Merge branch 'hellporno' of https://github.com/peugeot/youtube-dl into peugeot-hellporno

This commit is contained in:
Sergey M․ committed 2014-12-29 21:13:09 +06:00
commit 03d9aad87c
2 files changed
+53

No files matched your search

+1
View File
@@ -169,6 +169,7 @@ from .grooveshark import GroovesharkIE
from .groupon import GrouponIE
from .hark import HarkIE
from .heise import HeiseIE
from .hellporno import HellPornoIE
from .helsinki import HelsinkiIE
from .hentaistigma import HentaiStigmaIE
from .hornbunny import HornBunnyIE
+52
View File
@@ -0,0 +1,52 @@
from __future__ import unicode_literals
from .common import InfoExtractor
class HellPornoIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?hellporno\.com/videos/(?P<id>[^/]+)'
_TEST = {
'url': 'http://hellporno.com/videos/dixie-is-posing-with-naked-ass-very-erotic/',
'md5': '1fee339c610d2049699ef2aa699439f1',
'info_dict': {
'id': '149116',
'display_id': 'dixie-is-posing-with-naked-ass-very-erotic',
'ext': 'mp4',
'title': 'Dixie is posing with naked ass very erotic',
'thumbnail': 're:https?://.*\.jpg$',
'age_limit': 18,
}
}
def _real_extract(self, url):
display_id = self._match_id(url)
webpage = self._download_webpage(url, display_id)
video_id = self._html_search_regex(
r'video_id:\s*\'([^\']+)\'', webpage, 'id')
ext = self._html_search_regex(
r'postfix:\s*\'([^\']+)\'', webpage, 'ext')[1:]
video_url = self._html_search_regex(
r'video_url:\s*\'([^\']+)\'', webpage, 'video_url')
title = self._html_search_regex(
r'<title>([^<]+)\s*-\s*Hell Porno</title>', webpage, 'title')
thumbnail = self._html_search_regex(
r'preview_url:\s*\'([^\']+)\'',
webpage, 'thumbnail', fatal=False)
categories = self._html_search_meta(
'keywords', webpage, 'categories', default='').split(',')
return {
'id': video_id,
'display_id': display_id,
'url': video_url,
'title': title,
'ext': ext,
'thumbnail': thumbnail,
'categories': categories,
'age_limit': 18,
}