[nrk] Workaround geo restriction and improve error messages

This commit is contained in:
Sergey M․ 2016-11-13 20:54:34 +07:00
parent e58609b22c
commit 754e6c8322
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -1,6 +1,7 @@
# coding: utf-8
from __future__ import unicode_literals
import random
import re
from .common import InfoExtractor
@ -14,6 +15,24 @@
class NRKBaseIE(InfoExtractor):
_faked_ip = None
def _download_webpage(self, *args, **kwargs):
# NRK checks X-Forwarded-For HTTP header in order to figure out the
# origin of the client behind proxy. This allows to bypass geo
# restriction by faking this header's value to some Norway IP.
# We will do so once we encounter any geo restriction error.
if self._faked_ip:
kwargs.setdefault('headers', {})['X-Forwarded-For'] = self._faked_ip
return super(NRKBaseIE, self)._download_webpage(*args, **kwargs)
def _fake_ip(self):
# Use fake IP from 37.191.128.0/17 in order to workaround geo
# restriction
def octet(lb=0, ub=255):
return random.randint(lb, ub)
self._faked_ip = '37.191.%d.%d' % (octet(128), octet())
def _real_extract(self, url):
video_id = self._match_id(url)
@ -70,10 +89,22 @@ def video_id_and_title(idx):
}]
if not entries:
if data.get('usageRights', {}).get('isGeoBlocked'):
raise ExtractorError(
'NRK har ikke rettigheter til å vise dette programmet utenfor Norge',
expected=True)
message_type = data.get('messageType')
if message_type == 'ProgramIsGeoBlocked' and not self._faked_ip:
self.report_warning(
'Video is geo restricted, trying to fake IP')
self._fake_ip()
return self._real_extract(url)
MESSAGES = {
'ProgramRightsAreNotReady': 'Du kan dessverre ikke se eller høre programmet',
'ProgramRightsHasExpired': 'Programmet har gått ut',
'ProgramIsGeoBlocked': 'NRK har ikke rettigheter til å vise dette programmet utenfor Norge',
}
raise ExtractorError(
'%s said: %s' % (self.IE_NAME, MESSAGES.get(
message_type, message_type)),
expected=True)
conviva = data.get('convivaStatistics') or {}
series = conviva.get('seriesName') or data.get('seriesTitle')