[inc] Add extractor

This commit is contained in:
Déstin Reed 2017-01-08 19:28:16 +01:00 committed by Sergey M․
parent 022a5d663b
commit 7f0bdc7a31
No known key found for this signature in database
GPG Key ID: 2C393E0F18A9236D
2 changed files with 41 additions and 0 deletions

View File

@ -395,6 +395,7 @@ from .imgur import (
ImgurAlbumIE,
)
from .ina import InaIE
from .inc import IncIE
from .indavideo import (
IndavideoIE,
IndavideoEmbedIE,

View File

@ -0,0 +1,40 @@
from __future__ import unicode_literals
from .common import InfoExtractor
class IncIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?inc\.com(?:/[\w-]+)+/(?P<id>[\w-]+)(?:\.html)?'
_TESTS = [{
'url': 'http://www.inc.com/tip-sheet/bill-gates-says-these-5-books-will-make-you-smarter.html',
'md5': '7416739c9c16438c09fa35619d6ba5cb',
'info_dict': {
'id': '1_wqig47aq',
'ext': 'mov',
'title': 'Bill Gates Says These 5 Books Will Make You Smarter',
'description': 'md5:bea7ff6cce100886fc1995acb743237e',
'timestamp': 1474414430,
'upload_date': '20160920',
'uploader_id': 'video@inc.com',
},
}, {
'url': 'http://www.inc.com/video/david-whitford/founders-forum-tripadvisor-steve-kaufer-most-enjoyable-moment-for-entrepreneur.html',
'only_matching': True,
}]
def _real_extract(self, url):
display_id = self._match_id(url)
webpage = self._download_webpage(url, display_id)
partner_id = self._search_regex(
r'var\s+_bizo_data_partner_id\s*=\s*"(\d+)";',
webpage,
'partner id')
kaltura_id = self._parse_json(self._search_regex(
r'pageInfo\.videos\s*=\s*\[(.+)\];',
webpage,
'kaltura id'),
display_id)['vid_kaltura_id']
return self.url_result('kaltura:%s:%s' % (partner_id, kaltura_id), 'Kaltura')