362                     '__rev': self._search_regex(
 363                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 364                         'client revision', default='3944515'),
 365                     'fb_dtsg': self._search_regex(
 366                         r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
 367                         webpage, 'dtsg token', default=''),
 368                 }),
 369                 headers={
 370                     'Content-Type': 'application/x-www-form-urlencoded',
 371                 })
 372             tahoe_js_data = self._parse_json(
 373                 self._search_regex(
 374                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 375                     'tahoe js data', default='{}'),
 376                 video_id, fatal=False)
 377             video_data = extract_from_jsmods_instances(tahoe_js_data)
 378
 379         if not video_data:
 380             raise ExtractorError('Cannot parse data')
 381
 382         subtitles = {}
 383         formats = []
 384         for f in video_data:
 385             format_id = f['stream_type']
 386             if f and isinstance(f, dict):
 387                 f = [f]
 388             if not f or not isinstance(f, list):
 389                 continue
 390             for quality in ('sd', 'hd'):
 391                 for src_type in ('src', 'src_no_ratelimit'):
 392                     src = f[0].get('%s_%s' % (quality, src_type))
 393                     if src:
 394                         preference = -10 if format_id == 'progressive' else 0
 395                         if quality == 'hd':
 396                             preference += 5
 397                         formats.append({
 398                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 399                             'url': src,
 400                             'preference': preference,
 401                         })
 402             dash_manifest = f[0].get('dash_manifest')
 403             if dash_manifest:
 404                 formats.extend(self._parse_mpd_formats(
 405                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 406             subtitles_src = f[0].get('subtitles_src')
 407             if subtitles_src:
 408                 subtitles.setdefault('en', []).append({'url': subtitles_src})
 409         if not formats:
 410             raise ExtractorError('Cannot find video formats')
 411
 412         # Downloads with browser's User-Agent are rate limited. Working around
 413         # with non-browser User-Agent.
 414         for f in formats:
 415             f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
 416
 417         self._sort_formats(formats)
 418
 419         video_title = self._html_search_regex(
 420             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
 421             'title', default=None)
 422         if not video_title:
 423             video_title = self._html_search_regex(
 424                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 425                 webpage, 'alternative title', default=None)
 426         if not video_title:
 427             video_title = self._html_search_meta(
 428                 'description', webpage, 'title', default=None)
 429         if video_title:
 430             video_title = limit_length(video_title, 80)
 431         else:
 432             video_title = 'Facebook video #%s' % video_id
 433         uploader = clean_html(get_element_by_id(
 434             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
 435             r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader',
 436             default=None) or self._og_search_title(webpage, fatal=False)
 437         timestamp = int_or_none(self._search_regex(
 438             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 439             'timestamp', default=None))
 440         thumbnail = self._html_search_meta(['og:image', 'twitter:image'], webpage)
 441
 442         view_count = parse_count(self._search_regex(
 443             r'\bviewCount\s*:\s*["\']([\d,.]+)', webpage, 'view count',
 444             default=None))
 445
 446         info_dict = {
 447             'id': video_id,
 448             'title': video_title,
 449             'formats': formats,
 450             'uploader': uploader,
 451             'timestamp': timestamp,
 452             'thumbnail': thumbnail,
 453             'view_count': view_count,
 454             'subtitles': subtitles,
 455         }
 456
 457         return webpage, info_dict
 458
 459     def _real_extract(self, url):
 460         video_id = self._match_id(url)
 461
 462         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 463         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
 464
 465         if info_dict:
 466             return info_dict
 467
 468         if '/posts/' in url:
 469             video_id_json = self._search_regex(
 470                 r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])', webpage, 'video ids', group='ids',
 471                 default='')
 472             if video_id_json:
 473                 entries = [
 474                     self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
 475                     for vid in self._parse_json(video_id_json, video_id)]
 476                 return self.playlist_result(entries, video_id)
 477
 478             # Single Video?
 479             video_id = self._search_regex(r'video_id:\s*"([0-9]+)"', webpage, 'single video id')
 480             return self.url_result('facebook:%s' % video_id, FacebookIE.ie_key())
 481         else:
 482             _, info_dict = self._extract_from_url(
 483                 self._VIDEO_PAGE_TEMPLATE % video_id,
 484                 video_id, fatal_if_no_video=True)
 485             return info_dict
 486
 487
 488 class FacebookPluginsVideoIE(InfoExtractor):
 489     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 490
 491     _TESTS = [{
 492         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fgov.sg%2Fvideos%2F10154383743583686%2F&show_text=0&width=560',
 493         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 494         'info_dict': {
 495             'id': '10154383743583686',
 496             'ext': 'mp4',
 497             'title': 'What to do during the haze?',
 498             'uploader': 'Gov.sg',
 499             'upload_date': '20160826',
 500             'timestamp': 1472184808,
 501         },
 502         'add_ie': [FacebookIE.ie_key()],
 503     }, {
 504         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 505         'only_matching': True,
 506     }, {
 507         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 508         'only_matching': True,
 509     }]
 510
 511     def _real_extract(self, url):
 512         return self.url_result(
 513             compat_urllib_parse_unquote(self._match_id(url)),
 514             FacebookIE.ie_key())