_ Git - youtube-dl/blob - 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     sanitized_Request,
  24     try_get,
  25     urlencode_postdata,
  26 )
  27
  28
  29 class FacebookIE(InfoExtractor):
  30     _VALID_URL = r'''(?x)
  31                 (?:
  32                     https?://
  33                         (?:[\w-]+\.)?(?:facebook\.com|facebookcorewwwi\.onion)/
  34                         (?:[^#]*?\#!/)?
  35                         (?:
  36                             (?:
  37                                 video/video\.php|
  38                                 photo\.php|
  39                                 video\.php|
  40                                 video/embed|
  41                                 story\.php
  42                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  43                             [^/]+/videos/(?:[^/]+/)?|
  44                             [^/]+/posts/|
  45                             groups/[^/]+/permalink/
  46                         )|
  47                     facebook:
  48                 )
  49                 (?P<id>[0-9]+)
  50                 '''
  51     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  52     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  53     _NETRC_MACHINE = 'facebook'
  54     IE_NAME = 'facebook'
  55
  56     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
  57
  58     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  59     _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true'
  60
  61     _TESTS = [{
  62         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  63         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  64         'info_dict': {
  65             'id': '637842556329505',
  66             'ext': 'mp4',
  67             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  68             'uploader': 'Tennis on Facebook',
  69             'upload_date': '20140908',
  70             'timestamp': 1410199200,
  71         },
  72         'skip': 'Requires logging in',
  73     }, {
  74         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  75         'info_dict': {
  76             'id': '274175099429670',
  77             'ext': 'mp4',
  78             'title': 'Asif Nawab Butt posted a video to his Timeline.',
  79             'uploader': 'Asif Nawab Butt',
  80             'upload_date': '20140506',
  81             'timestamp': 1399398998,
  82             'thumbnail': r're:^https?://.*',
  83         },
  84         'expected_warnings': [
  85             'title'
  86         ]
  87     }, {
  88         'note': 'Video with DASH manifest',
  89         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  90         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
  91         'info_dict': {
  92             'id': '957955867617029',
  93             'ext': 'mp4',
  94             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
  95             'uploader': 'Demy de Zeeuw',
  96             'upload_date': '20160110',
  97             'timestamp': 1452431627,
  98         },
  99         'skip': 'Requires logging in',
 100     }, {
 101         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
 102         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
 103         'info_dict': {
 104             'id': '544765982287235',
 105             'ext': 'mp4',
 106             'title': '"What are you doing running in the snow?"',
 107             'uploader': 'FailArmy',
 108         },
 109         'skip': 'Video gone',
 110     }, {
 111         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 112         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 113         'info_dict': {
 114             'id': '1035862816472149',
 115             'ext': 'mp4',
 116             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 117             'uploader': 'S. Saint',
 118         },
 119         'skip': 'Video gone',
 120     }, {
 121         'note': 'swf params escaped',
 122         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 123         'md5': '97ba073838964d12c70566e0085c2b91',
 124         'info_dict': {
 125             'id': '10153664894881749',
 126             'ext': 'mp4',
 127             'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
 128             'thumbnail': r're:^https?://.*',
 129             'timestamp': 1456259628,
 130             'upload_date': '20160223',
 131             'uploader': 'Barack Obama',
 132         },
 133     }, {
 134         # have 1080P, but only up to 720p in swf params
 135         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
 136         'md5': '0d9813160b146b3bc8744e006027fcc6',
 137         'info_dict': {
 138             'id': '10155529876156509',
 139             'ext': 'mp4',
 140             'title': 'She survived the holocaust — and years later, she’s getting her citizenship s...',
 141             'timestamp': 1477818095,
 142             'upload_date': '20161030',
 143             'uploader': 'CNN',
 144             'thumbnail': r're:^https?://.*',
 145         },
 146     }, {
 147         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
 148         'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/',
 149         'info_dict': {
 150             'id': '1417995061575415',
 151             'ext': 'mp4',
 152             'title': 'md5:a7b86ca673f51800cd54687b7f4012fe',
 153             'timestamp': 1486648217,
 154             'upload_date': '20170209',
 155             'uploader': 'Yaroslav Korpan',
 156         },
 157         'params': {
 158             'skip_download': True,
 159         },
 160     }, {
 161         'url': 'https://www.facebook.com/LaGuiaDelVaron/posts/1072691702860471',
 162         'info_dict': {
 163             'id': '1072691702860471',
 164             'ext': 'mp4',
 165             'title': 'md5:ae2d22a93fbb12dad20dc393a869739d',
 166             'timestamp': 1477305000,
 167             'upload_date': '20161024',
 168             'uploader': 'La Guía Del Varón',
 169             'thumbnail': r're:^https?://.*',
 170         },
 171         'params': {
 172             'skip_download': True,
 173         },
 174     }, {
 175         'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/',
 176         'info_dict': {
 177             'id': '1396382447100162',
 178             'ext': 'mp4',
 179             'title': 'md5:e2d2700afdf84e121f5d0f999bad13a3',
 180             'timestamp': 1486035494,
 181             'upload_date': '20170202',
 182             'uploader': 'Elisabeth Ahtn',
 183         },
 184         'params': {
 185             'skip_download': True,
 186         },
 187     }, {
