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