1
0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-12-21 00:46:41 +00:00
youtube-dl/youtube_dl/extractor/goshgay.py
Yen Chi Hsuan 12355ac473
[goshgay] Fix extraction
isFamilyFriendly no longer exists in the webpage and I can't find
another indicator.
2016-04-12 17:23:00 +08:00

52 lines
1.5 KiB
Python

# -*- coding: utf-8 -*-
from __future__ import unicode_literals
from .common import InfoExtractor
from ..compat import (
compat_parse_qs,
)
from ..utils import (
parse_duration,
)
class GoshgayIE(InfoExtractor):
_VALID_URL = r'https?://www\.goshgay\.com/video(?P<id>\d+?)($|/)'
_TEST = {
'url': 'http://www.goshgay.com/video299069/diesel_sfw_xxx_video',
'md5': '4b6db9a0a333142eb9f15913142b0ed1',
'info_dict': {
'id': '299069',
'ext': 'flv',
'title': 'DIESEL SFW XXX Video',
'thumbnail': 're:^http://.*\.jpg$',
'duration': 80,
'age_limit': 18,
}
}
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
title = self._html_search_regex(
r'<h2>(.*?)<', webpage, 'title')
duration = parse_duration(self._html_search_regex(
r'<span class="duration">\s*-?\s*(.*?)</span>',
webpage, 'duration', fatal=False))
flashvars = compat_parse_qs(self._html_search_regex(
r'<embed.+?id="flash-player-embed".+?flashvars="([^"]+)"',
webpage, 'flashvars'))
thumbnail = flashvars.get('url_bigthumb', [None])[0]
video_url = flashvars['flv_url'][0]
return {
'id': video_id,
'url': video_url,
'title': title,
'thumbnail': thumbnail,
'duration': duration,
'age_limit': 18,
}