 188         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 189         'only_matching': True,
 190     }, {
 191         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 192         'only_matching': True,
 193     }, {
 194         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 195         'only_matching': True,
 196     }, {
 197         'url': 'facebook:544765982287235',
 198         'only_matching': True,
 199     }, {
 200         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 201         'only_matching': True,
 202     }, {
 203         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 204         'only_matching': True,
 205     }, {
 206         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
 207         'only_matching': True,
 208     }, {
 209         # no title
 210         'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
 211         'only_matching': True,
 212     }, {
 213         'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
 214         'info_dict': {
 215             'id': '359649331226507',
 216             'ext': 'mp4',
 217             'title': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
 218             'uploader': 'ESL One Dota 2',
 219         },
 220         'params': {
 221             'skip_download': True,
 222         },
 223     }]
 224
 225     @staticmethod
 226     def _extract_urls(webpage):
 227         urls = []
 228         for mobj in re.finditer(
 229                 r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
 230                 webpage):
 231             urls.append(mobj.group('url'))
 232         # Facebook API embed
 233         # see https://developers.facebook.com/docs/plugins/embedded-video-player
 234         for mobj in re.finditer(r'''(?x)<div[^>]+
 235                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
 236                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage):
 237             urls.append(mobj.group('url'))
 238         return urls
 239
 240     def _login(self):
 241         useremail, password = self._get_login_info()
 242         if useremail is None:
 243             return
 244
 245         login_page_req = sanitized_Request(self._LOGIN_URL)
 246         self._set_cookie('facebook.com', 'locale', 'en_US')
 247         login_page = self._download_webpage(login_page_req, None,
 248                                             note='Downloading login page',
 249                                             errnote='Unable to download login page')
 250         lsd = self._search_regex(
 251             r'<input type="hidden" name="lsd" value="([^"]*)"',
 252             login_page, 'lsd')
 253         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 254
 255         login_form = {
 256             'email': useremail,
 257             'pass': password,
 258             'lsd': lsd,
 259             'lgnrnd': lgnrnd,
 260             'next': 'http://facebook.com/home.php',
 261             'default_persistent': '0',
 262             'legacy_return': '1',
 263             'timezone': '-60',
 264             'trynum': '1',
 265         }
 266         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 267         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 268         try:
 269             login_results = self._download_webpage(request, None,
 270                                                    note='Logging in', errnote='unable to fetch login page')
 271             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 272                 error = self._html_search_regex(
 273                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 274                     login_results, 'login error', default=None, group='error')
 275                 if error:
 276                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 277                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 278                 return
 279
 280             fb_dtsg = self._search_regex(
 281                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 282             h = self._search_regex(
 283                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 284
 285             if not fb_dtsg or not h:
 286                 return
 287
 288             check_form = {
 289                 'fb_dtsg': fb_dtsg,
 290                 'h': h,
 291                 'name_action_selected': 'dont_save',
 292             }
 293             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 294             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 295             check_response = self._download_webpage(check_req, None,
 296                                                     note='Confirming login')
 297             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 298                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 299         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 300             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
 301             return
 302
 303     def _real_initialize(self):
 304         self._login()
 305
 306     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
 307         req = sanitized_Request(url)
 308         req.add_header('User-Agent', self._CHROME_USER_AGENT)
 309         webpage = self._download_webpage(req, video_id)
 310
 311         video_data = None
 312
 313         def extract_video_data(instances):
 314             for item in instances:
 315                 if item[1][0] == 'VideoConfig':
 316                     video_item = item[2][0]
 317                     if video_item.get('video_id'):
 318                         return video_item['videoData']
 319
 320         server_js_data = self._parse_json(self._search_regex(
 321             r'handleServerJS\(({.+})(?:\);|,")', webpage,
 322             'server js data', default='{}'), video_id, fatal=False)
 323
 324         if server_js_data:
 325             video_data = extract_video_data(server_js_data.get('instances', []))
 326
 327         def extract_from_jsmods_instances(js_data):
 328             if js_data:
 329                 return extract_video_data(try_get(
 330                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 331
 332         if not video_data:
 333             server_js_data = self._parse_json(
 334                 self._search_regex(
 335                     r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+(?:stream_pagelet|pagelet_group_mall|permalink_video_pagelet)',
 336                     webpage, 'js data', default='{}'),
 337                 video_id, transform_source=js_to_json, fatal=False)
 338             video_data = extract_from_jsmods_instances(server_js_data)
 339
 340         if not video_data:
