_ Git - youtube-dl/blob - youtube_dl/extractor/facebook.py

   1 from __future__ import unicode_literals
   2
   3 import json
   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     error_to_compat_str,
  17     ExtractorError,
  18     int_or_none,
  19     limit_length,
  20     sanitized_Request,
  21     urlencode_postdata,
  22     get_element_by_id,
  23     clean_html,
  24 )
  25
  26
  27 class FacebookIE(InfoExtractor):
  28     _VALID_URL = r'''(?x)
  29                 (?:
  30                     https?://
  31                         (?:[\w-]+\.)?facebook\.com/
  32                         (?:[^#]*?\#!/)?
  33                         (?:
  34                             (?:
  35                                 video/video\.php|
  36                                 photo\.php|
  37                                 video\.php|
  38                                 video/embed|
  39                                 story\.php
  40                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  41                             [^/]+/videos/(?:[^/]+/)?|
  42                             [^/]+/posts/|
  43                             groups/[^/]+/permalink/
  44                         )|
  45                     facebook:
  46                 )
  47                 (?P<id>[0-9]+)
  48                 '''
  49     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  50     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  51     _NETRC_MACHINE = 'facebook'
  52     IE_NAME = 'facebook'
  53
  54     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
  55
  56     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  57
  58     _TESTS = [{
  59         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  60         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  61         'info_dict': {
  62             'id': '637842556329505',
  63             'ext': 'mp4',
  64             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  65             'uploader': 'Tennis on Facebook',
  66             'upload_date': '20140908',
  67             'timestamp': 1410199200,
  68         }
  69     }, {
  70         'note': 'Video without discernible title',
  71         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  72         'info_dict': {
  73             'id': '274175099429670',
  74             'ext': 'mp4',
  75             'title': 'Facebook video #274175099429670',
  76             'uploader': 'Asif Nawab Butt',
  77             'upload_date': '20140506',
  78             'timestamp': 1399398998,
  79         },
  80         'expected_warnings': [
  81             'title'
  82         ]
  83     }, {
  84         'note': 'Video with DASH manifest',
  85         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  86         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
  87         'info_dict': {
  88             'id': '957955867617029',
  89             'ext': 'mp4',
  90             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
  91             'uploader': 'Demy de Zeeuw',
  92             'upload_date': '20160110',
  93             'timestamp': 1452431627,
  94         },
  95     }, {
  96         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
  97         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
  98         'info_dict': {
  99             'id': '544765982287235',
 100             'ext': 'mp4',
 101             'title': '"What are you doing running in the snow?"',
 102             'uploader': 'FailArmy',
 103         }
 104     }, {
 105         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
 106         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
 107         'info_dict': {
 108             'id': '1035862816472149',
 109             'ext': 'mp4',
 110             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 111             'uploader': 'S. Saint',
 112         },
 113     }, {
 114         'note': 'swf params escaped',
 115         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 116         'md5': '97ba073838964d12c70566e0085c2b91',
 117         'info_dict': {
 118             'id': '10153664894881749',
 119             'ext': 'mp4',
 120             'title': 'Facebook video #10153664894881749',
 121         },
 122     }, {
 123         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 124         'only_matching': True,
 125     }, {
 126         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 127         'only_matching': True,
 128     }, {
 129         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 130         'only_matching': True,
 131     }, {
 132         'url': 'facebook:544765982287235',
 133         'only_matching': True,
 134     }, {
 135         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
 136         'only_matching': True,
 137     }, {
 138         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
 139         'only_matching': True,
 140     }]
 141
 142     @staticmethod
 143     def _extract_url(webpage):
 144         mobj = re.search(
 145             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
 146         if mobj is not None:
 147             return mobj.group('url')
 148
 149         # Facebook API embed
 150         # see https://developers.facebook.com/docs/plugins/embedded-video-player
 151         mobj = re.search(r'''(?x)<div[^>]+
 152                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
 153                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
 154         if mobj is not None:
 155             return mobj.group('url')
 156
 157     def _login(self):
 158         (useremail, password) = self._get_login_info()
 159         if useremail is None:
 160             return
 161
 162         login_page_req = sanitized_Request(self._LOGIN_URL)
 163         self._set_cookie('facebook.com', 'locale', 'en_US')
 164         login_page = self._download_webpage(login_page_req, None,
 165                                             note='Downloading login page',
 166                                             errnote='Unable to download login page')
 167         lsd = self._search_regex(
 168             r'<input type="hidden" name="lsd" value="([^"]*)"',
 169             login_page, 'lsd')
 170         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 171
 172         login_form = {
 173             'email': useremail,
 174             'pass': password,
 175             'lsd': lsd,
 176             'lgnrnd': lgnrnd,
 177             'next': 'http://facebook.com/home.php',
 178             'default_persistent': '0',
 179             'legacy_return': '1',
 180             'timezone': '-60',
 181             'trynum': '1',
 182         }
 183         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 184         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 185         try:
 186             login_results = self._download_webpage(request, None,
 187                                                    note='Logging in', errnote='unable to fetch login page')
 188             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 189                 error = self._html_search_regex(
 190                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 191                     login_results, 'login error', default=None, group='error')
 192                 if error:
 193                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 194                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 195                 return
 196
 197             fb_dtsg = self._search_regex(
 198                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 199             h = self._search_regex(
 200                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 201
