Upstream/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-06-23 21:25:04 +02:00
Code Issues Projects Releases Wiki Activity

[SpankBang] Back-port changes to SpankBangIE from yt-dlp

Browse Source 
* improve title extraction
* add uploader_id
* update test
* but don't check file md5
This commit is contained in:
dirkf 2022-06-06 14:18:46 +01:00
parent 04fd3289d3
commit 30a954bad9
1 changed files with 19 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
youtube_dl/extractor/spankbang.py
View File

@ -26,19 +26,24 @@ class SpankBangIE(InfoExtractor):
) )
''' '''
_TESTS = [{ _TESTS = [{
'url': 'http://spankbang.com/3vvn/video/fantasy+solo', 'url': 'https://spankbang.com/56b3d/video/the+slut+maker+hmv',
'md5': '1cc433e1d6aa14bc376535b8679302f7', 'md5': '5039ba9d26f6124a7fdea6df2f21e765',
'info_dict': { 'info_dict': {
'id': '3vvn', 'id': '56b3d',
'ext': 'mp4', 'ext': 'mp4',
'title': 'fantasy solo', 'title': 'The Slut Maker HMV',
'description': 'dillion harper masturbates on a bed', 'description': 'Girls getting converted into cock slaves.',
'thumbnail': r're:^https?://.*\.jpg$', 'thumbnail': r're:^https?://.*\.jpg$',
'uploader': 'silly2587', 'uploader': 'Mindself',
'timestamp': 1422571989, 'uploader_id': 'mindself',
'upload_date': '20150129', 'timestamp': 1617109572,
'upload_date': '20210330',
'age_limit': 18, 'age_limit': 18,
} },
'params': {
# adaptive download
'skip_download': True,
},
}, { }, {
# 480p only # 480p only
'url': 'http://spankbang.com/1vt0/video/solvane+gangbang', 'url': 'http://spankbang.com/1vt0/video/solvane+gangbang',
@ -134,15 +139,15 @@ class SpankBangIE(InfoExtractor):
info = self._search_json_ld(webpage, video_id, default={}) info = self._search_json_ld(webpage, video_id, default={})
title = self._html_search_regex( title = self._html_search_regex(
r'(?s)<h1[^>]*>(.+?)</h1>', webpage, 'title', default=None) r'(?s)<h1[^>]+\btitle=["\']([^"]+)["\']>', webpage, 'title', default=None)
description = self._search_regex( description = self._search_regex(
r'<div[^>]+\bclass=["\']bottom[^>]+>\s*<p>[^<]*</p>\s*<p>([^<]+)', r'<div[^>]+\bclass=["\']bottom[^>]+>\s*<p>[^<]*</p>\s*<p>([^<]+)',
webpage, 'description', default=None) webpage, 'description', default=None)
thumbnail = self._og_search_thumbnail(webpage, default=None) thumbnail = self._og_search_thumbnail(webpage, default=None)
uploader = self._html_search_regex( uploader = self._html_search_regex(
(r'(?s)<li[^>]+class=["\']profile[^>]+>(.+?)</a>', r'<svg[^>]+\bclass="(?:[^"]*?user[^"]*?)">.*?</svg>([^<]+)', webpage, 'uploader', default=None)
r'class="user"[^>]*><img[^>]+>([^<]+)'), uploader_id = self._html_search_regex(
webpage, 'uploader', default=None) r'<a[^>]+href="/profile/([^"]+)"', webpage, 'uploader_id', default=None)
duration = parse_duration(self._search_regex( duration = parse_duration(self._search_regex(
r'<div[^>]+\bclass=["\']right_side[^>]+>\s*<span>([^<]+)', r'<div[^>]+\bclass=["\']right_side[^>]+>\s*<span>([^<]+)',
webpage, 'duration', default=None)) webpage, 'duration', default=None))
@ -157,6 +162,7 @@ class SpankBangIE(InfoExtractor):
'description': description, 'description': description,
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'uploader': uploader, 'uploader': uploader,
'uploader_id': uploader_id,
'duration': duration, 'duration': duration,
'view_count': view_count, 'view_count': view_count,
'formats': formats, 'formats': formats,