paradizelost/youtube-dl
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

add support for view count

Browse Source
This commit is contained in:
peugeot 2014-08-30 20:54:38 +02:00
parent 676e3ecf24
commit 12c82cf9cb
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
youtube_dl/extractor/vporn.py
View File

@ -4,6 +4,7 @@ from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..utils import int_or_none
class VpornIE(InfoExtractor): class VpornIE(InfoExtractor):
_VALID_URL = r'http?://(?:www\.)?vporn\.com/[a-z]+/(?P<title_dash>[a-z-]+)/(?P<id>\d+)/?' _VALID_URL = r'http?://(?:www\.)?vporn\.com/[a-z]+/(?P<title_dash>[a-z-]+)/(?P<id>\d+)/?'
@ -30,15 +31,18 @@ class VpornIE(InfoExtractor):
description = self._html_search_regex(r'<div class="description_txt">(.*?)</div>', webpage, 'description') description = self._html_search_regex(r'<div class="description_txt">(.*?)</div>', webpage, 'description')
thumbnail = 'http://www.vporn.com' + self._html_search_regex(r'flashvars.imageUrl = "(.*?)"', webpage, 'description') thumbnail = 'http://www.vporn.com' + self._html_search_regex(r'flashvars.imageUrl = "(.*?)"', webpage, 'description')
mobj = re.search( mobj = re.search(r'<span class="f_right">duration (?P<minutes>\d+) min (?P<seconds>\d+) sec </span>', webpage)
r'<span class="f_right">duration (?P<minutes>\d+) min (?P<seconds>\d+) sec </span>', webpage)
duration = int(mobj.group('minutes')) * 60 + int(mobj.group('seconds')) if mobj else None duration = int(mobj.group('minutes')) * 60 + int(mobj.group('seconds')) if mobj else None
mobj = re.search(r'<span>((?P<thousands>\d+),)?(?P<units>\d+) VIEWS</span>', webpage)
view_count = int(mobj.group('thousands')) * 1000 + int(mobj.group('units')) if mobj else None
return { return {
'id': video_id, 'id': video_id,
'url': video_url, 'url': video_url,
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'title': title, 'title': title,
'description': description, 'description': description,
'duration': duration, 'duration': int_or_none(duration),
'view_count': int_or_none(view_count),
} }