 202             if not fb_dtsg or not h:
 203                 return
 204
 205             check_form = {
 206                 'fb_dtsg': fb_dtsg,
 207                 'h': h,
 208                 'name_action_selected': 'dont_save',
 209             }
 210             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 211             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 212             check_response = self._download_webpage(check_req, None,
 213                                                     note='Confirming login')
 214             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 215                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 216         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 217             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
 218             return
 219
 220     def _real_initialize(self):
 221         self._login()
 222
 223     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
 224         req = sanitized_Request(url)
 225         req.add_header('User-Agent', self._CHROME_USER_AGENT)
 226         webpage = self._download_webpage(req, video_id)
 227
 228         video_data = None
 229
 230         BEFORE = '{swf.addParam(param[0], param[1]);});'
 231         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
 232         PATTERN = re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER)
 233
 234         for m in re.findall(PATTERN, webpage):
 235             swf_params = m.replace('\\\\', '\\').replace('\\"', '"')
 236             data = dict(json.loads(swf_params))
 237             params_raw = compat_urllib_parse_unquote(data['params'])
 238             video_data_candidate = json.loads(params_raw)['video_data']
 239             for _, f in video_data_candidate.items():
 240                 if not f:
 241                     continue
 242                 if isinstance(f, dict):
 243                     f = [f]
 244                 if not isinstance(f, list):
 245                     continue
 246                 if f[0].get('video_id') == video_id:
 247                     video_data = video_data_candidate
 248                     break
 249             if video_data:
 250                 break
 251
 252         def video_data_list2dict(video_data):
 253             ret = {}
 254             for item in video_data:
 255                 format_id = item['stream_type']
 256                 ret.setdefault(format_id, []).append(item)
 257             return ret
 258
 259         if not video_data:
 260             server_js_data = self._parse_json(self._search_regex(
 261                 r'handleServerJS\(({.+})(?:\);|,")', webpage, 'server js data', default='{}'), video_id)
 262             for item in server_js_data.get('instances', []):
 263                 if item[1][0] == 'VideoConfig':
 264                     video_data = video_data_list2dict(item[2][0]['videoData'])
 265                     break
 266
 267         if not video_data:
 268             if not fatal_if_no_video:
 269                 return webpage, False
 270             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 271             if m_msg is not None:
 272                 raise ExtractorError(
 273                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 274                     expected=True)
 275             else:
 276                 raise ExtractorError('Cannot parse data')
 277
 278         formats = []
 279         for format_id, f in video_data.items():
 280             if f and isinstance(f, dict):
 281                 f = [f]
 282             if not f or not isinstance(f, list):
 283                 continue
 284             for quality in ('sd', 'hd'):
 285                 for src_type in ('src', 'src_no_ratelimit'):
 286                     src = f[0].get('%s_%s' % (quality, src_type))
 287                     if src:
 288                         preference = -10 if format_id == 'progressive' else 0
 289                         if quality == 'hd':
 290                             preference += 5
 291                         formats.append({
 292                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 293                             'url': src,
 294                             'preference': preference,
 295                         })
 296             dash_manifest = f[0].get('dash_manifest')
 297             if dash_manifest:
 298                 formats.extend(self._parse_mpd_formats(
 299                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 300         if not formats:
 301             raise ExtractorError('Cannot find video formats')
 302
 303         self._sort_formats(formats)
 304
 305         video_title = self._html_search_regex(
 306             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
 307             default=None)
 308         if not video_title:
 309             video_title = self._html_search_regex(
 310                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 311                 webpage, 'alternative title', default=None)
 312             video_title = limit_length(video_title, 80)
 313         if not video_title:
 314             video_title = 'Facebook video #%s' % video_id
 315         uploader = clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
 316         timestamp = int_or_none(self._search_regex(
 317             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
 318             'timestamp', default=None))
 319
 320         info_dict = {
 321             'id': video_id,
 322             'title': video_title,
 323             'formats': formats,
 324             'uploader': uploader,
 325             'timestamp': timestamp,
 326         }
 327
 328         return webpage, info_dict
 329
 330     def _real_extract(self, url):
 331         video_id = self._match_id(url)
 332
 333         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 334         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
 335
 336         if info_dict:
 337             return info_dict
 338
 339         if '/posts/' in url:
 340             entries = [
 341                 self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
 342                 for vid in self._parse_json(
 343                     self._search_regex(
 344                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
 345                         webpage, 'video ids', group='ids'),
 346                     video_id)]
 347
 348             return self.playlist_result(entries, video_id)
 349         else:
 350             _, info_dict = self._extract_from_url(
 351                 self._VIDEO_PAGE_TEMPLATE % video_id,
 352                 video_id, fatal_if_no_video=True)
 353             return info_dict
 354
 355
 356 class FacebookPluginsVideoIE(InfoExtractor):
 357     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
 358
 359     _TESTS = [{
 360         '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',
 361         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
 362         'info_dict': {
 363             'id': '10154383743583686',
 364             'ext': 'mp4',
 365             'title': 'What to do during the haze?',
 366             'uploader': 'Gov.sg',
 367             'upload_date': '20160826',
 368             'timestamp': 1472184808,
 369         },
 370         'add_ie': [FacebookIE.ie_key()],
 371     }, {
 372         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
 373         'only_matching': True,
 374     }, {
 375         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
 376         'only_matching': True,
 377     }]
 378
 379     def _real_extract(self, url):
 380         return self.url_result(
 381             compat_urllib_parse_unquote(self._match_id(url)),
 382             FacebookIE.ie_key())