]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/facebook.py
Initiate new release.
[youtubedl] / youtube_dl / extractor / facebook.py
index 6315d40c50a700c741db25c5f04492bdae1ac809..610d6674592384922f9df7af4da5958592ce56bd 100644 (file)
@@ -20,6 +20,7 @@ from ..utils import (
     int_or_none,
     js_to_json,
     limit_length,
     int_or_none,
     js_to_json,
     limit_length,
+    parse_count,
     sanitized_Request,
     try_get,
     urlencode_postdata,
     sanitized_Request,
     try_get,
     urlencode_postdata,
@@ -56,6 +57,7 @@ class FacebookIE(InfoExtractor):
     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
 
     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
     _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
 
     _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
+    _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true&payloadtype=primary'
 
     _TESTS = [{
         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
 
     _TESTS = [{
         'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
@@ -67,17 +69,18 @@ class FacebookIE(InfoExtractor):
             'uploader': 'Tennis on Facebook',
             'upload_date': '20140908',
             'timestamp': 1410199200,
             'uploader': 'Tennis on Facebook',
             'upload_date': '20140908',
             'timestamp': 1410199200,
-        }
+        },
+        'skip': 'Requires logging in',
     }, {
     }, {
-        'note': 'Video without discernible title',
         'url': 'https://www.facebook.com/video.php?v=274175099429670',
         'info_dict': {
             'id': '274175099429670',
             'ext': 'mp4',
         'url': 'https://www.facebook.com/video.php?v=274175099429670',
         'info_dict': {
             'id': '274175099429670',
             'ext': 'mp4',
-            'title': 'Asif Nawab Butt posted a video to his Timeline.',
+            'title': 're:^Asif Nawab Butt posted a video',
             'uploader': 'Asif Nawab Butt',
             'upload_date': '20140506',
             'timestamp': 1399398998,
             'uploader': 'Asif Nawab Butt',
             'upload_date': '20140506',
             'timestamp': 1399398998,
+            'thumbnail': r're:^https?://.*',
         },
         'expected_warnings': [
             'title'
         },
         'expected_warnings': [
             'title'
@@ -94,6 +97,7 @@ class FacebookIE(InfoExtractor):
             'upload_date': '20160110',
             'timestamp': 1452431627,
         },
             'upload_date': '20160110',
             'timestamp': 1452431627,
         },
+        'skip': 'Requires logging in',
     }, {
         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
     }, {
         'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
         'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
@@ -121,19 +125,25 @@ class FacebookIE(InfoExtractor):
         'info_dict': {
             'id': '10153664894881749',
             'ext': 'mp4',
         'info_dict': {
             'id': '10153664894881749',
             'ext': 'mp4',
-            'title': 'Facebook video #10153664894881749',
+            'title': 'Average time to confirm recent Supreme Court nominees: 67 days Longest it\'s t...',
+            'thumbnail': r're:^https?://.*',
+            'timestamp': 1456259628,
+            'upload_date': '20160223',
+            'uploader': 'Barack Obama',
         },
     }, {
         # have 1080P, but only up to 720p in swf params
         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
         },
     }, {
         # have 1080P, but only up to 720p in swf params
         'url': 'https://www.facebook.com/cnn/videos/10155529876156509/',
-        'md5': '0d9813160b146b3bc8744e006027fcc6',
+        'md5': '9571fae53d4165bbbadb17a94651dcdc',
         'info_dict': {
             'id': '10155529876156509',
             'ext': 'mp4',
         'info_dict': {
             'id': '10155529876156509',
             'ext': 'mp4',
-            'title': 'Holocaust survivor becomes US citizen',
+            'title': 'She survived the holocaust — and years later, she’s getting her citizenship s...',
             'timestamp': 1477818095,
             'upload_date': '20161030',
             'uploader': 'CNN',
             'timestamp': 1477818095,
             'upload_date': '20161030',
             'uploader': 'CNN',
+            'thumbnail': r're:^https?://.*',
+            'view_count': int,
         },
     }, {
         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
         },
     }, {
         # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall
@@ -141,7 +151,7 @@ class FacebookIE(InfoExtractor):
         'info_dict': {
             'id': '1417995061575415',
             'ext': 'mp4',
         'info_dict': {
             'id': '1417995061575415',
             'ext': 'mp4',
-            'title': 'md5:a7b86ca673f51800cd54687b7f4012fe',
+            'title': 'md5:1db063d6a8c13faa8da727817339c857',
             'timestamp': 1486648217,
             'upload_date': '20170209',
             'uploader': 'Yaroslav Korpan',
             'timestamp': 1486648217,
             'upload_date': '20170209',
             'uploader': 'Yaroslav Korpan',
@@ -158,6 +168,7 @@ class FacebookIE(InfoExtractor):
             'timestamp': 1477305000,
             'upload_date': '20161024',
             'uploader': 'La Guía Del Varón',
             'timestamp': 1477305000,
             'upload_date': '20161024',
             'uploader': 'La Guía Del Varón',
+            'thumbnail': r're:^https?://.*',
         },
         'params': {
             'skip_download': True,
         },
         'params': {
             'skip_download': True,
@@ -167,7 +178,7 @@ class FacebookIE(InfoExtractor):
         'info_dict': {
             'id': '1396382447100162',
             'ext': 'mp4',
         'info_dict': {
             'id': '1396382447100162',
             'ext': 'mp4',
-            'title': 'md5:e2d2700afdf84e121f5d0f999bad13a3',
+            'title': 'md5:19a428bbde91364e3de815383b54a235',
             'timestamp': 1486035494,
             'upload_date': '20170202',
             'uploader': 'Elisabeth Ahtn',
             'timestamp': 1486035494,
             'upload_date': '20170202',
             'uploader': 'Elisabeth Ahtn',
@@ -196,25 +207,40 @@ class FacebookIE(InfoExtractor):
     }, {
         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
         'only_matching': True,
     }, {
         'url': 'https://www.facebookcorewwwi.onion/video.php?v=274175099429670',
         'only_matching': True,
+    }, {
+        # no title
+        'url': 'https://www.facebook.com/onlycleverentertainment/videos/1947995502095005/',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.facebook.com/WatchESLOne/videos/359649331226507/',
+        'info_dict': {
+            'id': '359649331226507',
+            'ext': 'mp4',
+            'title': '#ESLOne VoD - Birmingham Finals Day#1 Fnatic vs. @Evil Geniuses',
+            'uploader': 'ESL One Dota 2',
+        },
+        'params': {
+            'skip_download': True,
+        },
     }]
 
