w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[pornhub] Add support for subtitles (closes #16924)

Browse source
This commit is contained in:
Giuseppe Fabiano 2018-07-29 01:24:10 +02:00 committed by Sergey M
parent 1a88fc5a69
commit 4938c8d573
1 changed files with 34 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

34
youtube_dl/extractor/pornhub.py
View file

@ -18,6 +18,7 @@ from ..utils import (
orderedSet, orderedSet,
remove_quotes, remove_quotes,
str_to_int, str_to_int,
url_or_none,
) )
@ -68,6 +69,31 @@ class PornHubIE(InfoExtractor):
'params': { 'params': {
'skip_download': True, 'skip_download': True,
}, },
}, {
# subtitles
'url': 'https://www.pornhub.com/view_video.php?viewkey=ph5af5fef7c2aa7',
'info_dict': {
'id': 'ph5af5fef7c2aa7',
'ext': 'mp4',
'title': 'BFFS - Cute Teen Girls Share Cock On the Floor',
'uploader': 'BFFs',
'duration': 622,
'view_count': int,
'like_count': int,
'dislike_count': int,
'comment_count': int,
'age_limit': 18,
'tags': list,
'categories': list,
'subtitles': {
'en': [{
"ext": 'srt'
}]
},
},
'params': {
'skip_download': True,
},
}, { }, {
'url': 'http://www.pornhub.com/view_video.php?viewkey=ph557bbb6676d2d', 'url': 'http://www.pornhub.com/view_video.php?viewkey=ph557bbb6676d2d',
'only_matching': True, 'only_matching': True,
@ -139,12 +165,19 @@ class PornHubIE(InfoExtractor):
video_urls = [] video_urls = []
video_urls_set = set() video_urls_set = set()
subtitles = {}
flashvars = self._parse_json( flashvars = self._parse_json(
self._search_regex( self._search_regex(
r'var\s+flashvars_\d+\s*=\s*({.+?});', webpage, 'flashvars', default='{}'), r'var\s+flashvars_\d+\s*=\s*({.+?});', webpage, 'flashvars', default='{}'),
video_id) video_id)
if flashvars: if flashvars:
subtitle_url = url_or_none(flashvars.get('closedCaptionsFile'))
if subtitle_url:
subtitles.setdefault('en', []).append({
'url': subtitle_url,
'ext': 'srt',
})
thumbnail = flashvars.get('image_url') thumbnail = flashvars.get('image_url')
duration = int_or_none(flashvars.get('video_duration')) duration = int_or_none(flashvars.get('video_duration'))
media_definitions = flashvars.get('mediaDefinitions') media_definitions = flashvars.get('mediaDefinitions')
@ -256,6 +289,7 @@ class PornHubIE(InfoExtractor):
'age_limit': 18, 'age_limit': 18,
'tags': tags, 'tags': tags,
'categories': categories, 'categories': categories,
'subtitles': subtitles,
} }