]> gitweb @ CieloNegro.org - youtube-dl.git/blob - youtube_dl/extractor/facebook.py
[facebook] Fix title extraction (closes #11941)
[youtube-dl.git] / youtube_dl / extractor / facebook.py
1 from __future__ import unicode_literals
2
3 import re
4 import socket
5
6 from .common import InfoExtractor
7 from ..compat import (
8     compat_etree_fromstring,
9     compat_http_client,
10     compat_urllib_error,
11     compat_urllib_parse_unquote,
12     compat_urllib_parse_unquote_plus,
13 )
14 from ..utils import (
15     clean_html,
16     error_to_compat_str,
17     ExtractorError,
18     get_element_by_id,
19     int_or_none,
20     js_to_json,
21     limit_length,
22     sanitized_Request,
23     try_get,
24     urlencode_postdata,
25 )
26
27
28 class FacebookIE(InfoExtractor):
29     _VALID_URL = r'''(?x)
30                 (?:
31                     https?://
32                         (?:[\w-]+\.)?(?:facebook\.com|facebookcorewwwi\.onion)/
33                         (?:[^#]*?\#!/)?
34                         (?:
35                             (?:
36                                 video/video\.php|
37                                 photo\.php|
38                                 video\.php|
39                                 video/embed|
40                                 story\.php
41                             )\?(?:.*?)(?:v|video_id|story_fbid)=|
42                             [^/]+/videos/(?:[^/]+/)?|
43                             [^/]+/posts/|
44                             groups/[^/]+/permalink/
45                         )|
46                     facebook:
47                 )
48                 (?P<id>[0-9]+)
49                 '''
50     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
51     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
52     _NETRC_MACHINE = 'facebook'
53     IE_NAME = 'facebook'
54
55     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
56
57     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
58
59     _TESTS = [{
60         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
61         'md5': '6a40d33c0eccbb1af76cf0485a052659',
62         'info_dict': {
63             'id': '637842556329505',
64             'ext': 'mp4',
65             'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
66             'uploader': 'Tennis on Facebook',
67             'upload_date': '20140908',
68             'timestamp': 1410199200,
69         }
70     }, {
71         'note': 'Video without discernible title',
72         'url': 'https://www.facebook.com/video.php?v=274175099429670',
73         'info_dict': {
74             'id': '274175099429670',
75             'ext': 'mp4',
76             'title': 'Asif Nawab Butt posted a video to his Timeline.',
77             'uploader': 'Asif Nawab Butt',
78             'upload_date': '20140506',
79             'timestamp': 1399398998,
80         },
81         'expected_warnings': [
82             'title'
83         ]
84     }, {
85         'note': 'Video with DASH manifest',
86         'url': 'https://www.facebook.com/video.php?v=957955867617029',
87         'md5': 'b2c28d528273b323abe5c6ab59f0f030',
88         'info_dict': {
89             'id': '957955867617029',
90             'ext': 'mp4',
91             'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
92             'uploader': 'Demy de Zeeuw',
93             'upload_date': '20160110',
94             'timestamp': 1452431627,
95         },
96     }, {
97         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
98         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
99         'info_dict': {
100             'id': '544765982287235',
101             'ext': 'mp4',
102             'title': '"What are you doing running in the snow?"',
103             'uploader': 'FailArmy',
104         },
105         'skip': 'Video gone',
106     }, {
107         'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
108         'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
109         'info_dict': {
110             'id': '1035862816472149',
111             'ext': 'mp4',
112             'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
113             'uploader': 'S. Saint',
114         },
115         'skip': 'Video gone',
116     }, {
117         'note': 'swf params escaped',
118         'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
119         'md5': '97ba073838964d12c70566e0085c2b91',
120         'info_dict': {
121             'id': '10153664894881749',
122             'ext': 'mp4',
123             'title': 'Facebook video #10153664894881749',
124         },
125     }, {
126         # have 1080P, but only up to 720p in swf params
127         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
128         'md5': '0d9813160b146b3bc8744e006027fcc6',
129         'info_dict': {
130             'id': '10155529876156509',
131             'ext': 'mp4',
132             'title': 'Holocaust survivor becomes US citizen',
133             'timestamp': 1477818095,
134             'upload_date': '20161030',
135             'uploader': 'CNN',
136         },
137     }, {
138         'url': 'https://www.facebook.com/video.php?v=10204634152394104',
139         'only_matching': True,
140     }, {
141         'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
142         'only_matching': True,
143     }, {
144         'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
145         'only_matching': True,
146     }, {
