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     limit_length,
  19     sanitized_Request,
  20     urlencode_postdata,
  21     get_element_by_id,
  22     clean_html,
  23 )
  24
  25
  26 class FacebookIE(InfoExtractor):
  27     _VALID_URL = r'''(?x)
  28                 (?:
  29                     https?://
  30                         (?:\w+\.)?facebook\.com/
  31                         (?:[^#]*?\#!/)?
  32                         (?:
  33                             (?:
  34                                 video/video\.php|
  35                                 photo\.php|
  36                                 video\.php|
  37                                 video/embed|
  38                                 story\.php
  39                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
  40                             [^/]+/videos/(?:[^/]+/)?|
  41                             [^/]+/posts/
  42                         )|
  43                     facebook:
  44                 )
  45                 (?P<id>[0-9]+)
  46                 '''
  47     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
  48     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
  49     _NETRC_MACHINE = 'facebook'
  50     IE_NAME = 'facebook'
  51
  52     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
  53
  54     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
  55
  56     _TESTS = [{
  57         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
  58         'md5': '6a40d33c0eccbb1af76cf0485a052659',
  59         'info_dict': {
  60             'id': '637842556329505',
  61             'ext': 'mp4',
  62             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
  63             'uploader': 'Tennis on Facebook',
  64         }
  65     }, {
  66         'note': 'Video without discernible title',
  67         'url': 'https://www.facebook.com/video.php?v=274175099429670',
  68         'info_dict': {
  69             'id': '274175099429670',
  70             'ext': 'mp4',
  71             'title': 'Facebook video #274175099429670',
  72             'uploader': 'Asif Nawab Butt',
  73         },
  74         'expected_warnings': [
  75             'title'
  76         ]
  77     }, {
  78         'note': 'Video with DASH manifest',
  79         'url': 'https://www.facebook.com/video.php?v=957955867617029',
  80         'md5': '54706e4db4f5ad58fbad82dde1f1213f',
  81         'info_dict': {
  82             'id': '957955867617029',
  83             'ext': 'mp4',
  84             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
  85             'uploader': 'Demy de Zeeuw',
  86         },
  87     }, {
  88         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
  89         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
  90         'info_dict': {
  91             'id': '544765982287235',
  92             'ext': 'mp4',
  93             'title': '"What are you doing running in the snow?"',
  94             'uploader': 'FailArmy',
  95         }
  96     }, {
  97         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
  98         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
  99         'info_dict': {
 100             'id': '1035862816472149',
 101             'ext': 'mp4',
 102             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
 103             'uploader': 'S. Saint',
 104         },
 105     }, {
 106         'note': 'swf params escaped',
 107         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
 108         'md5': '97ba073838964d12c70566e0085c2b91',
 109         'info_dict': {
 110             'id': '10153664894881749',
 111             'ext': 'mp4',
 112             'title': 'Facebook video #10153664894881749',
 113         },
 114     }, {
 115         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
 116         'only_matching': True,
 117     }, {
 118         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
 119         'only_matching': True,
 120     }, {
 121         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
 122         'only_matching': True,
 123     }, {
 124         'url': 'facebook:544765982287235',
 125         'only_matching': True,
 126     }]
 127
 128     def _login(self):
 129         (useremail, password) = self._get_login_info()
 130         if useremail is None:
 131             return
 132
 133         login_page_req = sanitized_Request(self._LOGIN_URL)
 134         self._set_cookie('facebook.com', 'locale', 'en_US')
 135         login_page = self._download_webpage(login_page_req, None,
 136                                             note='Downloading login page',
 137                                             errnote='Unable to download login page')
 138         lsd = self._search_regex(
 139             r'<input type="hidden" name="lsd" value="([^"]*)"',
 140             login_page, 'lsd')
 141         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
 142
 143         login_form = {
 144             'email': useremail,
 145             'pass': password,
 146             'lsd': lsd,
 147             'lgnrnd': lgnrnd,
 148             'next': 'http://facebook.com/home.php',
 149             'default_persistent': '0',
 150             'legacy_return': '1',
 151             'timezone': '-60',
 152             'trynum': '1',
 153         }
 154         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
 155         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
 156         try:
 157             login_results = self._download_webpage(request, None,
 158                                                    note='Logging in', errnote='unable to fetch login page')
 159             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
 160                 error = self._html_search_regex(
 161                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
 162                     login_results, 'login error', default=None, group='error')
 163                 if error:
 164                     raise ExtractorError('Unable to login: %s' % error, expected=True)
 165                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
 166                 return
 167
