1
0
Fork 0
mirror of https://gitlab.com/dstftw/youtube-dl.git synced 2020-11-16 09:42:26 +00:00
youtube-dl/youtube_dl/extractor/sina.py
2013-07-18 15:31:50 +02:00

68 lines
2.5 KiB
Python

# coding: utf-8
import re
import xml.etree.ElementTree
from .common import InfoExtractor
from ..utils import (
compat_urllib_request,
compat_urllib_parse,
)
class SinaIE(InfoExtractor):
_VALID_URL = r'''https?://(.*?\.)?video\.sina\.com\.cn/
(
(.+?/(((?P<pseudo_id>\d+).html)|(.*?(\#|(vid=))(?P<id>\d+?)($|&))))
|
# This is used by external sites like Weibo
(api/sinawebApi/outplay.php/(?P<token>.+?)\.swf)
)
'''
_TEST = {
u'url': u'http://video.sina.com.cn/news/vlist/zt/chczlj2013/?opsubject_id=top12#110028898',
u'file': u'110028898.flv',
u'md5': u'd65dd22ddcf44e38ce2bf58a10c3e71f',
u'info_dict': {
u'title': u'《中国新闻》 朝鲜要求巴拿马立即释放被扣船员',
}
}
@classmethod
def suitable(cls, url):
return re.match(cls._VALID_URL, url, flags=re.VERBOSE) is not None
def _extract_video(self, video_id):
data = compat_urllib_parse.urlencode({'vid': video_id})
url_page = self._download_webpage('http://v.iask.com/v_play.php?%s' % data,
video_id, u'Downloading video url')
image_page = self._download_webpage(
'http://interface.video.sina.com.cn/interface/common/getVideoImage.php?%s' % data,
video_id, u'Downloading thumbnail info')
url_doc = xml.etree.ElementTree.fromstring(url_page.encode('utf-8'))
return {'id': video_id,
'url': url_doc.find('./durl/url').text,
'ext': 'flv',
'title': url_doc.find('./vname').text,
'thumbnail': image_page.split('=')[1],
}
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url, flags=re.VERBOSE)
video_id = mobj.group('id')
if mobj.group('token') is not None:
# The video id is in the redirected url
self.to_screen(u'Getting video id')
request = compat_urllib_request.Request(url)
request.get_method = lambda: 'HEAD'
(_, urlh) = self._download_webpage_handle(request, 'NA', False)
return self._real_extract(urlh.geturl())
elif video_id is None:
pseudo_id = mobj.group('pseudo_id')
webpage = self._download_webpage(url, pseudo_id)
video_id = self._search_regex(r'vid:\'(\d+?)\'', webpage, u'video id')
return self._extract_video(video_id)