147         'url': 'facebook:544765982287235',
148         'only_matching': True,
149     }, {
150         'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
151         'only_matching': True,
152     }, {
153         'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
154         'only_matching': True,
155     }, {
156         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
157         'only_matching': True,
158     }]
159
160     @staticmethod
161     def _extract_url(webpage):
162         mobj = re.search(
163             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
164         if mobj is not None:
165             return mobj.group('url')
166
167         # Facebook API embed
168         # see https://developers.facebook.com/docs/plugins/embedded-video-player
169         mobj = re.search(r'''(?x)<div[^>]+
170                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
171                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
172         if mobj is not None:
173             return mobj.group('url')
174
175     def _login(self):
176         (useremail, password) = self._get_login_info()
177         if useremail is None:
178             return
179
180         login_page_req = sanitized_Request(self._LOGIN_URL)
181         self._set_cookie('facebook.com', 'locale', 'en_US')
182         login_page = self._download_webpage(login_page_req, None,
183                                             note='Downloading login page',
184                                             errnote='Unable to download login page')
185         lsd = self._search_regex(
186             r'<input type="hidden" name="lsd" value="([^"]*)"',
187             login_page, 'lsd')
188         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
189
190         login_form = {
191             'email': useremail,
192             'pass': password,
193             'lsd': lsd,
194             'lgnrnd': lgnrnd,
195             'next': 'http://facebook.com/home.php',
196             'default_persistent': '0',
197             'legacy_return': '1',
198             'timezone': '-60',
199             'trynum': '1',
200         }
201         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
202         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
203         try:
204             login_results = self._download_webpage(request, None,
205                                                    note='Logging in', errnote='unable to fetch login page')
206             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
207                 error = self._html_search_regex(
208                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
209                     login_results, 'login error', default=None, group='error')
210                 if error:
211                     raise ExtractorError('Unable to login: %s' % error, expected=True)
212                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
213                 return
214
215             fb_dtsg = self._search_regex(
216                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
217             h = self._search_regex(
218                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
219
220             if not fb_dtsg or not h:
221                 return
222
223             check_form = {
224                 'fb_dtsg': fb_dtsg,
225                 'h': h,
226                 'name_action_selected': 'dont_save',
227             }
228             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
229             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
230             check_response = self._download_webpage(check_req, None,
231                                                     note='Confirming login')
232             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
233                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
234         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
235             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
236             return
237
238     def _real_initialize(self):
239         self._login()
240
241     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
242         req = sanitized_Request(url)
243         req.add_header('User-Agent', self._CHROME_USER_AGENT)
244         webpage = self._download_webpage(req, video_id)
245
246         video_data = None
247
248         def extract_video_data(instances):
249             for item in instances:
250                 if item[1][0] == 'VideoConfig':
251                     video_item = item[2][0]
252                     if video_item.get('video_id') == video_id:
253                         return video_item['videoData']
254
255         server_js_data = self._parse_json(self._search_regex(
256             r'handleServerJS\(({.+})(?:\);|,")', webpage,
257             'server js data', default='{}'), video_id, fatal=False)
258
259         if server_js_data:
260             video_data = extract_video_data(server_js_data.get('instances', []))
261
262         if not video_data:
263             server_js_data = self._parse_json(
264                 self._search_regex(
265                     r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+stream_pagelet',
266                     webpage, 'js data', default='{}'),
267                 video_id, transform_source=js_to_json, fatal=False)
268             if server_js_data:
269                 video_data = extract_video_data(try_get(
