]> git.bitcoin.ninja Git - youtube-dl/blob - youtube_dl/extractor/facebook.py
[brightcove:new] extract subtitles and strip video title
[youtube-dl] / 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                             groups/[^/]+/permalink/
43                         )|
44                     facebook:
45                 )
46                 (?P<id>[0-9]+)
47                 '''
48     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
49     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
50     _NETRC_MACHINE = 'facebook'
51     IE_NAME = 'facebook'
52
53     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
54
55     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
56
57     _TESTS = [{
58         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
59         'md5': '6a40d33c0eccbb1af76cf0485a052659',
60         'info_dict': {
61             'id': '637842556329505',
62             'ext': 'mp4',
63             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
64             'uploader': 'Tennis on Facebook',
65         }
66     }, {
67         'note': 'Video without discernible title',
68         'url': 'https://www.facebook.com/video.php?v=274175099429670',
69         'info_dict': {
70             'id': '274175099429670',
71             'ext': 'mp4',
72             'title': 'Facebook video #274175099429670',
73             'uploader': 'Asif Nawab Butt',
74         },
75         'expected_warnings': [
76             'title'
77         ]
78     }, {
79         'note': 'Video with DASH manifest',
80         'url': 'https://www.facebook.com/video.php?v=957955867617029',
81         'md5': '54706e4db4f5ad58fbad82dde1f1213f',
82         'info_dict': {
83             'id': '957955867617029',
84             'ext': 'mp4',
85             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
86             'uploader': 'Demy de Zeeuw',
87         },
88     }, {
89         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
90         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
91         'info_dict': {
92             'id': '544765982287235',
93             'ext': 'mp4',
94             'title': '"What are you doing running in the snow?"',
95             'uploader': 'FailArmy',
96         }
97     }, {
98         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
99         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
100         'info_dict': {
101             'id': '1035862816472149',
102             'ext': 'mp4',
103             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
104             'uploader': 'S. Saint',
105         },
106     }, {
107         'note': 'swf params escaped',
108         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
109         'md5': '97ba073838964d12c70566e0085c2b91',
110         'info_dict': {
111             'id': '10153664894881749',
112             'ext': 'mp4',
113             'title': 'Facebook video #10153664894881749',
114         },
115     }, {
116         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
117         'only_matching': True,
118     }, {
119         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
120         'only_matching': True,
121     }, {
122         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
123         'only_matching': True,
124     }, {
125         'url': 'facebook:544765982287235',
126         'only_matching': True,
127     }, {
128         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
129         'only_matching': True,
130     }]
131
132     def _login(self):
133         (useremail, password) = self._get_login_info()
134         if useremail is None:
135             return
136
137         login_page_req = sanitized_Request(self._LOGIN_URL)
138         self._set_cookie('facebook.com', 'locale', 'en_US')
139         login_page = self._download_webpage(login_page_req, None,
140                                             note='Downloading login page',
141                                             errnote='Unable to download login page')
142         lsd = self._search_regex(
143             r'<input type="hidden" name="lsd" value="([^"]*)"',
144             login_page, 'lsd')
145         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
146
147         login_form = {
148             'email': useremail,
149             'pass': password,
150             'lsd': lsd,
151             'lgnrnd': lgnrnd,
152             'next': 'http://facebook.com/home.php',
153             'default_persistent': '0',
154             'legacy_return': '1',
155             'timezone': '-60',
156             'trynum': '1',
157         }
158         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
159         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
160         try:
161             login_results = self._download_webpage(request, None,
162                                                    note='Logging in', errnote='unable to fetch login page')
163             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
164                 error = self._html_search_regex(
165                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
166                     login_results, 'login error', default=None, group='error')
167                 if error:
168                     raise ExtractorError('Unable to login: %s' % error, expected=True)
169                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
170                 return
171
172             fb_dtsg = self._search_regex(
173                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
174             h = self._search_regex(
175                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
176
177             if not fb_dtsg or not h:
178                 return
179
180             check_form = {
181                 'fb_dtsg': fb_dtsg,
182                 'h': h,
183                 'name_action_selected': 'dont_save',
184             }
185             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
186             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
187             check_response = self._download_webpage(check_req, None,
188                                                     note='Confirming login')
189             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
190                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
191         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
192             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
193             return
194
195     def _real_initialize(self):
196         self._login()
197
198     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
199         req = sanitized_Request(url)
200         req.add_header('User-Agent', self._CHROME_USER_AGENT)
201         webpage = self._download_webpage(req, video_id)
202
203         video_data = None
204
205         BEFORE = '{swf.addParam(param[0], param[1]);});'
206         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
207         m = re.search(re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER), webpage)
208         if m:
209             swf_params = m.group(1).replace('\\\\', '\\').replace('\\"', '"')
210             data = dict(json.loads(swf_params))
211             params_raw = compat_urllib_parse_unquote(data['params'])
212             video_data = json.loads(params_raw)['video_data']
213
214         def video_data_list2dict(video_data):
215             ret = {}
216             for item in video_data:
217                 format_id = item['stream_type']
218                 ret.setdefault(format_id, []).append(item)
219             return ret
220
221         if not video_data:
222             server_js_data = self._parse_json(self._search_regex(
223                 r'handleServerJS\(({.+})\);', webpage, 'server js data', default='{}'), video_id)
224             for item in server_js_data.get('instances', []):
225                 if item[1][0] == 'VideoConfig':
226                     video_data = video_data_list2dict(item[2][0]['videoData'])
227                     break
228
229         if not video_data:
230             if not fatal_if_no_video:
231                 return webpage, False
232             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
233             if m_msg is not None:
234                 raise ExtractorError(
235                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
236                     expected=True)
237             else:
238                 raise ExtractorError('Cannot parse data')
239
240         formats = []
241         for format_id, f in video_data.items():
242             if not f or not isinstance(f, list):
243                 continue
244             for quality in ('sd', 'hd'):
245                 for src_type in ('src', 'src_no_ratelimit'):
246                     src = f[0].get('%s_%s' % (quality, src_type))
247                     if src:
248                         preference = -10 if format_id == 'progressive' else 0
249                         if quality == 'hd':
250                             preference += 5
251                         formats.append({
252                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
253                             'url': src,
254                             'preference': preference,
255                         })
256             dash_manifest = f[0].get('dash_manifest')
257             if dash_manifest:
258                 formats.extend(self._parse_mpd_formats(
259                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
260         if not formats:
261             raise ExtractorError('Cannot find video formats')
262
263         self._sort_formats(formats)
264
265         video_title = self._html_search_regex(
266             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
267             default=None)
268         if not video_title:
269             video_title = self._html_search_regex(
270                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
271                 webpage, 'alternative title', default=None)
272             video_title = limit_length(video_title, 80)
273         if not video_title:
274             video_title = 'Facebook video #%s' % video_id
275         uploader = clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
276
277         info_dict = {
278             'id': video_id,
279             'title': video_title,
280             'formats': formats,
281             'uploader': uploader,
282         }
283
284         return webpage, info_dict
285
286     def _real_extract(self, url):
287         video_id = self._match_id(url)
288
289         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
290         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
291
292         if info_dict:
293             return info_dict
294
295         if '/posts/' in url:
296             entries = [
297                 self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
298                 for vid in self._parse_json(
299                     self._search_regex(
300                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
301                         webpage, 'video ids', group='ids'),
302                     video_id)]
303
304             return self.playlist_result(entries, video_id)
305         else:
306             _, info_dict = self._extract_from_url(
307                 self._VIDEO_PAGE_TEMPLATE % video_id,
308                 video_id, fatal_if_no_video=True)
309             return info_dict