From ff4007891fde74212eb0898bb04c14b2de92ed03 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sat, 4 Feb 2017 18:51:59 +0700 Subject: [PATCH] [nrk] Improve geo restriction detection and use geo bypass mechanism --- youtube_dl/extractor/nrk.py | 36 +++++------------------------------- 1 file changed, 5 insertions(+), 31 deletions(-) diff --git a/youtube_dl/extractor/nrk.py b/youtube_dl/extractor/nrk.py index fc3c0cd3cc..78ece33e12 100644 --- a/youtube_dl/extractor/nrk.py +++ b/youtube_dl/extractor/nrk.py @@ -1,7 +1,6 @@ # coding: utf-8 from __future__ import unicode_literals -import random import re from .common import InfoExtractor @@ -15,25 +14,6 @@ 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): video_id = self._match_id(url) @@ -44,8 +24,6 @@ def _real_extract(self, url): title = data.get('fullTitle') or data.get('mainTitle') or data['title'] video_id = data.get('id') or video_id - http_headers = {'X-Forwarded-For': self._faked_ip} if self._faked_ip else {} - entries = [] conviva = data.get('convivaStatistics') or {} @@ -90,7 +68,6 @@ def video_id_and_title(idx): 'duration': duration, 'subtitles': subtitles, 'formats': formats, - 'http_headers': http_headers, }) if not entries: @@ -107,19 +84,16 @@ def video_id_and_title(idx): }] 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 = { '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', } + 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( '%s said: %s' % (self.IE_NAME, MESSAGES.get( message_type, message_type)),