yt-dlp/youtube_dl/extractor/vrak.py
2017-03-03 23:54:21 +07:00

68 lines
2.6 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from .brightcove import BrightcoveNewIE
class VrakIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?vrak\.tv/videos\?.*?target=(?P<id>[0-9\.]+).*'
_TEST = {
'url': 'http://www.vrak.tv/videos?target=1.2240923&filtre=emission&id=1.1806721',
'md5': 'c5d5ce237bca3b1e990ce1b48d1f0948',
'info_dict': {
'id': '5231040869001',
'ext': 'mp4',
'title': 'Référendums américains, animés japonais et hooligans russes',
'upload_date': '20161201',
'description': 'This video file has been uploaded automatically using Oprah. It should be updated with real description soon.',
'timestamp': 1480628425,
'uploader_id': '2890187628001',
}
}
def _real_extract(self, url):
url_id = self._match_id(url)
webpage = self._download_webpage(url, url_id)
result = {}
result['title'] = self._html_search_regex(
r'<h3 class="videoTitle">(.+?)</h3>', webpage, 'title')
# Inspired from BrightcoveNewIE._extract_url()
entries = []
for account_id, player_id, _, video_id in re.findall(
# account_id, player_id and embed from:
# <div class="video-player [...]
# data-publisher-id="2890187628001"
# data-player-id="VkSnGw3cx"
# video id is extracted from weird CMS Java/Javascript notation:
# RW java.lang.String value = '5231040869001';
# Need to use backtrack to pin to a ref since video is in grid
# layout with others
r'''(?sx)
<div[^>]+
data-publisher-id=["\'](\d+)["\']
[^>]*
data-player-id=["\']([^"\']+)["\']
[^>]*
refId&quot;:&quot;([^&]+)&quot;
[^>]*
>.*?
</div>.*?
RW\ java\.lang\.String\ value\ =\ \'brightcove\.article\.\d+\.\3\'
[^>]*
RW\ java\.lang\.String\ value\ =\ \'(\d+)\'
''', webpage):
entries.append(
'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
% (account_id, player_id, 'default', video_id))
if entries:
result = self.url_result(entries[0], BrightcoveNewIE.ie_key())
return result