 341             if not fatal_if_no_video:
 342                 return webpage, False
 343             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 344             if m_msg is not None:
 345                 raise ExtractorError(
 346                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 347                     expected=True)
 348             elif '>You must log in to continue' in webpage:
 349                 self.raise_login_required()
 350
 351             # Video info not in first request, do a secondary request using
 352             # tahoe player specific URL
 353             tahoe_data = self._download_webpage(
 354                 self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
 355                 data=urlencode_postdata({
 356                     '__user': 0,
 357                     '__a': 1,
 358                     '__pc': self._search_regex(
 359                         r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
 360                         'pkg cohort', default='PHASED:DEFAULT'),
 361                     '__rev': self._search_regex(
 362                         r'client_revision["\']\s*:\s*(\d+),', webpage,
 363                         'client revision', default='3944515'),
 364                 }),
 365                 headers={
 366                     'Content-Type': 'application/x-www-form-urlencoded',
 367                 })
 368             tahoe_js_data = self._parse_json(
 369                 self._search_regex(
 370                     r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
 371                     'tahoe js data', default='{}'),
 372                 video_id, fatal=False)
 373             video_data = extract_from_jsmods_instances(tahoe_js_data)
 374
 375         if not video_data:
 376             raise ExtractorError('Cannot parse data')
 377
 378         formats = []
 379         for f in video_data:
 380             format_id = f['stream_type']
 381             if f and isinstance(f, dict):
 382                 f = [f]
 383             if not f or not isinstance(f, list):
 384                 continue
 385             for quality in ('sd', 'hd'):
 386                 for src_type in ('src', 'src_no_ratelimit'):
 387                     src = f[0].get('%s_%s' % (quality, src_type))
 388                     if src:
 389                         preference = -10 if format_id == 'progressive' else 0
 390                         if quality == 'hd':
 391                             preference += 5
 392                         formats.append({
 393                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 394                             'url': src,
 395                             'preference': preference,
 396                         })
 397             dash_manifest = f[0].get('dash_manifest')
 398             if dash_manifest:
 399                 formats.extend(self._parse_mpd_formats(
 400                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 401         if not formats:
 402             raise ExtractorError('Cannot find video formats')
 403
 404         self._sort_formats(formats)
 405
 406         video_title = self._html_search_regex(
 407             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
 408             'title', default=None)
 409         if not video_title:
 410             video_title = self._html_search_regex(
 411                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 412                 webpage, 'alternative title', default=None)
 413         if not video_title:
 414             video_title = self._html_search_meta(
 415                 'description', webpage, 'title', default=None)
 416         if video_title:
 417             video_title = limit_length(video_title, 80)
 418         else:
 419             video_title = 'Facebook video #%s' % video_id
 420         uploader = clean_html(get_element_by_id(
 421             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
 422             r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader',
 423             fatal=False) or self._og_search_title(webpage, fatal=False)
 424         timestamp = int_or_none(self._search_regex(
 425             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 426             'timestamp', default=None))
 427         thumbnail = self._og_search_thumbnail(webpage)
 428
 429         info_dict = {
 430             'id': video_id,
 431             'title': video_title,
 432             'formats': formats,
 433             'uploader': uploader,
 434             'timestamp': timestamp,
 435             'thumbnail': thumbnail,
 436         }
 437
 438         return webpage, info_dict
 439
 440     def _real_extract(self, url):
 441         video_id = self._match_id(url)
 442
 443         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 444         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
 445
 446         if info_dict:
 447             return info_dict
 448
 449         if '/posts/' in url:
 450             entries = [
 451                 self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
 452                 for vid in self._parse_json(
 453                     self._search_regex(
 454                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
 455                         webpage, 'video ids', group='ids'),
 456                     video_id)]
 457
 458             return self.playlist_result(entries, video_id)
 459         else:
 460             _, info_dict = self._extract_from_url(
 461                 self._VIDEO_PAGE_TEMPLATE % video_id,
 462                 video_id, fatal_if_no_video=True)
 463             return info_dict
 464
 465
 466 class FacebookPluginsVideoIE(InfoExtractor):
 467     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 468
 469     _TESTS = [{
 470         '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',
 471         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 472         'info_dict': {
 473             'id': '10154383743583686',
 474             'ext': 'mp4',
 475             'title': 'What to do during the haze?',
 476             'uploader': 'Gov.sg',
 477             'upload_date': '20160826',
 478             'timestamp': 1472184808,
 479         },
 480         'add_ie': [FacebookIE.ie_key()],
 481     }, {
 482         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 483         'only_matching': True,
 484     }, {
 485         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 486         'only_matching': True,
 487     }]
 488
 489     def _real_extract(self, url):
 490         return self.url_result(
 491             compat_urllib_parse_unquote(self._match_id(url)),
 492             FacebookIE.ie_key())