mirror of
https://codeberg.org/polarisfm/youtube-dl
synced 2024-11-23 08:54:33 +01:00
120 lines
4.4 KiB
Python
120 lines
4.4 KiB
Python
from __future__ import unicode_literals
|
|
|
|
import re
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
ExtractorError,
|
|
int_or_none,
|
|
merge_dicts,
|
|
str_to_int,
|
|
unified_strdate,
|
|
url_or_none,
|
|
)
|
|
|
|
|
|
class RedTubeIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:(?:www\.)?redtube\.com/|embed\.redtube\.com/\?.*?\bid=)(?P<id>[0-9]+)'
|
|
_TESTS = [{
|
|
'url': 'http://www.redtube.com/66418',
|
|
'md5': 'fc08071233725f26b8f014dba9590005',
|
|
'info_dict': {
|
|
'id': '66418',
|
|
'ext': 'mp4',
|
|
'title': 'Sucked on a toilet',
|
|
'upload_date': '20110811',
|
|
'duration': 596,
|
|
'view_count': int,
|
|
'age_limit': 18,
|
|
}
|
|
}, {
|
|
'url': 'http://embed.redtube.com/?bgcolor=000000&id=1443286',
|
|
'only_matching': True,
|
|
}]
|
|
|
|
@staticmethod
|
|
def _extract_urls(webpage):
|
|
return re.findall(
|
|
r'<iframe[^>]+?src=["\'](?P<url>(?:https?:)?//embed\.redtube\.com/\?.*?\bid=\d+)',
|
|
webpage)
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
webpage = self._download_webpage(
|
|
'http://www.redtube.com/%s' % video_id, video_id)
|
|
|
|
if any(s in webpage for s in ['video-deleted-info', '>This video has been removed']):
|
|
raise ExtractorError('Video %s has been removed' % video_id, expected=True)
|
|
|
|
info = self._search_json_ld(webpage, video_id, default={})
|
|
|
|
if not info.get('title'):
|
|
info['title'] = self._html_search_regex(
|
|
(r'<h(\d)[^>]+class="(?:video_title_text|videoTitle)[^"]*">(?P<title>(?:(?!\1).)+)</h\1>',
|
|
r'(?:videoTitle|title)\s*:\s*(["\'])(?P<title>(?:(?!\1).)+)\1',),
|
|
webpage, 'title', group='title',
|
|
default=None) or self._og_search_title(webpage)
|
|
|
|
formats = []
|
|
sources = self._parse_json(
|
|
self._search_regex(
|
|
r'sources\s*:\s*({.+?})', webpage, 'source', default='{}'),
|
|
video_id, fatal=False)
|
|
if sources and isinstance(sources, dict):
|
|
for format_id, format_url in sources.items():
|
|
if format_url:
|
|
formats.append({
|
|
'url': format_url,
|
|
'format_id': format_id,
|
|
'height': int_or_none(format_id),
|
|
})
|
|
medias = self._parse_json(
|
|
self._search_regex(
|
|
r'mediaDefinition\s*:\s*(\[.+?\])', webpage,
|
|
'media definitions', default='{}'),
|
|
video_id, fatal=False)
|
|
if medias and isinstance(medias, list):
|
|
for media in medias:
|
|
format_url = url_or_none(media.get('videoUrl'))
|
|
if not format_url:
|
|
continue
|
|
format_id = media.get('quality')
|
|
formats.append({
|
|
'url': format_url,
|
|
'format_id': format_id,
|
|
'height': int_or_none(format_id),
|
|
})
|
|
if not formats:
|
|
video_url = self._html_search_regex(
|
|
r'<source src="(.+?)" type="video/mp4">', webpage, 'video URL')
|
|
formats.append({'url': video_url})
|
|
self._sort_formats(formats)
|
|
|
|
thumbnail = self._og_search_thumbnail(webpage)
|
|
upload_date = unified_strdate(self._search_regex(
|
|
r'<span[^>]+>(?:ADDED|Published on) ([^<]+)<',
|
|
webpage, 'upload date', default=None))
|
|
duration = int_or_none(self._og_search_property(
|
|
'video:duration', webpage, default=None) or self._search_regex(
|
|
r'videoDuration\s*:\s*(\d+)', webpage, 'duration', default=None))
|
|
view_count = str_to_int(self._search_regex(
|
|
(r'<div[^>]*>Views</div>\s*<div[^>]*>\s*([\d,.]+)',
|
|
r'<span[^>]*>VIEWS</span>\s*</td>\s*<td>\s*([\d,.]+)',
|
|
r'<span[^>]+\bclass=["\']video_view_count[^>]*>\s*([\d,.]+)'),
|
|
webpage, 'view count', default=None))
|
|
|
|
# No self-labeling, but they describe themselves as
|
|
# "Home of Videos Porno"
|
|
age_limit = 18
|
|
|
|
return merge_dicts(info, {
|
|
'id': video_id,
|
|
'ext': 'mp4',
|
|
'thumbnail': thumbnail,
|
|
'upload_date': upload_date,
|
|
'duration': duration,
|
|
'view_count': view_count,
|
|
'age_limit': age_limit,
|
|
'formats': formats,
|
|
})
|