[SexyKarma] Add new extractor
This commit is contained in:
parent
09a42738fc
commit
4740864508
|
@ -316,6 +316,7 @@ from .sbs import SBSIE
|
||||||
from .scivee import SciVeeIE
|
from .scivee import SciVeeIE
|
||||||
from .screencast import ScreencastIE
|
from .screencast import ScreencastIE
|
||||||
from .servingsys import ServingSysIE
|
from .servingsys import ServingSysIE
|
||||||
|
from .sexykarma import SexyKarmaIE
|
||||||
from .shared import SharedIE
|
from .shared import SharedIE
|
||||||
from .sharesix import ShareSixIE
|
from .sharesix import ShareSixIE
|
||||||
from .sina import SinaIE
|
from .sina import SinaIE
|
||||||
|
|
|
@ -0,0 +1,43 @@
|
||||||
|
# coding: utf-8
|
||||||
|
from __future__ import unicode_literals
|
||||||
|
|
||||||
|
from .common import InfoExtractor
|
||||||
|
|
||||||
|
|
||||||
|
class SexyKarmaIE(InfoExtractor):
|
||||||
|
_VALID_URL = r'https?://(?:www\.)?sexykarma\.com/gonewild/video/(?P<id>[a-zA-Z0-9\-]+)(.html)'
|
||||||
|
_TESTS = [{
|
||||||
|
'url': 'http://www.sexykarma.com/gonewild/video/taking-a-quick-pee-yHI70cOyIHt.html',
|
||||||
|
'md5': 'b9798e7d1ef1765116a8f516c8091dbd',
|
||||||
|
'info_dict': {
|
||||||
|
'id': 'taking-a-quick-pee-yHI70cOyIHt',
|
||||||
|
'ext': 'mp4',
|
||||||
|
'title': 'Taking a quick pee.',
|
||||||
|
'uploader': 'wildginger7',
|
||||||
|
}
|
||||||
|
}, {
|
||||||
|
'url': 'http://www.sexykarma.com/gonewild/video/pot-pixie-tribute-8Id6EZPbuHf.html',
|
||||||
|
'md5': 'dd216c68d29b49b12842b9babe762a5d',
|
||||||
|
'info_dict': {
|
||||||
|
'id': 'pot-pixie-tribute-8Id6EZPbuHf',
|
||||||
|
'ext': 'mp4',
|
||||||
|
'title': 'pot_pixie tribute',
|
||||||
|
'uploader': 'banffite',
|
||||||
|
}
|
||||||
|
}]
|
||||||
|
|
||||||
|
def _real_extract(self, url):
|
||||||
|
video_id = self._match_id(url)
|
||||||
|
|
||||||
|
# TODO more code goes here, for example ...
|
||||||
|
webpage = self._download_webpage(url, video_id)
|
||||||
|
title = self._html_search_regex(r'<h2 class="he2"><span>(.*?)</span>', webpage, 'title')
|
||||||
|
uploader = self._html_search_regex(r'class="aupa">\n*(.*?)</a>', webpage, 'uploader')
|
||||||
|
url = self._html_search_regex(r'<p><a href="(.*?)" ?\n*target="_blank"><font color', webpage, 'url')
|
||||||
|
|
||||||
|
return {
|
||||||
|
'id': video_id,
|
||||||
|
'title': title,
|
||||||
|
'uploader': uploader,
|
||||||
|
'url': url
|
||||||
|
}
|
Loading…
Reference in New Issue