270                     server_js_data, lambda x: x['jsmods']['instances'],
271                     list) or [])
272
273         if not video_data:
274             if not fatal_if_no_video:
275                 return webpage, False
276             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
277             if m_msg is not None:
278                 raise ExtractorError(
279                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
280                     expected=True)
281             elif '>You must log in to continue' in webpage:
282                 self.raise_login_required()
283             else:
284                 raise ExtractorError('Cannot parse data')
285
286         formats = []
287         for f in video_data:
288             format_id = f['stream_type']
289             if f and isinstance(f, dict):
290                 f = [f]
291             if not f or not isinstance(f, list):
292                 continue
293             for quality in ('sd', 'hd'):
294                 for src_type in ('src', 'src_no_ratelimit'):
295                     src = f[0].get('%s_%s' % (quality, src_type))
296                     if src:
297                         preference = -10 if format_id == 'progressive' else 0
298                         if quality == 'hd':
299                             preference += 5
300                         formats.append({
301                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
302                             'url': src,
303                             'preference': preference,
304                         })
305             dash_manifest = f[0].get('dash_manifest')
306             if dash_manifest:
307                 formats.extend(self._parse_mpd_formats(
308                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
309         if not formats:
310             raise ExtractorError('Cannot find video formats')
311
312         self._sort_formats(formats)
313
314         video_title = self._html_search_regex(
315             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
316             default=None)
317         if not video_title:
318             video_title = self._html_search_regex(
319                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
320                 webpage, 'alternative title', default=None)
321         if not video_title:
322             video_title = self._html_search_meta(
323                 'description', webpage, 'title')
324         if video_title:
325             video_title = limit_length(video_title, 80)
326         else:
327             video_title = 'Facebook video #%s' % video_id
328         uploader = clean_html(get_element_by_id(
329             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
330             r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader', fatal=False)
331         timestamp = int_or_none(self._search_regex(
332             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
333             'timestamp', default=None))
334
335         info_dict = {
336             'id': video_id,
337             'title': video_title,
338             'formats': formats,
339             'uploader': uploader,
340             'timestamp': timestamp,
341         }
342
343         return webpage, info_dict
344
345     def _real_extract(self, url):
346         video_id = self._match_id(url)
347
348         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
349         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
350
351         if info_dict:
352             return info_dict
353
354         if '/posts/' in url:
355             entries = [
356                 self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
357                 for vid in self._parse_json(
358                     self._search_regex(
359                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
360                         webpage, 'video ids', group='ids'),
361                     video_id)]
362
363             return self.playlist_result(entries, video_id)
364         else:
365             _, info_dict = self._extract_from_url(
366                 self._VIDEO_PAGE_TEMPLATE % video_id,
367                 video_id, fatal_if_no_video=True)
368             return info_dict
369
370
371 class FacebookPluginsVideoIE(InfoExtractor):
372     _VALID_URL = r'https?://(?:[\w-]+\.)?facebook\.com/plugins/video\.php\?.*?\bhref=(?P<id>https.+)'
373
374     _TESTS = [{
375         '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',
376         'md5': '5954e92cdfe51fe5782ae9bda7058a07',
377         'info_dict': {
378             'id': '10154383743583686',
379             'ext': 'mp4',
380             'title': 'What to do during the haze?',
381             'uploader': 'Gov.sg',
382             'upload_date': '20160826',
383             'timestamp': 1472184808,
384         },
385         'add_ie': [FacebookIE.ie_key()],
386     }, {
387         'url': 'https://www.facebook.com/plugins/video.php?href=https%3A%2F%2Fwww.facebook.com%2Fvideo.php%3Fv%3D10204634152394104',
388         'only_matching': True,
389     }, {
390         'url': 'https://www.facebook.com/plugins/video.php?href=https://www.facebook.com/gov.sg/videos/10154383743583686/&show_text=0&width=560',
391         'only_matching': True,
392     }]
393
394     def _real_extract(self, url):
395         return self.url_result(
396             compat_urllib_parse_unquote(self._match_id(url)),
397             FacebookIE.ie_key())