[MDR] Relax _VALID_URL and playerURL matching and update _TESTS

Ref: #12169
This commit is contained in:
Yen Chi Hsuan 2017-02-26 17:24:54 +08:00
parent b3aec47665
commit a5cf17989b
No known key found for this signature in database
GPG key ID: 7F902A182457CA23
2 changed files with 17 additions and 4 deletions

View file

@ -1,6 +1,7 @@
version <unreleased> version <unreleased>
Extractors Extractors
+ [MDR] Recognize more URL patterns (#12169)
* [vevo] Fix extraction for videos with the new streams/streamsV3 format * [vevo] Fix extraction for videos with the new streams/streamsV3 format
(#11719) (#11719)
+ [njpwworld] Add new extractor (#11561) + [njpwworld] Add new extractor (#11561)

View file

@ -14,7 +14,7 @@ from ..utils import (
class MDRIE(InfoExtractor): class MDRIE(InfoExtractor):
IE_DESC = 'MDR.DE and KiKA' IE_DESC = 'MDR.DE and KiKA'
_VALID_URL = r'https?://(?:www\.)?(?:mdr|kika)\.de/(?:.*)/[a-z]+-?(?P<id>\d+)(?:_.+?)?\.html' _VALID_URL = r'https?://(?:www\.)?(?:mdr|kika)\.de/(?:.*)/[a-z-]+-?(?P<id>\d+)(?:_.+?)?\.html'
_TESTS = [{ _TESTS = [{
# MDR regularly deletes its videos # MDR regularly deletes its videos
@ -31,6 +31,7 @@ class MDRIE(InfoExtractor):
'duration': 250, 'duration': 250,
'uploader': 'MITTELDEUTSCHER RUNDFUNK', 'uploader': 'MITTELDEUTSCHER RUNDFUNK',
}, },
'skip': '404 not found',
}, { }, {
'url': 'http://www.kika.de/baumhaus/videos/video19636.html', 'url': 'http://www.kika.de/baumhaus/videos/video19636.html',
'md5': '4930515e36b06c111213e80d1e4aad0e', 'md5': '4930515e36b06c111213e80d1e4aad0e',
@ -41,6 +42,7 @@ class MDRIE(InfoExtractor):
'duration': 134, 'duration': 134,
'uploader': 'KIKA', 'uploader': 'KIKA',
}, },
'skip': '404 not found',
}, { }, {
'url': 'http://www.kika.de/sendungen/einzelsendungen/weihnachtsprogramm/videos/video8182.html', 'url': 'http://www.kika.de/sendungen/einzelsendungen/weihnachtsprogramm/videos/video8182.html',
'md5': '5fe9c4dd7d71e3b238f04b8fdd588357', 'md5': '5fe9c4dd7d71e3b238f04b8fdd588357',
@ -49,11 +51,21 @@ class MDRIE(InfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'Beutolomäus und der geheime Weihnachtswunsch', 'title': 'Beutolomäus und der geheime Weihnachtswunsch',
'description': 'md5:b69d32d7b2c55cbe86945ab309d39bbd', 'description': 'md5:b69d32d7b2c55cbe86945ab309d39bbd',
'timestamp': 1450950000, 'timestamp': 1482541200,
'upload_date': '20151224', 'upload_date': '20161224',
'duration': 4628, 'duration': 4628,
'uploader': 'KIKA', 'uploader': 'KIKA',
}, },
}, {
# audio with alternative playerURL pattern
'url': 'http://www.mdr.de/kultur/videos-und-audios/audio-radio/operation-mindfuck-robert-wilson100.html',
'info_dict': {
'id': '100',
'ext': 'mp4',
'title': 'Feature: Operation Mindfuck - Robert Anton Wilson',
'duration': 3239,
'uploader': 'MITTELDEUTSCHER RUNDFUNK',
},
}, { }, {
'url': 'http://www.kika.de/baumhaus/sendungen/video19636_zc-fea7f8a0_zs-4bf89c60.html', 'url': 'http://www.kika.de/baumhaus/sendungen/video19636_zc-fea7f8a0_zs-4bf89c60.html',
'only_matching': True, 'only_matching': True,
@ -71,7 +83,7 @@ class MDRIE(InfoExtractor):
webpage = self._download_webpage(url, video_id) webpage = self._download_webpage(url, video_id)
data_url = self._search_regex( data_url = self._search_regex(
r'(?:dataURL|playerXml(?:["\'])?)\s*:\s*(["\'])(?P<url>.+/(?:video|audio)-?[0-9]+-avCustom\.xml)\1', r'(?:dataURL|playerXml(?:["\'])?)\s*:\s*(["\'])(?P<url>.+?-avCustom\.xml)\1',
webpage, 'data url', group='url').replace(r'\/', '/') webpage, 'data url', group='url').replace(r'\/', '/')
doc = self._download_xml( doc = self._download_xml(