[odnoklassniki] Add extractor (Closes #5075)
This commit is contained in:
parent
781a7ef60a
commit
4ffbf77886
2 changed files with 86 additions and 0 deletions
|
@ -346,6 +346,7 @@ from .ntvde import NTVDeIE
|
|||
from .ntvru import NTVRuIE
|
||||
from .nytimes import NYTimesIE
|
||||
from .nuvid import NuvidIE
|
||||
from .odnoklassniki import OdnoklassnikiIE
|
||||
from .oktoberfesttv import OktoberfestTVIE
|
||||
from .ooyala import OoyalaIE
|
||||
from .openfilm import OpenFilmIE
|
||||
|
|
85
youtube_dl/extractor/odnoklassniki.py
Normal file
85
youtube_dl/extractor/odnoklassniki.py
Normal file
|
@ -0,0 +1,85 @@
|
|||
# coding: utf-8
|
||||
from __future__ import unicode_literals
|
||||
|
||||
from .common import InfoExtractor
|
||||
from ..utils import (
|
||||
unified_strdate,
|
||||
int_or_none,
|
||||
qualities,
|
||||
)
|
||||
|
||||
|
||||
class OdnoklassnikiIE(InfoExtractor):
|
||||
_VALID_URL = r'https?://(?:odnoklassniki|ok)\.ru/(?:video|web-api/video/moviePlayer)/(?P<id>\d+)'
|
||||
_TESTS = [{
|
||||
'url': 'http://ok.ru/video/20079905452',
|
||||
'md5': '8e24ad2da6f387948e7a7d44eb8668fe',
|
||||
'info_dict': {
|
||||
'id': '20079905452',
|
||||
'ext': 'mp4',
|
||||
'title': 'Культура меняет нас (прекрасный ролик!))',
|
||||
'duration': 100,
|
||||
'upload_date': '20141207',
|
||||
'uploader_id': '330537914540',
|
||||
'uploader': 'Виталий Добровольский',
|
||||
'like_count': int,
|
||||
'age_limit': 0,
|
||||
},
|
||||
}, {
|
||||
'url': 'http://ok.ru/web-api/video/moviePlayer/20079905452',
|
||||
'only_matching': True,
|
||||
}]
|
||||
|
||||
def _real_extract(self, url):
|
||||
video_id = self._match_id(url)
|
||||
|
||||
webpage = self._download_webpage(url, video_id)
|
||||
|
||||
player = self._parse_json(
|
||||
self._search_regex(
|
||||
r"OKVideo\.start\(({.+?})\s*,\s*'VideoAutoplay_player'", webpage, 'player'),
|
||||
video_id)
|
||||
|
||||
metadata = self._parse_json(player['flashvars']['metadata'], video_id)
|
||||
|
||||
movie = metadata['movie']
|
||||
title = movie['title']
|
||||
thumbnail = movie.get('poster')
|
||||
duration = int_or_none(movie.get('duration'))
|
||||
|
||||
author = metadata.get('author', {})
|
||||
uploader_id = author.get('id')
|
||||
uploader = author.get('name')
|
||||
|
||||
upload_date = unified_strdate(self._html_search_meta(
|
||||
'ya:ovs:upload_date', webpage, 'upload date'))
|
||||
|
||||
age_limit = None
|
||||
adult = self._html_search_meta(
|
||||
'ya:ovs:adult', webpage, 'age limit')
|
||||
if adult:
|
||||
age_limit = 18 if adult == 'true' else 0
|
||||
|
||||
like_count = int_or_none(metadata.get('likeCount'))
|
||||
|
||||
quality = qualities(('mobile', 'lowest', 'low', 'sd', 'hd'))
|
||||
|
||||
formats = [{
|
||||
'url': f['url'],
|
||||
'ext': 'mp4',
|
||||
'format_id': f['name'],
|
||||
'quality': quality(f['name']),
|
||||
} for f in metadata['videos']]
|
||||
|
||||
return {
|
||||
'id': video_id,
|
||||
'title': title,
|
||||
'thumbnail': thumbnail,
|
||||
'duration': duration,
|
||||
'upload_date': upload_date,
|
||||
'uploader': uploader,
|
||||
'uploader_id': uploader_id,
|
||||
'like_count': like_count,
|
||||
'age_limit': age_limit,
|
||||
'formats': formats,
|
||||
}
|
Loading…
Reference in a new issue