1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-23 02:32:16 +00:00
youtube-dl/youtube_dl/extractor/fivemin.py

89 lines
3 KiB
Python
Raw Normal View History

from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
compat_str,
compat_urllib_parse,
ExtractorError,
)
class FiveMinIE(InfoExtractor):
IE_NAME = '5min'
_VALID_URL = r'''(?x)
(?:https?://[^/]*?5min\.com/Scripts/PlayerSeed\.js\?(.*?&)?playList=|
5min:)
(?P<id>\d+)
'''
_TESTS = [
{
# From http://www.engadget.com/2013/11/15/ipad-mini-retina-display-review/
'url': 'http://pshared.5min.com/Scripts/PlayerSeed.js?sid=281&width=560&height=345&playList=518013791',
'md5': '4f7b0b79bf1a470e5004f7112385941d',
'info_dict': {
'id': '518013791',
'ext': 'mp4',
'title': 'iPad Mini with Retina Display Review',
},
},
{
# From http://on.aol.com/video/how-to-make-a-next-level-fruit-salad-518086247
'url': '5min:518086247',
'md5': 'e539a9dd682c288ef5a498898009f69e',
'info_dict': {
'id': '518086247',
'ext': 'mp4',
'title': 'How to Make a Next-Level Fruit Salad',
},
},
]
@classmethod
def _build_result(cls, video_id):
return cls.url_result('5min:%s' % video_id, cls.ie_key())
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
embed_url = 'https://embed.5min.com/playerseed/?playList=%s' % video_id
embed_page = self._download_webpage(embed_url, video_id,
'Downloading embed page')
sid = self._search_regex(r'sid=(\d+)', embed_page, 'sid')
query = compat_urllib_parse.urlencode({
'func': 'GetResults',
'playlist': video_id,
'sid': sid,
'isPlayerSeed': 'true',
'url': embed_url,
})
response = self._download_json(
'https://syn.5min.com/handlers/SenseHandler.ashx?' + query,
video_id)
if not response['success']:
err_msg = response['errorMessage']
if err_msg == 'ErrorVideoUserNotGeo':
msg = 'Video not available from your location'
else:
msg = 'Aol said: %s' % err_msg
raise ExtractorError(msg, expected=True, video_id=video_id)
info = response['binding'][0]
second_id = compat_str(int(video_id[:-2]) + 1)
formats = []
for quality, height in [(1, 320), (2, 480), (4, 720), (8, 1080)]:
if any(r['ID'] == quality for r in info['Renditions']):
formats.append({
'format_id': compat_str(quality),
'url': 'http://avideos.5min.com/%s/%s/%s_%s.mp4' % (second_id[-3:], second_id, video_id, quality),
'height': height,
})
return {
'id': video_id,
'title': info['Title'],
'formats': formats,
}