]> gitweb @ CieloNegro.org - youtube-dl.git/blob - youtube_dl/extractor/facebook.py
cdb093262a00387386a0a3243554f46386276733
[youtube-dl.git] / 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     @staticmethod
133     def _extract_url(webpage):
134         mobj = re.search(
135             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
136         if mobj is not None:
137             return mobj.group('url')
138
139         # Facebook API embed
140         # see https://developers.facebook.com/docs/plugins/embedded-video-player
141         mobj = re.search(r'''(?x)<div[^>]+
142                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
143                 data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
144         if mobj is not None:
145             return mobj.group('url')
146
147     def _login(self):
148         (useremail, password) = self._get_login_info()
149         if useremail is None:
150             return
151
152         login_page_req = sanitized_Request(self._LOGIN_URL)
153         self._set_cookie('facebook.com', 'locale', 'en_US')
154         login_page = self._download_webpage(login_page_req, None,
155                                             note='Downloading login page',
156                                             errnote='Unable to download login page')
157         lsd = self._search_regex(
158             r'<input type="hidden" name="lsd" value="([^"]*)"',
159             login_page, 'lsd')
160         lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')
161
162         login_form = {
163             'email': useremail,
164             'pass': password,
165             'lsd': lsd,
166             'lgnrnd': lgnrnd,
167             'next': 'http://facebook.com/home.php',
168             'default_persistent': '0',
169             'legacy_return': '1',
170             'timezone': '-60',
171             'trynum': '1',
172         }
173         request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))
174         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
175         try:
176             login_results = self._download_webpage(request, None,
177                                                    note='Logging in', errnote='unable to fetch login page')
178             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
179                 error = self._html_search_regex(
180                     r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
181                     login_results, 'login error', default=None, group='error')
182                 if error:
183                     raise ExtractorError('Unable to login: %s' % error, expected=True)
184                 self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
185                 return
186
187             fb_dtsg = self._search_regex(
188                 r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
189             h = self._search_regex(
190                 r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
191
192             if not fb_dtsg or not h:
193                 return
194
195             check_form = {
196                 'fb_dtsg': fb_dtsg,
197                 'h': h,
198                 'name_action_selected': 'dont_save',
199             }
200             check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))
201             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
202             check_response = self._download_webpage(check_req, None,
203                                                     note='Confirming login')
204             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
205                 self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')
206         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
207             self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))
208             return
209
210     def _real_initialize(self):
211         self._login()
212
213     def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
214         req = sanitized_Request(url)
215         req.add_header('User-Agent', self._CHROME_USER_AGENT)
216         webpage = self._download_webpage(req, video_id)
217
218         video_data = None
219
220         BEFORE = '{swf.addParam(param[0], param[1]);});'
221         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
222         PATTERN = re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER)
223
224         for m in re.findall(PATTERN, webpage):
225             swf_params = m.replace('\\\\', '\\').replace('\\"', '"')
226             data = dict(json.loads(swf_params))
227             params_raw = compat_urllib_parse_unquote(data['params'])
228             video_data_candidate = json.loads(params_raw)['video_data']
229             for _, f in video_data_candidate.items():
230                 if not f:
231                     continue
232                 if isinstance(f, dict):
233                     f = [f]
234                 if not isinstance(f, list):
235                     continue
236                 if f[0].get('video_id') == video_id:
237                     video_data = video_data_candidate
238                     break
239             if video_data:
240                 break
241
242         def video_data_list2dict(video_data):
243             ret = {}
244             for item in video_data:
245                 format_id = item['stream_type']
246                 ret.setdefault(format_id, []).append(item)
247             return ret
248
249         if not video_data:
250             server_js_data = self._parse_json(self._search_regex(
251                 r'handleServerJS\(({.+})\);', webpage, 'server js data', default='{}'), video_id)
252             for item in server_js_data.get('instances', []):
253                 if item[1][0] == 'VideoConfig':
254                     video_data = video_data_list2dict(item[2][0]['videoData'])
255                     break
256
257         if not video_data:
258             if not fatal_if_no_video:
259                 return webpage, False
260             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
261             if m_msg is not None:
262                 raise ExtractorError(
263                     'The video is not available, Facebook said: "%s"' % m_msg.group(1),
264                     expected=True)
265             else:
266                 raise ExtractorError('Cannot parse data')
267
268         formats = []
269         for format_id, f in video_data.items():
270             if f and isinstance(f, dict):
271                 f = [f]
272             if not f or not isinstance(f, list):
273                 continue
274             for quality in ('sd', 'hd'):
275                 for src_type in ('src', 'src_no_ratelimit'):
276                     src = f[0].get('%s_%s' % (quality, src_type))
277                     if src:
278                         preference = -10 if format_id == 'progressive' else 0
279                         if quality == 'hd':
280                             preference += 5
281                         formats.append({
282                             'format_id': '%s_%s_%s' % (format_id, quality, src_type),
283                             'url': src,
284                             'preference': preference,
285                         })
286             dash_manifest = f[0].get('dash_manifest')
287             if dash_manifest:
288                 formats.extend(self._parse_mpd_formats(
289                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
290         if not formats:
291             raise ExtractorError('Cannot find video formats')
292
293         self._sort_formats(formats)
294
295         video_title = self._html_search_regex(
296             r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
297             default=None)
298         if not video_title:
299             video_title = self._html_search_regex(
300                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
301                 webpage, 'alternative title', default=None)
302             video_title = limit_length(video_title, 80)
303         if not video_title:
304             video_title = 'Facebook video #%s' % video_id
305         uploader = clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))
306
307         info_dict = {
308             'id': video_id,
309             'title': video_title,
310             'formats': formats,
311             'uploader': uploader,
312         }
313
314         return webpage, info_dict
315
316     def _real_extract(self, url):
317         video_id = self._match_id(url)
318
319         real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
320         webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
321
322         if info_dict:
323             return info_dict
324
325         if '/posts/' in url:
326             entries = [
327                 self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
328                 for vid in self._parse_json(
329                     self._search_regex(
330                         r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
331                         webpage, 'video ids', group='ids'),
332                     video_id)]
333
334             return self.playlist_result(entries, video_id)
335         else:
336             _, info_dict = self._extract_from_url(
337                 self._VIDEO_PAGE_TEMPLATE % video_id,
338                 video_id, fatal_if_no_video=True)
339             return info_dict