 168             fb_dtsg = self._search_regex(
 169                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
 170             h = self._search_regex(
 171                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
 172
 173             if not fb_dtsg or not h:
 174                 return
 175
 176             check_form = {
 177                 'fb_dtsg': fb_dtsg,
 178                 'h': h,
 179                 'name_action_selected': 'dont_save',
 180             }
 181             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
 182             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
 183             check_response = self._download_webpage(check_req, None,
 184                                                     note='Confirming login')
 185             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
 186                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
 187         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 188             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
 189             return
 190
 191     def _real_initialize(self):
 192         self._login()
 193
 194     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
 195         req = sanitized_Request(url)
 196         req.add_header('User-Agent', self._CHROME_USER_AGENT)
 197         webpage = self._download_webpage(req, video_id)
 198
 199         video_data = None
 200
 201         BEFORE = '{swf.addParam(param[0], param[1]);});'
 202         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
 203         m = re.search(re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER), webpage)
 204         if m:
 205             swf_params = m.group(1).replace('\\\\', '\\').replace('\\"', '"')
 206             data = dict(json.loads(swf_params))
 207             params_raw = compat_urllib_parse_unquote(data['params'])
 208             video_data = json.loads(params_raw)['video_data']
 209
 210         def video_data_list2dict(video_data):
 211             ret = {}
 212             for item in video_data:
 213                 format_id = item['stream_type']
 214                 ret.setdefault(format_id, []).append(item)
 215             return ret
 216
 217         if not video_data:
 218             server_js_data = self._parse_json(self._search_regex(
 219                 r'handleServerJS\(({.+})\);', webpage, 'server js data', default='{}'), video_id)
 220             for item in server_js_data.get('instances', []):
 221                 if item[1][0] == 'VideoConfig':
 222                     video_data = video_data_list2dict(item[2][0]['videoData'])
 223                     break
 224
 225         if not video_data:
 226             if not fatal_if_no_video:
 227                 return webpage, False
 228             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
 229             if m_msg is not None:
 230                 raise ExtractorError(
 231                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
 232                     expected=True)
 233             else:
 234                 raise ExtractorError('Cannot parse data')
 235
 236         formats = []
 237         for format_id, f in video_data.items():
 238             if not f or not isinstance(f, list):
 239                 continue
 240             for quality in ('sd', 'hd'):
 241                 for src_type in ('src', 'src_no_ratelimit'):
 242                     src = f[0].get('%s_%s' % (quality, src_type))
 243                     if src:
 244                         preference = -10 if format_id == 'progressive' else 0
 245                         if quality == 'hd':
 246                             preference += 5
 247                         formats.append({
 248                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
 249                             'url': src,
 250                             'preference': preference,
 251                         })
 252             dash_manifest = f[0].get('dash_manifest')
 253             if dash_manifest:
 254                 formats.extend(self._parse_mpd_formats(
 255                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
 256         if not formats:
 257             raise ExtractorError('Cannot find video formats')
 258
 259         self._sort_formats(formats)
 260
 261         video_title = self._html_search_regex(
 262             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
 263             default=None)
 264         if not video_title:
 265             video_title = self._html_search_regex(
 266                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
 267                 webpage, 'alternative title', default=None)
 268             video_title = limit_length(video_title, 80)
 269         if not video_title:
 270             video_title = 'Facebook video #%s' % video_id
 271         uploader = clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
 272
 273         info_dict = {
 274             'id': video_id,
 275             'title': video_title,
 276             'formats': formats,
 277             'uploader': uploader,
 278         }
 279
 280         return webpage, info_dict
 281
 282     def _real_extract(self, url):
 283         video_id = self._match_id(url)
 284
 285         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
 286         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
 287
 288         if info_dict:
 289             return info_dict
 290
 291         if '/posts/' in url:
 292             entries = [
 293                 self.url_result('facebook:%s' % video_id, FacebookIE.ie_key())
 294                 for video_id in self._parse_json(
 295                     self._search_regex(
 296                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
 297                         webpage, 'video ids', group='ids'),
 298                     video_id)]
 299
 300             return self.playlist_result(entries, video_id)
 301         else:
 302             _, info_dict = self._extract_from_url(
 303                 self._VIDEO_PAGE_TEMPLATE % video_id,
 304                 video_id, fatal_if_no_video=True)
 305             return info_dict