mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-12-15 07:42:42 +00:00
61 lines
2.7 KiB
Python
61 lines
2.7 KiB
Python
import re
|
|
|
|
from .common import InfoExtractor
|
|
|
|
from ..utils import (
|
|
ExtractorError,
|
|
unified_strdate,
|
|
)
|
|
|
|
class NormalbootsIE(InfoExtractor):
|
|
_VALID_URL = r'(?:http://)?(?:www\.)?normalboots\.com/video/(?P<videoid>[0-9a-z-]*)/?$'
|
|
_TEST = {
|
|
u'url': u'http://normalboots.com/video/home-alone-games-jontron/',
|
|
u'file': u'home-alone-games-jontron.mp4',
|
|
u'md5': u'8bf6de238915dd501105b44ef5f1e0f6',
|
|
u'info_dict': {
|
|
u'title': u'Home Alone Games - JonTron - NormalBoots',
|
|
u'description': u'Jon is late for Christmas. Typical. Thanks to: Paul Ritchey for Co-Writing/Filming: http://www.youtube.com/user/ContinueShow Michael Azzi for Christmas Intro Animation: http://michafrar.tumblr.com/ Jerrod Waters for Christmas Intro Music: http://www.youtube.com/user/xXJerryTerryXx Casey Ormond for \u2018Tense Battle Theme\u2019:\xa0http://www.youtube.com/Kiamet/',
|
|
u'uploader': u'JonTron',
|
|
u'upload_date': u'20140125',
|
|
}
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
mobj = re.match(self._VALID_URL, url)
|
|
if mobj is None:
|
|
raise ExtractorError(u'Invalid URL: %s' % url)
|
|
video_id = mobj.group('videoid')
|
|
|
|
info = {
|
|
'id': video_id,
|
|
'uploader': None,
|
|
'upload_date': None,
|
|
}
|
|
|
|
if url[:4] != 'http':
|
|
url = 'http://' + url
|
|
|
|
webpage = self._download_webpage(url, video_id)
|
|
video_title = self._og_search_title(webpage)
|
|
video_description = self._og_search_description(webpage)
|
|
video_thumbnail = self._og_search_thumbnail(webpage)
|
|
video_uploader = self._html_search_regex(r'Posted\sby\s<a\shref="[A-Za-z0-9/]*">(?P<uploader>[A-Za-z]*)\s</a>',
|
|
webpage, 'uploader')
|
|
raw_upload_date = self._html_search_regex('<span style="text-transform:uppercase; font-size:inherit;">[A-Za-z]+, (?P<date>.*)</span>',
|
|
webpage, 'date')
|
|
video_upload_date = unified_strdate(raw_upload_date)
|
|
video_upload_date = unified_strdate(raw_upload_date)
|
|
|
|
player_url = self._html_search_regex(r'<iframe\swidth="[0-9]+"\sheight="[0-9]+"\ssrc="(?P<url>[\S]+)"', webpage, 'url')
|
|
player_page = self._download_webpage(player_url, video_id)
|
|
video_url = u'http://player.screenwavemedia.com/' + self._html_search_regex(r"'file':\s'(?P<file>[0-9A-Za-z-_\.]+)'", player_page, 'file')
|
|
|
|
info['url'] = video_url
|
|
info['title'] = video_title
|
|
info['description'] = video_description
|
|
info['thumbnail'] = video_thumbnail
|
|
info['uploader'] = video_uploader
|
|
info['upload_date'] = video_upload_date
|
|
|
|
return info
|