You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
l1ving_youtube-dl/youtube_dl/extractor/manyvids.py

49 lines
1.5 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
from .common import InfoExtractor
from ..utils import int_or_none
class ManyVidsIE(InfoExtractor):
_VALID_URL = r'(?i)https?://(?:www\.)?manyvids\.com/video/(?P<id>\d+)'
_TEST = {
'url': 'https://www.manyvids.com/Video/133957/everthing-about-me/',
'md5': '03f11bb21c52dd12a05be21a5c7dcc97',
'info_dict': {
'id': '133957',
'ext': 'mp4',
'title': 'everthing about me (Preview)',
'view_count': int,
'like_count': int,
},
}
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
video_url = self._search_regex(
r'data-(?:video-filepath|meta-video)\s*=s*(["\'])(?P<url>(?:(?!\1).)+)\1',
webpage, 'video URL', group='url')
title = '%s (Preview)' % self._html_search_regex(
r'<h2[^>]+class="m-a-0"[^>]*>([^<]+)', webpage, 'title')
like_count = int_or_none(self._search_regex(
r'data-likes=["\'](\d+)', webpage, 'like count', default=None))
view_count = int_or_none(self._html_search_regex(
r'(?s)<span[^>]+class="views-wrapper"[^>]*>(.+?)</span', webpage,
'view count', default=None))
return {
'id': video_id,
'title': title,
'view_count': view_count,
'like_count': like_count,
'formats': [{
'url': video_url,
}],
}