     @staticmethod
     }]
 
     @staticmethod
-    def _extract_url(webpage):
-        mobj = re.search(
-            r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
-        if mobj is not None:
-            return mobj.group('url')
-
+    def _extract_urls(webpage):
+        urls = []
+        for mobj in re.finditer(
+                r'<iframe[^>]+?src=(["\'])(?P<url>https?://www\.facebook\.com/(?:video/embed|plugins/video\.php).+?)\1',
+                webpage):
+            urls.append(mobj.group('url'))
         # Facebook API embed
         # see https://developers.facebook.com/docs/plugins/embedded-video-player
         # Facebook API embed
         # see https://developers.facebook.com/docs/plugins/embedded-video-player
-        mobj = re.search(r'''(?x)<div[^>]+
+        for mobj in re.finditer(r'''(?x)<div[^>]+
                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
                 class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
-                data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
-        if mobj is not None:
-            return mobj.group('url')
+                data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage):
+            urls.append(mobj.group('url'))
+        return urls
 
     def _login(self):
 
     def _login(self):
-        (useremail, password) = self._get_login_info()
+        useremail, password = self._get_login_info()
         if useremail is None:
             return
 
         if useremail is None:
             return
 
@@ -300,16 +326,18 @@ class FacebookIE(InfoExtractor):
         if server_js_data:
             video_data = extract_video_data(server_js_data.get('instances', []))
 
         if server_js_data:
             video_data = extract_video_data(server_js_data.get('instances', []))
 
+        def extract_from_jsmods_instances(js_data):
+            if js_data:
+                return extract_video_data(try_get(
+                    js_data, lambda x: x['jsmods']['instances'], list) or [])
+
         if not video_data:
             server_js_data = self._parse_json(
                 self._search_regex(
         if not video_data:
             server_js_data = self._parse_json(
                 self._search_regex(
-                    r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+(?:stream_pagelet|pagelet_group_mall|permalink_video_pagelet)',
+                    r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_\d+)',
                     webpage, 'js data', default='{}'),
                 video_id, transform_source=js_to_json, fatal=False)
                     webpage, 'js data', default='{}'),
                 video_id, transform_source=js_to_json, fatal=False)
-            if server_js_data:
-                video_data = extract_video_data(try_get(
-                    server_js_data, lambda x: x['jsmods']['instances'],
-                    list) or [])
+            video_data = extract_from_jsmods_instances(server_js_data)
 
         if not video_data:
             if not fatal_if_no_video:
 
         if not video_data:
             if not fatal_if_no_video:
@@ -321,9 +349,37 @@ class FacebookIE(InfoExtractor):
                     expected=True)
             elif '>You must log in to continue' in webpage:
                 self.raise_login_required()
                     expected=True)
             elif '>You must log in to continue' in webpage:
                 self.raise_login_required()
-            else:
-                raise ExtractorError('Cannot parse data')
 
