midgard/youtube-dl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[nrk] Improve geo restriction detection and use geo bypass mechanism

Browse source
This commit is contained in:
Sergey M․ 2017-02-04 18:51:59 +07:00 committed by Sergey M
parent 28200e654b
commit ff4007891f
1 changed files with 5 additions and 31 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

36
youtube_dl/extractor/nrk.py
View file

@ -1,7 +1,6 @@
# coding: utf-8 # coding: utf-8
from __future__ import unicode_literals from __future__ import unicode_literals
import random
import re import re
from .common import InfoExtractor from .common import InfoExtractor
@ -15,25 +14,6 @@ from ..utils import (
class NRKBaseIE(InfoExtractor): class NRKBaseIE(InfoExtractor):
_faked_ip = None
def _download_webpage_handle(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:
# NB: str is intentional
kwargs.setdefault(str('headers'), {})['X-Forwarded-For'] = self._faked_ip
return super(NRKBaseIE, self)._download_webpage_handle(*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): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)
@ -44,8 +24,6 @@ class NRKBaseIE(InfoExtractor):
title = data.get('fullTitle') or data.get('mainTitle') or data['title'] title = data.get('fullTitle') or data.get('mainTitle') or data['title']
video_id = data.get('id') or video_id video_id = data.get('id') or video_id
http_headers = {'X-Forwarded-For': self._faked_ip} if self._faked_ip else {}
entries = [] entries = []
conviva = data.get('convivaStatistics') or {} conviva = data.get('convivaStatistics') or {}
@ -90,7 +68,6 @@ class NRKBaseIE(InfoExtractor):
'duration': duration, 'duration': duration,
'subtitles': subtitles, 'subtitles': subtitles,
'formats': formats, 'formats': formats,
'http_headers': http_headers,
}) })
if not entries: if not entries:
@ -107,19 +84,16 @@ class NRKBaseIE(InfoExtractor):
}] }]
if not entries: if not entries:
message_type = data.get('messageType', '')
# Can be ProgramIsGeoBlocked or ChannelIsGeoBlocked*
if 'IsGeoBlocked' in message_type 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 = { MESSAGES = {
'ProgramRightsAreNotReady': 'Du kan dessverre ikke se eller høre programmet', 'ProgramRightsAreNotReady': 'Du kan dessverre ikke se eller høre programmet',
'ProgramRightsHasExpired': 'Programmet har gått ut', 'ProgramRightsHasExpired': 'Programmet har gått ut',
'ProgramIsGeoBlocked': 'NRK har ikke rettigheter til å vise dette programmet utenfor Norge', 'ProgramIsGeoBlocked': 'NRK har ikke rettigheter til å vise dette programmet utenfor Norge',
} }
message_type = data.get('messageType', '')
# Can be ProgramIsGeoBlocked or ChannelIsGeoBlocked*
if 'IsGeoBlocked' in message_type:
self.raise_geo_restricted(
msg=MESSAGES.get('ProgramIsGeoBlocked'), countries=['NO'])
raise ExtractorError( raise ExtractorError(
'%s said: %s' % (self.IE_NAME, MESSAGES.get( '%s said: %s' % (self.IE_NAME, MESSAGES.get(
message_type, message_type)), message_type, message_type)),