mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-30 20:30:27 +00:00
54 lines
1.9 KiB
Python
54 lines
1.9 KiB
Python
# coding: utf-8
|
|
from __future__ import unicode_literals
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import unified_strdate
|
|
|
|
|
|
class GlideIE(InfoExtractor):
|
|
IE_DESC = 'Glide mobile video messages (glide.me)'
|
|
_VALID_URL = r'https?://share\.glide\.me/(?P<id>[A-Za-z0-9\-=_+]+)'
|
|
_TEST = {
|
|
'url': 'http://share.glide.me/UZF8zlmuQbe4mr+7dCiQ0w==',
|
|
'md5': '4466372687352851af2d131cfaa8a4c7',
|
|
'info_dict': {
|
|
'id': 'UZF8zlmuQbe4mr+7dCiQ0w==',
|
|
'ext': 'mp4',
|
|
'title': 'Damon Timm\'s Glide message',
|
|
'thumbnail': 're:^https?://.*?\.cloudfront\.net/.*\.jpg$',
|
|
'uploader': 'Damon Timm',
|
|
'upload_date': '20140919',
|
|
}
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
|
|
webpage = self._download_webpage(url, video_id)
|
|
|
|
title = self._html_search_regex(
|
|
r'<title>(.+?)</title>', webpage, 'title')
|
|
video_url = self._proto_relative_url(self._search_regex(
|
|
r'<source[^>]+src=(["\'])(?P<url>.+?)\1',
|
|
webpage, 'video URL', default=None,
|
|
group='url')) or self._og_search_video_url(webpage)
|
|
thumbnail = self._proto_relative_url(self._search_regex(
|
|
r'<img[^>]+id=["\']video-thumbnail["\'][^>]+src=(["\'])(?P<url>.+?)\1',
|
|
webpage, 'thumbnail url', default=None,
|
|
group='url')) or self._og_search_thumbnail(webpage)
|
|
uploader = self._search_regex(
|
|
r'<div[^>]+class=["\']info-name["\'][^>]*>([^<]+)',
|
|
webpage, 'uploader', fatal=False)
|
|
upload_date = unified_strdate(self._search_regex(
|
|
r'<div[^>]+class="info-date"[^>]*>([^<]+)',
|
|
webpage, 'upload date', fatal=False))
|
|
|
|
return {
|
|
'id': video_id,
|
|
'title': title,
|
|
'url': video_url,
|
|
'thumbnail': thumbnail,
|
|
'uploader': uploader,
|
|
'upload_date': upload_date,
|
|
}
|