 
+            # Video info not in first request, do a secondary request using
+            # tahoe player specific URL
+            tahoe_data = self._download_webpage(
+                self._VIDEO_PAGE_TAHOE_TEMPLATE % video_id, video_id,
+                data=urlencode_postdata({
+                    '__a': 1,
+                    '__pc': self._search_regex(
+                        r'pkg_cohort["\']\s*:\s*["\'](.+?)["\']', webpage,
+                        'pkg cohort', default='PHASED:DEFAULT'),
+                    '__rev': self._search_regex(
+                        r'client_revision["\']\s*:\s*(\d+),', webpage,
+                        'client revision', default='3944515'),
+                    'fb_dtsg': self._search_regex(
+                        r'"DTSGInitialData"\s*,\s*\[\]\s*,\s*{\s*"token"\s*:\s*"([^"]+)"',
+                        webpage, 'dtsg token', default=''),
+                }),
+                headers={
+                    'Content-Type': 'application/x-www-form-urlencoded',
+                })
+            tahoe_js_data = self._parse_json(
+                self._search_regex(
+                    r'for\s+\(\s*;\s*;\s*\)\s*;(.+)', tahoe_data,
+                    'tahoe js data', default='{}'),
+                video_id, fatal=False)
+            video_data = extract_from_jsmods_instances(tahoe_js_data)
+
+        if not video_data:
+            raise ExtractorError('Cannot parse data')
+
+        subtitles = {}
         formats = []
         for f in video_data:
             format_id = f['stream_type']
         formats = []
         for f in video_data:
             format_id = f['stream_type']
@@ -347,31 +403,45 @@ class FacebookIE(InfoExtractor):
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
+            subtitles_src = f[0].get('subtitles_src')
+            if subtitles_src:
+                subtitles.setdefault('en', []).append({'url': subtitles_src})
         if not formats:
             raise ExtractorError('Cannot find video formats')
 
         if not formats:
             raise ExtractorError('Cannot find video formats')
 
+        # Downloads with browser's User-Agent are rate limited. Working around
+        # with non-browser User-Agent.
+        for f in formats:
+            f.setdefault('http_headers', {})['User-Agent'] = 'facebookexternalhit/1.1'
+
         self._sort_formats(formats)
 
         video_title = self._html_search_regex(
         self._sort_formats(formats)
 
         video_title = self._html_search_regex(
-            r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
-            default=None)
+            r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage,
+            'title', default=None)
         if not video_title:
             video_title = self._html_search_regex(
                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
                 webpage, 'alternative title', default=None)
         if not video_title:
             video_title = self._html_search_meta(
         if not video_title:
             video_title = self._html_search_regex(
                 r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
                 webpage, 'alternative title', default=None)
         if not video_title:
             video_title = self._html_search_meta(
-                'description', webpage, 'title')
+                'description', webpage, 'title', default=None)
         if video_title:
             video_title = limit_length(video_title, 80)
         else:
             video_title = 'Facebook video #%s' % video_id
         uploader = clean_html(get_element_by_id(
             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
         if video_title:
             video_title = limit_length(video_title, 80)
         else:
             video_title = 'Facebook video #%s' % video_id
         uploader = clean_html(get_element_by_id(
             'fbPhotoPageAuthorName', webpage)) or self._search_regex(
-            r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader', fatal=False)
+            r'ownerName\s*:\s*"([^"]+)"', webpage, 'uploader',
+            default=None) or self._og_search_title(webpage, fatal=False)
         timestamp = int_or_none(self._search_regex(
             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
             'timestamp', default=None))
         timestamp = int_or_none(self._search_regex(
             r'<abbr[^>]+data-utime=["\'](\d+)', webpage,
             'timestamp', default=None))
+        thumbnail = self._html_search_meta(['og:image', 'twitter:image'], webpage)
+
+        view_count = parse_count(self._search_regex(
+            r'\bviewCount\s*:\s*["\']([\d,.]+)', webpage, 'view count',
+            default=None))
 
         info_dict = {
             'id': video_id,
 
         info_dict = {
             'id': video_id,
@@ -379,6 +449,9 @@ class FacebookIE(InfoExtractor):
             'formats': formats,
             'uploader': uploader,
             'timestamp': timestamp,
             'formats': formats,
             'uploader': uploader,
             'timestamp': timestamp,
+            'thumbnail': thumbnail,
+            'view_count': view_count,
+            'subtitles': subtitles,
         }
 
         return webpage, info_dict
         }
 
         return webpage, info_dict
@@ -393,15 +466,18 @@ class FacebookIE(InfoExtractor):
             return info_dict
 
         if '/posts/' in url:
             return info_dict
 
         if '/posts/' in url:
-            entries = [
-                self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
-                for vid in self._parse_json(
-                    self._search_regex(
-                        r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
-                        webpage, 'video ids', group='ids'),
-                    video_id)]
-
-            return self.playlist_result(entries, video_id)
+            video_id_json = self._search_regex(
+                r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])', webpage, 'video ids', group='ids',
+                default='')
+            if video_id_json:
+                entries = [
+                    self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
+                    for vid in self._parse_json(video_id_json, video_id)]
+                return self.playlist_result(entries, video_id)
+
+            # Single Video?
+            video_id = self._search_regex(r'video_id:\s*"([0-9]+)"', webpage, 'single video id')
+            return self.url_result('facebook:%s' % video_id, FacebookIE.ie_key())
         else:
             _, info_dict = self._extract_from_url(
                 self._VIDEO_PAGE_TEMPLATE % video_id,
         else:
             _, info_dict = self._extract_from_url(
                 self._VIDEO_PAGE_TEMPLATE % video_id,