X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/51057713322a0d2ff542520876a4c675f26c57ac..c8c68895a8818fe160e81ce9c6f8baea38fe3051:/youtube_dl/extractor/yourporn.py?ds=sidebyside diff --git a/youtube_dl/extractor/yourporn.py b/youtube_dl/extractor/yourporn.py index 8a2d5f6..9834749 100644 --- a/youtube_dl/extractor/yourporn.py +++ b/youtube_dl/extractor/yourporn.py @@ -1,6 +1,7 @@ from __future__ import unicode_literals from .common import InfoExtractor +from ..compat import compat_str from ..utils import ( parse_duration, urljoin, @@ -8,9 +9,9 @@ from ..utils import ( class YourPornIE(InfoExtractor): - _VALID_URL = r'https?://(?:www\.)?(?:yourporn\.sexy|sxyprn\.com)/post/(?P[^/?#&.]+)' + _VALID_URL = r'https?://(?:www\.)?sxyprn\.com/post/(?P[^/?#&.]+)' _TESTS = [{ - 'url': 'https://yourporn.sexy/post/57ffcb2e1179b.html', + 'url': 'https://sxyprn.com/post/57ffcb2e1179b.html', 'md5': '6f8682b6464033d87acaa7a8ff0c092e', 'info_dict': { 'id': '57ffcb2e1179b', @@ -33,11 +34,19 @@ class YourPornIE(InfoExtractor): webpage = self._download_webpage(url, video_id) - video_url = urljoin(url, self._parse_json( + parts = self._parse_json( self._search_regex( r'data-vnfo=(["\'])(?P{.+?})\1', webpage, 'data info', group='data'), - video_id)[video_id]).replace('/cdn/', '/cdn5/') + video_id)[video_id].split('/') + + num = 0 + for c in parts[6] + parts[7]: + if c.isnumeric(): + num += int(c) + parts[5] = compat_str(int(parts[5]) - num) + parts[1] += '8' + video_url = urljoin(url, '/'.join(parts)) title = (self._search_regex( r'<[^>]+\bclass=["\']PostEditTA[^>]+>([^<]+)', webpage, 'title', @@ -54,4 +63,5 @@ class YourPornIE(InfoExtractor): 'thumbnail': thumbnail, 'duration': duration, 'age_limit': 18, + 'ext': 'mp4', }