diff options
Diffstat (limited to 'youtube_dl/extractor/facebook.py')
| -rw-r--r-- | youtube_dl/extractor/facebook.py | 411 |
1 files changed, 339 insertions, 72 deletions
diff --git a/youtube_dl/extractor/facebook.py b/youtube_dl/extractor/facebook.py index b69c1ede0..04650af39 100644 --- a/youtube_dl/extractor/facebook.py +++ b/youtube_dl/extractor/facebook.py @@ -1,6 +1,7 @@ # coding: utf-8 from __future__ import unicode_literals +import json import re import socket @@ -8,6 +9,7 @@ from .common import InfoExtractor from ..compat import ( compat_etree_fromstring, compat_http_client, + compat_str, compat_urllib_error, compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus, @@ -16,13 +18,17 @@ from ..utils import ( clean_html, error_to_compat_str, ExtractorError, + float_or_none, get_element_by_id, int_or_none, js_to_json, limit_length, + parse_count, + qualities, sanitized_Request, try_get, urlencode_postdata, + urljoin, ) @@ -38,11 +44,13 @@ class FacebookIE(InfoExtractor): photo\.php| video\.php| video/embed| - story\.php + story\.php| + watch(?:/live)?/? )\?(?:.*?)(?:v|video_id|story_fbid)=| [^/]+/videos/(?:[^/]+/)?| [^/]+/posts/| - groups/[^/]+/permalink/ + groups/[^/]+/permalink/| + watchparty/ )| facebook: ) @@ -53,9 +61,8 @@ class FacebookIE(InfoExtractor): _NETRC_MACHINE = 'facebook' IE_NAME = 'facebook' - _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', @@ -67,17 +74,19 @@ class FacebookIE(InfoExtractor): 'uploader': 'Tennis on Facebook', 'upload_date': '20140908', 'timestamp': 1410199200, - } + }, + 'skip': 'Requires logging in', }, { - 'note': 'Video without discernible title', + # data.video '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, + 'thumbnail': r're:^https?://.*', }, 'expected_warnings': [ 'title' @@ -94,6 +103,7 @@ class FacebookIE(InfoExtractor): 'upload_date': '20160110', 'timestamp': 1452431627, }, + 'skip': 'Requires logging in', }, { 'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570', 'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6', @@ -121,27 +131,35 @@ class FacebookIE(InfoExtractor): '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 + # data.video.story.attachments[].media 'url': 'https://www.facebook.com/cnn/videos/10155529876156509/', - 'md5': '0d9813160b146b3bc8744e006027fcc6', + 'md5': '9571fae53d4165bbbadb17a94651dcdc', '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', + 'thumbnail': r're:^https?://.*', + 'view_count': int, }, }, { # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall + # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media 'url': 'https://www.facebook.com/yaroslav.korpan/videos/1417995061575415/', 'info_dict': { 'id': '1417995061575415', 'ext': 'mp4', - 'title': 'md5:a7b86ca673f51800cd54687b7f4012fe', + 'title': 'md5:1db063d6a8c13faa8da727817339c857', 'timestamp': 1486648217, 'upload_date': '20170209', 'uploader': 'Yaroslav Korpan', @@ -158,16 +176,18 @@ class FacebookIE(InfoExtractor): 'timestamp': 1477305000, 'upload_date': '20161024', 'uploader': 'La Guía Del Varón', + 'thumbnail': r're:^https?://.*', }, 'params': { 'skip_download': True, }, }, { + # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media 'url': 'https://www.facebook.com/groups/1024490957622648/permalink/1396382447100162/', 'info_dict': { 'id': '1396382447100162', 'ext': 'mp4', - 'title': 'md5:e2d2700afdf84e121f5d0f999bad13a3', + 'title': 'md5:19a428bbde91364e3de815383b54a235', 'timestamp': 1486035494, 'upload_date': '20170202', 'uploader': 'Elisabeth Ahtn', @@ -182,43 +202,117 @@ class FacebookIE(InfoExtractor): 'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf', 'only_matching': True, }, { + # data.mediaset.currMedia.edges 'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater', 'only_matching': True, }, { + # data.video.story.attachments[].media 'url': 'facebook:544765982287235', 'only_matching': True, }, { + # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.media 'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/', 'only_matching': True, }, { + # data.video.creation_story.attachments[].media 'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/', 'only_matching': True, }, { + # data.video '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, + }, { + # data.video + '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, + }, + }, { + # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media + 'url': 'https://www.facebook.com/100033620354545/videos/106560053808006/', + 'info_dict': { + 'id': '106560053808006', + }, + 'playlist_count': 2, + }, { + # data.video.story.attachments[].media + 'url': 'https://www.facebook.com/watch/?v=647537299265662', + 'only_matching': True, + }, { + # data.node.comet_sections.content.story.attachments[].style_type_renderer.attachment.all_subattachments.nodes[].media + 'url': 'https://www.facebook.com/PankajShahLondon/posts/10157667649866271', + 'info_dict': { + 'id': '10157667649866271', + }, + 'playlist_count': 3, + }, { + # data.nodes[].comet_sections.content.story.attachments[].style_type_renderer.attachment.media + 'url': 'https://m.facebook.com/Alliance.Police.Department/posts/4048563708499330', + 'info_dict': { + 'id': '117576630041613', + 'ext': 'mp4', + # TODO: title can be extracted from video page + 'title': 'Facebook video #117576630041613', + 'uploader_id': '189393014416438', + 'upload_date': '20201123', + 'timestamp': 1606162592, + }, + 'skip': 'Requires logging in', + }, { + # node.comet_sections.content.story.attached_story.attachments.style_type_renderer.attachment.media + 'url': 'https://www.facebook.com/groups/ateistiskselskab/permalink/10154930137678856/', + 'info_dict': { + 'id': '211567722618337', + 'ext': 'mp4', + 'title': 'Facebook video #211567722618337', + 'uploader_id': '127875227654254', + 'upload_date': '20161122', + 'timestamp': 1479793574, + }, + }, { + # data.video.creation_story.attachments[].media + 'url': 'https://www.facebook.com/watch/live/?v=1823658634322275', + 'only_matching': True, + }, { + 'url': 'https://www.facebook.com/watchparty/211641140192478', + 'info_dict': { + 'id': '211641140192478', + }, + 'playlist_count': 1, + 'skip': 'Requires logging in', }] + _SUPPORTED_PAGLETS_REGEX = r'(?:pagelet_group_mall|permalink_video_pagelet|hyperfeed_story_id_[0-9a-f]+)' + _api_config = { + 'graphURI': '/api/graphql/' + } @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 - mobj = re.search(r'''(?x)<div[^>]+ + for mobj in re.finditer(r'''(?x)<div[^>]+ 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): - (useremail, password) = self._get_login_info() + useremail, password = self._get_login_info() if useremail is None: return @@ -283,52 +377,235 @@ class FacebookIE(InfoExtractor): def _real_initialize(self): self._login() - def _extract_from_url(self, url, video_id, fatal_if_no_video=True): - req = sanitized_Request(url) - req.add_header('User-Agent', self._CHROME_USER_AGENT) - webpage = self._download_webpage(req, video_id) + def _extract_from_url(self, url, video_id): + webpage = self._download_webpage( + url.replace('://m.facebook.com/', '://www.facebook.com/'), video_id) video_data = None def extract_video_data(instances): + video_data = [] for item in instances: - if item[1][0] == 'VideoConfig': + if try_get(item, lambda x: x[1][0]) == 'VideoConfig': video_item = item[2][0] if video_item.get('video_id'): - return video_item['videoData'] + video_data.append(video_item['videoData']) + return video_data server_js_data = self._parse_json(self._search_regex( - r'handleServerJS\(({.+})(?:\);|,")', webpage, - 'server js data', default='{}'), video_id, fatal=False) + [r'handleServerJS\(({.+})(?:\);|,")', r'\bs\.handle\(({.+?})\);'], + webpage, 'server js data', default='{}'), video_id, fatal=False) 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 []) + + def extract_dash_manifest(video, formats): + dash_manifest = video.get('dash_manifest') + if dash_manifest: + formats.extend(self._parse_mpd_formats( + compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest)))) + + def process_formats(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) + + def extract_relay_data(_filter): + return self._parse_json(self._search_regex( + r'handleWithCustomApplyEach\([^,]+,\s*({.*?%s.*?})\);' % _filter, + webpage, 'replay data', default='{}'), video_id, fatal=False) or {} + + def extract_relay_prefetched_data(_filter): + replay_data = extract_relay_data(_filter) + for require in (replay_data.get('require') or []): + if require[0] == 'RelayPrefetchedStreamCache': + return try_get(require, lambda x: x[3][1]['__bbox']['result']['data'], dict) or {} + 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)', - 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 []) + server_js_data = self._parse_json(self._search_regex([ + r'bigPipe\.onPageletArrive\(({.+?})\)\s*;\s*}\s*\)\s*,\s*["\']onPageletArrive\s+' + self._SUPPORTED_PAGLETS_REGEX, + r'bigPipe\.onPageletArrive\(({.*?id\s*:\s*"%s".*?})\);' % self._SUPPORTED_PAGLETS_REGEX + ], webpage, 'js data', default='{}'), video_id, js_to_json, False) + video_data = extract_from_jsmods_instances(server_js_data) + + if not video_data: + data = extract_relay_prefetched_data( + r'"(?:dash_manifest|playable_url(?:_quality_hd)?)"\s*:\s*"[^"]+"') + if data: + entries = [] + + def parse_graphql_video(video): + formats = [] + q = qualities(['sd', 'hd']) + for (suffix, format_id) in [('', 'sd'), ('_quality_hd', 'hd')]: + playable_url = video.get('playable_url' + suffix) + if not playable_url: + continue + formats.append({ + 'format_id': format_id, + 'quality': q(format_id), + 'url': playable_url, + }) + extract_dash_manifest(video, formats) + process_formats(formats) + v_id = video.get('videoId') or video.get('id') or video_id + info = { + 'id': v_id, + 'formats': formats, + 'thumbnail': try_get(video, lambda x: x['thumbnailImage']['uri']), + 'uploader_id': try_get(video, lambda x: x['owner']['id']), + 'timestamp': int_or_none(video.get('publish_time')), + 'duration': float_or_none(video.get('playable_duration_in_ms'), 1000), + } + description = try_get(video, lambda x: x['savable_description']['text']) + title = video.get('name') + if title: + info.update({ + 'title': title, + 'description': description, + }) + else: + info['title'] = description or 'Facebook video #%s' % v_id + entries.append(info) + + def parse_attachment(attachment, key='media'): + media = attachment.get(key) or {} + if media.get('__typename') == 'Video': + return parse_graphql_video(media) + + nodes = data.get('nodes') or [] + node = data.get('node') or {} + if not nodes and node: + nodes.append(node) + for node in nodes: + story = try_get(node, lambda x: x['comet_sections']['content']['story'], dict) or {} + attachments = try_get(story, [ + lambda x: x['attached_story']['attachments'], + lambda x: x['attachments'] + ], list) or [] + for attachment in attachments: + attachment = try_get(attachment, lambda x: x['style_type_renderer']['attachment'], dict) + ns = try_get(attachment, lambda x: x['all_subattachments']['nodes'], list) or [] + for n in ns: + parse_attachment(n) + parse_attachment(attachment) + + edges = try_get(data, lambda x: x['mediaset']['currMedia']['edges'], list) or [] + for edge in edges: + parse_attachment(edge, key='node') + + video = data.get('video') or {} + if video: + attachments = try_get(video, [ + lambda x: x['story']['attachments'], + lambda x: x['creation_story']['attachments'] + ], list) or [] + for attachment in attachments: + parse_attachment(attachment) + if not entries: + parse_graphql_video(video) + + return self.playlist_result(entries, video_id) if not video_data: - if not fatal_if_no_video: - return webpage, False m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage) if m_msg is not None: raise ExtractorError( 'The video is not available, Facebook said: "%s"' % m_msg.group(1), expected=True) - elif '>You must log in to continue' in webpage: + elif any(p in webpage for p in ( + '>You must log in to continue', + 'id="login_form"', + 'id="loginbutton"')): self.raise_login_required() - else: - raise ExtractorError('Cannot parse data') + + if not video_data and '/watchparty/' in url: + post_data = { + 'doc_id': 3731964053542869, + 'variables': json.dumps({ + 'livingRoomID': video_id, + }), + } + + prefetched_data = extract_relay_prefetched_data(r'"login_data"\s*:\s*{') + if prefetched_data: + lsd = try_get(prefetched_data, lambda x: x['login_data']['lsd'], dict) + if lsd: + post_data[lsd['name']] = lsd['value'] + + relay_data = extract_relay_data(r'\[\s*"RelayAPIConfigDefaults"\s*,') + for define in (relay_data.get('define') or []): + if define[0] == 'RelayAPIConfigDefaults': + self._api_config = define[2] + + living_room = self._download_json( + urljoin(url, self._api_config['graphURI']), video_id, + data=urlencode_postdata(post_data))['data']['living_room'] + + entries = [] + for edge in (try_get(living_room, lambda x: x['recap']['watched_content']['edges']) or []): + video = try_get(edge, lambda x: x['node']['video']) or {} + v_id = video.get('id') + if not v_id: + continue + v_id = compat_str(v_id) + entries.append(self.url_result( + self._VIDEO_PAGE_TEMPLATE % v_id, + self.ie_key(), v_id, video.get('name'))) + + return self.playlist_result(entries, video_id) + + if not video_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') + + if len(video_data) > 1: + entries = [] + for v in video_data: + video_url = v[0].get('video_url') + if not video_url: + continue + entries.append(self.url_result(urljoin( + url, video_url), self.ie_key(), v[0].get('video_id'))) + return self.playlist_result(entries, video_id) + video_data = video_data[0] formats = [] + subtitles = {} for f in video_data: format_id = f['stream_type'] if f and isinstance(f, dict): @@ -347,14 +624,14 @@ class FacebookIE(InfoExtractor): 'url': src, 'preference': preference, }) - dash_manifest = f[0].get('dash_manifest') - if dash_manifest: - formats.extend(self._parse_mpd_formats( - compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest)))) + extract_dash_manifest(f[0], formats) + 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') - self._sort_formats(formats) + process_formats(formats) video_title = self._html_search_regex( r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, @@ -372,10 +649,16 @@ class FacebookIE(InfoExtractor): 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)) + 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, @@ -383,34 +666,18 @@ class FacebookIE(InfoExtractor): 'formats': formats, 'uploader': uploader, 'timestamp': timestamp, + 'thumbnail': thumbnail, + 'view_count': view_count, + 'subtitles': subtitles, } - return webpage, info_dict + return info_dict def _real_extract(self, url): video_id = self._match_id(url) real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url - webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False) - - if info_dict: - 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) - else: - _, info_dict = self._extract_from_url( - self._VIDEO_PAGE_TEMPLATE % video_id, - video_id, fatal_if_no_video=True) - return info_dict + return self._extract_from_url(real_url, video_id) class FacebookPluginsVideoIE(InfoExtractor): |
