X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Ffacebook.py;h=789dd79d5474ea8f7c9d18df244c686076e3e646;hb=a61ce71468cb222338ccd8039dc631f3619dc585;hp=8bbca4f56f055a9034bcf0b9d06f0131f127bdde;hpb=9d082e7cb81c63dc5c2e616cd7ca7237a5e0d642;p=youtube-dl diff --git a/youtube_dl/extractor/facebook.py b/youtube_dl/extractor/facebook.py index 8bbca4f56..789dd79d5 100644 --- a/youtube_dl/extractor/facebook.py +++ b/youtube_dl/extractor/facebook.py @@ -20,6 +20,7 @@ from ..utils import ( int_or_none, js_to_json, limit_length, + parse_count, sanitized_Request, try_get, urlencode_postdata, @@ -56,7 +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' - _VIDEO_PAGE_TAHOE_TEMPLATE = 'https://www.facebook.com/video/tahoe/async/%s/?chain=true&isvideo=true' + _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', @@ -75,7 +76,7 @@ class FacebookIE(InfoExtractor): '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, @@ -133,7 +134,7 @@ class FacebookIE(InfoExtractor): }, { # 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', @@ -142,6 +143,7 @@ class FacebookIE(InfoExtractor): 'upload_date': '20161030', 'uploader': 'CNN', 'thumbnail': r're:^https?://.*', + 'view_count': int, }, }, { # bigPipe.onPageletArrive ... onPageletArrive pagelet_group_mall @@ -149,7 +151,7 @@ class FacebookIE(InfoExtractor): 'info_dict': { 'id': '1417995061575415', 'ext': 'mp4', - 'title': 'md5:a7b86ca673f51800cd54687b7f4012fe', + 'title': 'md5:1db063d6a8c13faa8da727817339c857', 'timestamp': 1486648217, 'upload_date': '20170209', 'uploader': 'Yaroslav Korpan', @@ -176,7 +178,7 @@ class FacebookIE(InfoExtractor): 'info_dict': { 'id': '1396382447100162', 'ext': 'mp4', - 'title': 'md5:e2d2700afdf84e121f5d0f999bad13a3', + 'title': 'md5:19a428bbde91364e3de815383b54a235', 'timestamp': 1486035494, 'upload_date': '20170202', 'uploader': 'Elisabeth Ahtn', @@ -324,47 +326,58 @@ class FacebookIE(InfoExtractor): 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( 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 []) + video_data = extract_from_jsmods_instances(server_js_data) if not video_data: - # video info not in first request, do a secondary request using tahoe player specific url + if not fatal_if_no_video: + return webpage, False + m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*">