[oe1] Add support for oe1.orf.at.
This commit is contained in:
parent
cb3ac1c610
commit
f0da3f1ef9
|
@ -177,6 +177,7 @@ from .normalboots import NormalbootsIE
|
||||||
from .novamov import NovaMovIE
|
from .novamov import NovaMovIE
|
||||||
from .nowness import NownessIE
|
from .nowness import NownessIE
|
||||||
from .nowvideo import NowVideoIE
|
from .nowvideo import NowVideoIE
|
||||||
|
from .oe1 import OE1IE
|
||||||
from .ooyala import OoyalaIE
|
from .ooyala import OoyalaIE
|
||||||
from .orf import ORFIE
|
from .orf import ORFIE
|
||||||
from .parliamentliveuk import ParliamentLiveUKIE
|
from .parliamentliveuk import ParliamentLiveUKIE
|
||||||
|
|
|
@ -0,0 +1,38 @@
|
||||||
|
# coding: utf-8
|
||||||
|
from __future__ import unicode_literals
|
||||||
|
import calendar
|
||||||
|
import datetime
|
||||||
|
import json
|
||||||
|
import re
|
||||||
|
|
||||||
|
from .common import InfoExtractor
|
||||||
|
|
||||||
|
# audios on oe1.orf.at are only available for 7 days, so we can't
|
||||||
|
# add tests.
|
||||||
|
|
||||||
|
|
||||||
|
class OE1IE(InfoExtractor):
|
||||||
|
_VALID_URL = r'http://oe1\.orf\.at/programm/(?P<id>\d+)'
|
||||||
|
|
||||||
|
def _real_extract(self, url):
|
||||||
|
mobj = re.match(self._VALID_URL, url)
|
||||||
|
show_id = mobj.group('id')
|
||||||
|
data = json.loads(self._download_webpage(
|
||||||
|
'http://oe1.orf.at/programm/%s/konsole' % show_id,
|
||||||
|
show_id
|
||||||
|
))
|
||||||
|
|
||||||
|
timestamp = datetime.datetime.strptime('%s %s' % (
|
||||||
|
data['item']['day_label'],
|
||||||
|
data['item']['time']
|
||||||
|
), '%d.%m.%Y %H:%M')
|
||||||
|
unix_timestamp = calendar.timegm(timestamp.utctimetuple())
|
||||||
|
|
||||||
|
return {
|
||||||
|
'id': show_id,
|
||||||
|
'title': data['item']['title'],
|
||||||
|
'url': data['item']['url_stream'],
|
||||||
|
'ext': 'mp3',
|
||||||
|
'description': data['item']['info'],
|
||||||
|
'timestamp': unix_timestamp
|
||||||
|
}
|
Loading…
Reference in New Issue