1
0
mirror of https://github.com/l1ving/youtube-dl synced 2024-11-22 10:25:34 +08:00
youtube-dl/youtube_dl/extractor/liveleak.py
Philipp Hagemeister 7eeb5bef24 [liveleak] Simplify
2014-01-28 21:57:38 +01:00

45 lines
1.4 KiB
Python

from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
)
class LiveLeakIE(InfoExtractor):
_VALID_URL = r'^(?:http://)?(?:\w+\.)?liveleak\.com/view\?(?:.*?)i=(?P<video_id>[\w_]+)(?:.*)'
_TEST = {
'url': 'http://www.liveleak.com/view?i=757_1364311680',
'file': '757_1364311680.mp4',
'md5': '0813c2430bea7a46bf13acf3406992f4',
'info_dict': {
'description': 'extremely bad day for this guy..!',
'uploader': 'ljfriel2',
'title': 'Most unlucky car accident'
}
}
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('video_id')
webpage = self._download_webpage(url, video_id)
video_url = self._search_regex(
r'file: "(.*?)",', webpage, 'video URL')
video_title = self._og_search_title(webpage).replace('LiveLeak.com -', '').strip()
video_description = self._og_search_description(webpage)
video_uploader = self._html_search_regex(
r'By:.*?(\w+)</a>', webpage, 'uploader', fatal=False)
return {
'id': video_id,
'url': video_url,
'ext': 'mp4',
'title': video_title,
'description': video_description,
'uploader': video_uploader
}