1
0
mirror of https://github.com/l1ving/youtube-dl synced 2024-12-22 20:13:31 +08:00
youtube-dl/youtube_dl/extractor/cultureunplugged.py
2016-08-20 20:02:49 +01:00

71 lines
2.4 KiB
Python

from __future__ import unicode_literals
import re
import time
from .common import InfoExtractor
from ..utils import (
int_or_none,
HEADRequest,
)
class CultureUnpluggedIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?cultureunplugged\.com/documentary/watch-online/play/(?P<id>\d+)(?:/(?P<display_id>[^/]+))?'
_TESTS = [{
'url': 'http://www.cultureunplugged.com/documentary/watch-online/play/53662/The-Next--Best-West',
'md5': 'ac6c093b089f7d05e79934dcb3d228fc',
'info_dict': {
'id': '53662',
'display_id': 'The-Next--Best-West',
'ext': 'mp4',
'title': 'The Next, Best West',
'description': 'md5:0423cd00833dea1519cf014e9d0903b1',
'thumbnail': 're:^https?://.*\.jpg$',
'creator': 'Coldstream Creative',
'duration': 2203,
'view_count': int,
}
}, {
'url': 'http://www.cultureunplugged.com/documentary/watch-online/play/53662',
'only_matching': True,
}]
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
display_id = mobj.group('display_id') or video_id
# request setClientTimezone.php to get PHPSESSID cookie which is need to get valid json data in the next request
self._request_webpage(HEADRequest(
'http://www.cultureunplugged.com/setClientTimezone.php?timeOffset=%d' % -(time.timezone / 3600)), display_id)
movie_data = self._download_json(
'http://www.cultureunplugged.com/movie-data/cu-%s.json' % video_id, display_id)
video_url = movie_data['url']
title = movie_data['title']
description = movie_data.get('synopsis')
creator = movie_data.get('producer')
duration = int_or_none(movie_data.get('duration'))
view_count = int_or_none(movie_data.get('views'))
thumbnails = [{
'url': movie_data['%s_thumb' % size],
'id': size,
'preference': preference,
} for preference, size in enumerate((
'small', 'large')) if movie_data.get('%s_thumb' % size)]
return {
'id': video_id,
'display_id': display_id,
'url': video_url,
'title': title,
'description': description,
'creator': creator,
'duration': duration,
'view_count': view_count,
'thumbnails': thumbnails,
}