pepepper/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-11-14 22:37:59 +09:00
Code Issues Projects Releases Wiki Activity

[WDR] the other wdrmaus.de pages also changed to the new player

Browse Source
This commit is contained in:
Boris Wachtmeister 2016-05-26 19:58:55 +02:00
parent 33a1ff7113
commit 949fc42e00
2 changed files with 15 additions and 75 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
youtube_dl/extractor/extractors.py
View File

@ -923,7 +923,6 @@ from .wat import WatIE
from .wdr import (
WDRIE,
WDRMobileIE,
WDRMausIE,
)
from .webofstories import (
WebOfStoriesIE,

89
youtube_dl/extractor/wdr.py
View File

@ -4,10 +4,6 @@ from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..compat import (
compat_parse_qs,
compat_urlparse,
)
from ..utils import (
determine_ext,
js_to_json,
@ -18,7 +14,7 @@ from ..utils import (
class WDRIE(InfoExtractor):
_CURRENT_MAUS_URL = r'https?://www.wdrmaus.de/aktuelle-sendung/(wdr|index).php5'
_CURRENT_MAUS_URL = r'https?://(?:www\.)wdrmaus.de/(?:[^/]+/){1,2}[^/?#]+\.php5'
_PAGE_REGEX = r'/mediathek/(?P<media_type>[^/]+)/(?P<type>[^/]+)/(?P<display_id>.+)\.html'
_VALID_URL = r'(?P<page_url>https?://(?:www\d\.)?wdr\d?\.de)' + _PAGE_REGEX + '|' + _CURRENT_MAUS_URL
@ -92,6 +88,20 @@ class WDRIE(InfoExtractor):
},
'skip': 'The id changes from week to week because of the new episode'
},
{
'url': 'http://www.wdrmaus.de/sachgeschichten/sachgeschichten/achterbahn.php5',
'md5': 'ca365705551e4bd5217490f3b0591290',
'info_dict': {
'id': 'mdb-186083',
'ext': 'flv',
'upload_date': '20130919',
'title': 'Sachgeschichte - Achterbahn ',
'description': '- Die Sendung mit der Maus -',
},
'params': {
'skip_download': True, # the file has different versions :(
},
},
]
def _real_extract(self, url):
@ -222,72 +232,3 @@ class WDRMobileIE(InfoExtractor):
'User-Agent': 'mobile',
},
}
class WDRMausIE(InfoExtractor):
_VALID_URL = 'https?://(?:www\.)?wdrmaus\.de/(?:[^/]+/){,2}(?P<id>[^/?#]+)((?<!index)\.php5|/(?:$|[?#]))'
IE_DESC = 'Sendung mit der Maus'
_TESTS = [{
'url': 'http://www.wdrmaus.de/sachgeschichten/sachgeschichten/achterbahn.php5',
'md5': '178b432d002162a14ccb3e0876741095',
'info_dict': {
'id': 'achterbahn',
'ext': 'mp4',
'thumbnail': 're:^http://.+\.jpg',
'upload_date': '20131001',
'title': '19.09.2013 - Achterbahn',
}
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
param_code = self._html_search_regex(
r'<a href="\?startVideo=1&amp;([^"]+)"', webpage, 'parameters')
title_date = self._search_regex(
r'<div class="sendedatum"><p>Sendedatum:\s*([0-9\.]+)</p>',
webpage, 'air date')
title_str = self._html_search_regex(
r'<h1>(.*?)</h1>', webpage, 'title')
title = '%s - %s' % (title_date, title_str)
upload_date = unified_strdate(
self._html_search_meta('dc.date', webpage))
fields = compat_parse_qs(param_code)
video_url = fields['firstVideo'][0]
thumbnail = compat_urlparse.urljoin(url, fields['startPicture'][0])
formats = [{
'format_id': 'rtmp',
'url': video_url,
}]
jscode = self._download_webpage(
'http://www.wdrmaus.de/codebase/js/extended-medien.min.js',
video_id, fatal=False,
note='Downloading URL translation table',
errnote='Could not download URL translation table')
if jscode:
for m in re.finditer(
r"stream:\s*'dslSrc=(?P<stream>[^']+)',\s*download:\s*'(?P<dl>[^']+)'\s*\}",
jscode):
if video_url.startswith(m.group('stream')):
http_url = video_url.replace(
m.group('stream'), m.group('dl'))
formats.append({
'format_id': 'http',
'url': http_url,
})
break
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'formats': formats,
'thumbnail': thumbnail,
'upload_date': upload_date,
}