midgard/youtube-dl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[gaskrank] Improve (closes #12493)

Browse source
This commit is contained in:
Sergey M․ 2017-05-28 06:46:52 +07:00
parent 5a6d1da442
commit 7e688d2f6a
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D
1 changed files with 36 additions and 38 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

74
youtube_dl/extractor/gaskrank.py
View file

@ -11,50 +11,48 @@ from ..utils import (
class GaskrankIE(InfoExtractor): class GaskrankIE(InfoExtractor):
"""InfoExtractor for gaskrank.tv""" _VALID_URL = r'https?://(?:www\.)?gaskrank\.tv/tv/(?P<categories>[^/]+)/(?P<id>[^/]+)\.htm'
_VALID_URL = r'https?://(?:www\.)?gaskrank\.tv/tv/(?P<categories>[^/]+)/(?P<id>[^/]+)\.html?' _TESTS = [{
_TESTS = [ 'url': 'http://www.gaskrank.tv/tv/motorrad-fun/strike-einparken-durch-anfaenger-crash-mit-groesserem-flurschaden.htm',
{ 'md5': '1ae88dbac97887d85ebd1157a95fc4f9',
'url': 'http://www.gaskrank.tv/tv/motorrad-fun/strike-einparken-durch-anfaenger-crash-mit-groesserem-flurschaden.htm', 'info_dict': {
'md5': '1ae88dbac97887d85ebd1157a95fc4f9', 'id': '201601/26955',
'info_dict': { 'ext': 'mp4',
'id': '201601/26955', 'title': 'Strike! Einparken können nur Männer - Flurschaden hält sich in Grenzen *lol*',
'ext': 'mp4', 'thumbnail': r're:^https?://.*\.jpg$',
'title': 'Strike! Einparken können nur Männer - Flurschaden hält sich in Grenzen *lol*', 'categories': ['motorrad-fun'],
'thumbnail': r're:^https?://.*\.jpg$', 'display_id': 'strike-einparken-durch-anfaenger-crash-mit-groesserem-flurschaden',
'categories': ['motorrad-fun'], 'uploader_id': 'Bikefun',
'display_id': 'strike-einparken-durch-anfaenger-crash-mit-groesserem-flurschaden', 'upload_date': '20170110',
'uploader_id': 'Bikefun', 'uploader_url': None,
'upload_date': '20170110',
'uploader_url': None,
}
},
{
'url': 'http://www.gaskrank.tv/tv/racing/isle-of-man-tt-2011-michael-du-15920.htm',
'md5': 'c33ee32c711bc6c8224bfcbe62b23095',
'info_dict': {
'id': '201106/15920',
'ext': 'mp4',
'title': 'Isle of Man - Michael Dunlop vs Guy Martin - schwindelig kucken',
'thumbnail': r're:^https?://.*\.jpg$',
'categories': ['racing'],
'display_id': 'isle-of-man-tt-2011-michael-du-15920',
'uploader_id': 'IOM',
'upload_date': '20160506',
'uploader_url': 'www.iomtt.com',
}
} }
] }, {
'url': 'http://www.gaskrank.tv/tv/racing/isle-of-man-tt-2011-michael-du-15920.htm',
'md5': 'c33ee32c711bc6c8224bfcbe62b23095',
'info_dict': {
'id': '201106/15920',
'ext': 'mp4',
'title': 'Isle of Man - Michael Dunlop vs Guy Martin - schwindelig kucken',
'thumbnail': r're:^https?://.*\.jpg$',
'categories': ['racing'],
'display_id': 'isle-of-man-tt-2011-michael-du-15920',
'uploader_id': 'IOM',
'upload_date': '20170523',
'uploader_url': 'www.iomtt.com',
}
}]
def _real_extract(self, url): def _real_extract(self, url):
"""extract information from gaskrank.tv"""
display_id = self._match_id(url) display_id = self._match_id(url)
webpage = self._download_webpage(url, display_id) webpage = self._download_webpage(url, display_id)
categories = [re.match(self._VALID_URL, url).group('categories')]
title = self._og_search_title(webpage, default=None) or self._html_search_meta( title = self._og_search_title(
webpage, default=None) or self._html_search_meta(
'title', webpage, fatal=True) 'title', webpage, fatal=True)
categories = [re.match(self._VALID_URL, url).group('categories')]
mobj = re.search( mobj = re.search(
r'Video von:\s*(?P<uploader_id>[^|]*?)\s*\|\s*vom:\s*(?P<upload_date>[0-9][0-9]\.[0-9][0-9]\.[0-9][0-9][0-9][0-9])', r'Video von:\s*(?P<uploader_id>[^|]*?)\s*\|\s*vom:\s*(?P<upload_date>[0-9][0-9]\.[0-9][0-9]\.[0-9][0-9][0-9][0-9])',
webpage) webpage)
@ -83,7 +81,7 @@ class GaskrankIE(InfoExtractor):
video_id = self._search_regex( video_id = self._search_regex(
r'https?://movies\.gaskrank\.tv/([^-]*?)(-[^\.]*)?\.mp4', r'https?://movies\.gaskrank\.tv/([^-]*?)(-[^\.]*)?\.mp4',
webpage, 'video id') webpage, 'video id', default=display_id)
entry = self._parse_html5_media_entries(url, webpage, video_id)[0] entry = self._parse_html5_media_entries(url, webpage, video_id)[0]
entry.update({ entry.update({