mirror of
https://github.com/l1ving/youtube-dl
synced 2024-11-23 16:53:00 +08:00
46 lines
1.6 KiB
Python
46 lines
1.6 KiB
Python
# -*- coding: utf-8 -*-
|
||
from __future__ import unicode_literals
|
||
|
||
from .common import InfoExtractor
|
||
|
||
|
||
class RUHDIE(InfoExtractor):
|
||
_VALID_URL = r'http://(?:www\.)?ruhd\.ru/play\.php\?vid=(?P<id>\d+)'
|
||
_TEST = {
|
||
'url': 'http://www.ruhd.ru/play.php?vid=207',
|
||
'md5': 'd1a9ec4edf8598e3fbd92bb16072ba83',
|
||
'info_dict': {
|
||
'id': '207',
|
||
'ext': 'divx',
|
||
'title': 'КОТ бааааам',
|
||
'description': 'классный кот)',
|
||
'thumbnail': 're:^http://.*\.jpg$',
|
||
}
|
||
}
|
||
|
||
def _real_extract(self, url):
|
||
video_id = self._match_id(url)
|
||
webpage = self._download_webpage(url, video_id)
|
||
|
||
video_url = self._html_search_regex(
|
||
r'<param name="src" value="([^"]+)"', webpage, 'video url')
|
||
title = self._html_search_regex(
|
||
r'<title>([^<]+) RUHD.ru - Видео Высокого качества №1 в России!</title>',
|
||
webpage, 'title')
|
||
description = self._html_search_regex(
|
||
r'(?s)<div id="longdesc">(.+?)<span id="showlink">',
|
||
webpage, 'description', fatal=False)
|
||
thumbnail = self._html_search_regex(
|
||
r'<param name="previewImage" value="([^"]+)"',
|
||
webpage, 'thumbnail', fatal=False)
|
||
if thumbnail:
|
||
thumbnail = 'http://www.ruhd.ru' + thumbnail
|
||
|
||
return {
|
||
'id': video_id,
|
||
'url': video_url,
|
||
'title': title,
|
||
'description': description,
|
||
'thumbnail': thumbnail,
|
||
}
|