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