X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Ftwitch.py;h=0db2dca41c7e7e914c66956987d2f921b81b1791;hb=0ec9d4e565c1471c1234634bb3be0c7c7662d864;hp=b7a72a7bdccf74bc3027e0930d262d5c173a4aec;hpb=008661069b0c33614ebdd7c9bcf838256db1b127;p=youtube-dl
diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py
index b7a72a7bd..0db2dca41 100644
--- a/youtube_dl/extractor/twitch.py
+++ b/youtube_dl/extractor/twitch.py
@@ -4,26 +4,39 @@ from __future__ import unicode_literals
import itertools
import re
import random
+import json
from .common import InfoExtractor
from ..compat import (
+ compat_kwargs,
+ compat_parse_qs,
compat_str,
- compat_urllib_parse,
- compat_urllib_request,
+ compat_urllib_parse_urlencode,
+ compat_urllib_parse_urlparse,
)
from ..utils import (
+ clean_html,
ExtractorError,
+ int_or_none,
+ orderedSet,
+ parse_duration,
parse_iso8601,
+ try_get,
+ unified_timestamp,
+ update_url_query,
+ url_or_none,
+ urljoin,
)
class TwitchBaseIE(InfoExtractor):
- _VALID_URL_BASE = r'https?://(?:www\.)?twitch\.tv'
+ _VALID_URL_BASE = r'https?://(?:(?:www|go|m)\.)?twitch\.tv'
_API_BASE = 'https://api.twitch.tv'
- _USHER_BASE = 'http://usher.twitch.tv'
- _LOGIN_URL = 'https://secure.twitch.tv/login'
- _LOGIN_POST_URL = 'https://passport.twitch.tv/authorize'
+ _USHER_BASE = 'https://usher.ttvnw.net'
+ _LOGIN_FORM_URL = 'https://www.twitch.tv/login'
+ _LOGIN_POST_URL = 'https://passport.twitch.tv/login'
+ _CLIENT_ID = 'kimne78kx3ncx6brgo4mv6wki5h1ko'
_NETRC_MACHINE = 'twitch'
def _handle_error(self, response):
@@ -35,16 +48,13 @@ class TwitchBaseIE(InfoExtractor):
'%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
expected=True)
- def _download_json(self, url, video_id, note='Downloading JSON metadata'):
- headers = {
- 'Referer': 'http://api.twitch.tv/crossdomain/receiver.html?v=2',
- 'X-Requested-With': 'XMLHttpRequest',
- }
- for cookie in self._downloader.cookiejar:
- if cookie.name == 'api_token':
- headers['Twitch-Api-Token'] = cookie.value
- request = compat_urllib_request.Request(url, headers=headers)
- response = super(TwitchBaseIE, self)._download_json(request, video_id, note)
+ def _call_api(self, path, item_id, *args, **kwargs):
+ headers = kwargs.get('headers', {}).copy()
+ headers['Client-ID'] = self._CLIENT_ID
+ kwargs['headers'] = headers
+ response = self._download_json(
+ '%s/%s' % (self._API_BASE, path), item_id,
+ *args, **compat_kwargs(kwargs))
self._handle_error(response)
return response
@@ -52,53 +62,97 @@ class TwitchBaseIE(InfoExtractor):
self._login()
def _login(self):
- (username, password) = self._get_login_info()
+ username, password = self._get_login_info()
if username is None:
return
- login_page = self._download_webpage(
- self._LOGIN_URL, None, 'Downloading login page')
+ def fail(message):
+ raise ExtractorError(
+ 'Unable to login. Twitch said: %s' % message, expected=True)
- login_form = dict(re.findall(
- r']+action=(["\'])(?P.+?)\1', page,
+ 'post url', default=self._LOGIN_POST_URL, group='url')
+ post_url = urljoin(page_url, post_url)
- request = compat_urllib_request.Request(
- self._LOGIN_POST_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
- request.add_header('Referer', self._LOGIN_URL)
- response = self._download_webpage(
- request, None, 'Logging in as %s' % username)
+ headers = {
+ 'Referer': page_url,
+ 'Origin': page_url,
+ 'Content-Type': 'text/plain;charset=UTF-8',
+ }
- m = re.search(
- r"id=([\"'])login_error_message\1[^>]*>(?P[^<]+)", response)
- if m:
- raise ExtractorError(
- 'Unable to login: %s' % m.group('msg').strip(), expected=True)
+ response = self._download_json(
+ post_url, None, note, data=json.dumps(form).encode(),
+ headers=headers, expected_status=400)
+ error = response.get('error_description') or response.get('error_code')
+ if error:
+ fail(error)
+
+ if 'Authenticated successfully' in response.get('message', ''):
+ return None, None
+
+ redirect_url = urljoin(
+ post_url,
+ response.get('redirect') or response['redirect_path'])
+ return self._download_webpage_handle(
+ redirect_url, None, 'Downloading login redirect page',
+ headers=headers)
+
+ login_page, handle = self._download_webpage_handle(
+ self._LOGIN_FORM_URL, None, 'Downloading login page')
+
+ # Some TOR nodes and public proxies are blocked completely
+ if 'blacklist_message' in login_page:
+ fail(clean_html(login_page))
+
+ redirect_page, handle = login_step(
+ login_page, handle, 'Logging in', {
+ 'username': username,
+ 'password': password,
+ 'client_id': self._CLIENT_ID,
+ })
+
+ # Successful login
+ if not redirect_page:
+ return
+
+ if re.search(r'(?i)