mirror of
https://github.com/ytdl-org/youtube-dl.git
synced 2024-11-21 22:51:24 +01:00
183 lines
6.3 KiB
Python
183 lines
6.3 KiB
Python
from __future__ import unicode_literals
|
|
|
|
import re
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
float_or_none,
|
|
int_or_none,
|
|
merge_dicts,
|
|
str_or_none,
|
|
str_to_int,
|
|
url_or_none,
|
|
)
|
|
|
|
|
|
class SpankwireIE(InfoExtractor):
|
|
_VALID_URL = r'''(?x)
|
|
https?://
|
|
(?:www\.)?spankwire\.com/
|
|
(?:
|
|
[^/]+/video|
|
|
EmbedPlayer\.aspx/?\?.*?\bArticleId=
|
|
)
|
|
(?P<id>\d+)
|
|
'''
|
|
_TESTS = [{
|
|
# download URL pattern: */<height>P_<tbr>K_<video_id>.mp4
|
|
'url': 'http://www.spankwire.com/Buckcherry-s-X-Rated-Music-Video-Crazy-Bitch/video103545/',
|
|
'md5': '5aa0e4feef20aad82cbcae3aed7ab7cd',
|
|
'info_dict': {
|
|
'id': '103545',
|
|
'ext': 'mp4',
|
|
'title': 'Buckcherry`s X Rated Music Video Crazy Bitch',
|
|
'description': 'Crazy Bitch X rated music video.',
|
|
'duration': 222,
|
|
'uploader': 'oreusz',
|
|
'uploader_id': '124697',
|
|
'timestamp': 1178587885,
|
|
'upload_date': '20070508',
|
|
'average_rating': float,
|
|
'view_count': int,
|
|
'comment_count': int,
|
|
'age_limit': 18,
|
|
'categories': list,
|
|
'tags': list,
|
|
},
|
|
}, {
|
|
# download URL pattern: */mp4_<format_id>_<video_id>.mp4
|
|
'url': 'http://www.spankwire.com/Titcums-Compiloation-I/video1921551/',
|
|
'md5': '09b3c20833308b736ae8902db2f8d7e6',
|
|
'info_dict': {
|
|
'id': '1921551',
|
|
'ext': 'mp4',
|
|
'title': 'Titcums Compiloation I',
|
|
'description': 'cum on tits',
|
|
'uploader': 'dannyh78999',
|
|
'uploader_id': '3056053',
|
|
'upload_date': '20150822',
|
|
'age_limit': 18,
|
|
},
|
|
'params': {
|
|
'proxy': '127.0.0.1:8118'
|
|
},
|
|
'skip': 'removed',
|
|
}, {
|
|
'url': 'https://www.spankwire.com/EmbedPlayer.aspx/?ArticleId=156156&autostart=true',
|
|
'only_matching': True,
|
|
}]
|
|
|
|
@staticmethod
|
|
def _extract_urls(webpage):
|
|
return re.findall(
|
|
r'<iframe[^>]+\bsrc=["\']((?:https?:)?//(?:www\.)?spankwire\.com/EmbedPlayer\.aspx/?\?.*?\bArticleId=\d+)',
|
|
webpage)
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
|
|
video = self._download_json(
|
|
'https://www.spankwire.com/api/video/%s.json' % video_id, video_id)
|
|
|
|
title = video['title']
|
|
|
|
formats = []
|
|
videos = video.get('videos')
|
|
if isinstance(videos, dict):
|
|
for format_id, format_url in videos.items():
|
|
video_url = url_or_none(format_url)
|
|
if not format_url:
|
|
continue
|
|
height = int_or_none(self._search_regex(
|
|
r'(\d+)[pP]', format_id, 'height', default=None))
|
|
m = re.search(
|
|
r'/(?P<height>\d+)[pP]_(?P<tbr>\d+)[kK]', video_url)
|
|
if m:
|
|
tbr = int(m.group('tbr'))
|
|
height = height or int(m.group('height'))
|
|
else:
|
|
tbr = None
|
|
formats.append({
|
|
'url': video_url,
|
|
'format_id': '%dp' % height if height else format_id,
|
|
'height': height,
|
|
'tbr': tbr,
|
|
})
|
|
m3u8_url = url_or_none(video.get('HLS'))
|
|
if m3u8_url:
|
|
formats.extend(self._extract_m3u8_formats(
|
|
m3u8_url, video_id, 'mp4', entry_protocol='m3u8_native',
|
|
m3u8_id='hls', fatal=False))
|
|
self._sort_formats(formats, ('height', 'tbr', 'width', 'format_id'))
|
|
|
|
view_count = str_to_int(video.get('viewed'))
|
|
|
|
thumbnails = []
|
|
for preference, t in enumerate(('', '2x'), start=0):
|
|
thumbnail_url = url_or_none(video.get('poster%s' % t))
|
|
if not thumbnail_url:
|
|
continue
|
|
thumbnails.append({
|
|
'url': thumbnail_url,
|
|
'preference': preference,
|
|
})
|
|
|
|
def extract_names(key):
|
|
entries_list = video.get(key)
|
|
if not isinstance(entries_list, list):
|
|
return
|
|
entries = []
|
|
for entry in entries_list:
|
|
name = str_or_none(entry.get('name'))
|
|
if name:
|
|
entries.append(name)
|
|
return entries
|
|
|
|
categories = extract_names('categories')
|
|
tags = extract_names('tags')
|
|
|
|
uploader = None
|
|
info = {}
|
|
|
|
webpage = self._download_webpage(
|
|
'https://www.spankwire.com/_/video%s/' % video_id, video_id,
|
|
fatal=False)
|
|
if webpage:
|
|
info = self._search_json_ld(webpage, video_id, default={})
|
|
thumbnail_url = None
|
|
if 'thumbnail' in info:
|
|
thumbnail_url = url_or_none(info['thumbnail'])
|
|
del info['thumbnail']
|
|
if not thumbnail_url:
|
|
thumbnail_url = self._og_search_thumbnail(webpage)
|
|
if thumbnail_url:
|
|
thumbnails.append({
|
|
'url': thumbnail_url,
|
|
'preference': 10,
|
|
})
|
|
uploader = self._html_search_regex(
|
|
r'(?s)by\s*<a[^>]+\bclass=["\']uploaded__by[^>]*>(.+?)</a>',
|
|
webpage, 'uploader', fatal=False)
|
|
if not view_count:
|
|
view_count = str_to_int(self._search_regex(
|
|
r'data-views=["\']([\d,.]+)', webpage, 'view count',
|
|
fatal=False))
|
|
|
|
return merge_dicts({
|
|
'id': video_id,
|
|
'title': title,
|
|
'description': video.get('description'),
|
|
'duration': int_or_none(video.get('duration')),
|
|
'thumbnails': thumbnails,
|
|
'uploader': uploader,
|
|
'uploader_id': str_or_none(video.get('userId')),
|
|
'timestamp': int_or_none(video.get('time_approved_on')),
|
|
'average_rating': float_or_none(video.get('rating')),
|
|
'view_count': view_count,
|
|
'comment_count': int_or_none(video.get('comments')),
|
|
'age_limit': 18,
|
|
'categories': categories,
|
|
'tags': tags,
|
|
'formats': formats,
|
|
}, info)
|