mirror of
https://github.com/ytdl-org/youtube-dl.git
synced 2024-01-07 17:16:08 +00:00
Merge branch 'remitamine-shahid'
This commit is contained in:
commit
06c085ab6e
|
@ -523,6 +523,7 @@ from .senateisvp import SenateISVPIE
|
||||||
from .servingsys import ServingSysIE
|
from .servingsys import ServingSysIE
|
||||||
from .sexu import SexuIE
|
from .sexu import SexuIE
|
||||||
from .sexykarma import SexyKarmaIE
|
from .sexykarma import SexyKarmaIE
|
||||||
|
from .shahid import ShahidIE
|
||||||
from .shared import SharedIE
|
from .shared import SharedIE
|
||||||
from .sharesix import ShareSixIE
|
from .sharesix import ShareSixIE
|
||||||
from .sina import SinaIE
|
from .sina import SinaIE
|
||||||
|
|
107
youtube_dl/extractor/shahid.py
Normal file
107
youtube_dl/extractor/shahid.py
Normal file
|
@ -0,0 +1,107 @@
|
||||||
|
# coding: utf-8
|
||||||
|
from __future__ import unicode_literals
|
||||||
|
|
||||||
|
from .common import InfoExtractor
|
||||||
|
from ..compat import compat_urllib_parse
|
||||||
|
from ..utils import (
|
||||||
|
ExtractorError,
|
||||||
|
int_or_none,
|
||||||
|
parse_iso8601,
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
|
class ShahidIE(InfoExtractor):
|
||||||
|
_VALID_URL = r'https?://shahid\.mbc\.net/ar/episode/(?P<id>\d+)/?'
|
||||||
|
_TESTS = [{
|
||||||
|
'url': 'https://shahid.mbc.net/ar/episode/90574/%D8%A7%D9%84%D9%85%D9%84%D9%83-%D8%B9%D8%A8%D8%AF%D8%A7%D9%84%D9%84%D9%87-%D8%A7%D9%84%D8%A5%D9%86%D8%B3%D8%A7%D9%86-%D8%A7%D9%84%D9%85%D9%88%D8%B3%D9%85-1-%D9%83%D9%84%D9%8A%D8%A8-3.html',
|
||||||
|
'info_dict': {
|
||||||
|
'id': '90574',
|
||||||
|
'ext': 'm3u8',
|
||||||
|
'title': 'الملك عبدالله الإنسان الموسم 1 كليب 3',
|
||||||
|
'description': 'الفيلم الوثائقي - الملك عبد الله الإنسان',
|
||||||
|
'duration': 2972,
|
||||||
|
'timestamp': 1422057420,
|
||||||
|
'upload_date': '20150123',
|
||||||
|
},
|
||||||
|
'params': {
|
||||||
|
# m3u8 download
|
||||||
|
'skip_download': True,
|
||||||
|
}
|
||||||
|
}, {
|
||||||
|
# shahid plus subscriber only
|
||||||
|
'url': 'https://shahid.mbc.net/ar/episode/90511/%D9%85%D8%B1%D8%A7%D9%8A%D8%A7-2011-%D8%A7%D9%84%D9%85%D9%88%D8%B3%D9%85-1-%D8%A7%D9%84%D8%AD%D9%84%D9%82%D8%A9-1.html',
|
||||||
|
'only_matching': True
|
||||||
|
}]
|
||||||
|
|
||||||
|
def _handle_error(self, response):
|
||||||
|
if not isinstance(response, dict):
|
||||||
|
return
|
||||||
|
error = response.get('error')
|
||||||
|
if error:
|
||||||
|
raise ExtractorError(
|
||||||
|
'%s returned error: %s' % (self.IE_NAME, '\n'.join(error.values())),
|
||||||
|
expected=True)
|
||||||
|
|
||||||
|
def _download_json(self, url, video_id, note='Downloading JSON metadata'):
|
||||||
|
response = super(ShahidIE, self)._download_json(url, video_id, note)['data']
|
||||||
|
self._handle_error(response)
|
||||||
|
return response
|
||||||
|
|
||||||
|
def _real_extract(self, url):
|
||||||
|
video_id = self._match_id(url)
|
||||||
|
|
||||||
|
webpage = self._download_webpage(url, video_id)
|
||||||
|
|
||||||
|
api_vars = {
|
||||||
|
'id': video_id,
|
||||||
|
'type': 'player',
|
||||||
|
'url': 'http://api.shahid.net/api/v1_1',
|
||||||
|
'playerType': 'episode',
|
||||||
|
}
|
||||||
|
|
||||||
|
flashvars = self._search_regex(
|
||||||
|
r'var\s+flashvars\s*=\s*({[^}]+})', webpage, 'flashvars', default=None)
|
||||||
|
if flashvars:
|
||||||
|
for key in api_vars.keys():
|
||||||
|
value = self._search_regex(
|
||||||
|
r'\b%s\s*:\s*(?P<q>["\'])(?P<value>.+?)(?P=q)' % key,
|
||||||
|
flashvars, 'type', default=None, group='value')
|
||||||
|
if value:
|
||||||
|
api_vars[key] = value
|
||||||
|
|
||||||
|
player = self._download_json(
|
||||||
|
'https://shahid.mbc.net/arContent/getPlayerContent-param-.id-%s.type-%s.html'
|
||||||
|
% (video_id, api_vars['type']), video_id, 'Downloading player JSON')
|
||||||
|
|
||||||
|
formats = self._extract_m3u8_formats(player['url'], video_id, 'mp4')
|
||||||
|
|
||||||
|
video = self._download_json(
|
||||||
|
'%s/%s/%s?%s' % (
|
||||||
|
api_vars['url'], api_vars['playerType'], api_vars['id'],
|
||||||
|
compat_urllib_parse.urlencode({
|
||||||
|
'apiKey': 'sh@hid0nlin3',
|
||||||
|
'hash': 'b2wMCTHpSmyxGqQjJFOycRmLSex+BpTK/ooxy6vHaqs=',
|
||||||
|
}).encode('utf-8')),
|
||||||
|
video_id, 'Downloading video JSON')
|
||||||
|
|
||||||
|
video = video[api_vars['playerType']]
|
||||||
|
|
||||||
|
title = video['title']
|
||||||
|
description = video.get('description')
|
||||||
|
thumbnail = video.get('thumbnailUrl')
|
||||||
|
duration = int_or_none(video.get('duration'))
|
||||||
|
timestamp = parse_iso8601(video.get('referenceDate'))
|
||||||
|
categories = [
|
||||||
|
category['name']
|
||||||
|
for category in video.get('genres', []) if 'name' in category]
|
||||||
|
|
||||||
|
return {
|
||||||
|
'id': video_id,
|
||||||
|
'title': title,
|
||||||
|
'description': description,
|
||||||
|
'thumbnail': thumbnail,
|
||||||
|
'duration': duration,
|
||||||
|
'timestamp': timestamp,
|
||||||
|
'categories': categories,
|
||||||
|
'formats': formats,
|
||||||
|
}
|
Loading…
Reference in a new issue