[twitch] Fix authentication (closes #17024)

[youtube-dl] / youtube_dl / extractor / twitch.py

diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py
index ec96ae5064e759b8632620c4462a894cb9c6435c..cbb7354039b86c4ac08fb8d4c8bb3d5174e084a0 100644 (file)
--- a/youtube_dl/extractor/twitch.py
+++ b/youtube_dl/extractor/twitch.py
@@ -4,6 +4,7 @@ from __future__ import unicode_literals
 import itertools
 import re
 import random
+import json
 
 from .common import InfoExtractor
 from ..compat import (
@@ -26,7 +27,7 @@ from ..utils import (
     try_get,
     unified_timestamp,
     update_url_query,
-    urlencode_postdata,
+    url_or_none,
     urljoin,
 )
 
@@ -36,7 +37,8 @@ class TwitchBaseIE(InfoExtractor):
 
     _API_BASE = 'https://api.twitch.tv'
     _USHER_BASE = 'https://usher.ttvnw.net'
-    _LOGIN_URL = 'https://www.twitch.tv/login'
+    _LOGIN_FORM_URL = 'https://www.twitch.tv/login'
+    _LOGIN_POST_URL = 'https://passport.twitch.tv/login'
     _CLIENT_ID = 'jzkbprff40iqj646a697cyrvl0zt2m6'
     _NETRC_MACHINE = 'twitch'
 
@@ -61,7 +63,7 @@ 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
 
@@ -76,21 +78,25 @@ class TwitchBaseIE(InfoExtractor):
             page_url = urlh.geturl()
             post_url = self._search_regex(
                 r'<form[^>]+action=(["\'])(?P<url>.+?)\1', page,
-                'post url', default=page_url, group='url')
+                'post url', default=self._LOGIN_POST_URL, group='url')
             post_url = urljoin(page_url, post_url)
 
-            headers = {'Referer': page_url}
+            headers = {
+                'Referer': page_url,
+                'Origin': page_url,
+                'Content-Type': 'text/plain;charset=UTF-8'
+            }
 
             try:
                 response = self._download_json(
                     post_url, None, note,
-                    data=urlencode_postdata(form),
+                    data=json.dumps(form).encode(),
                     headers=headers)
             except ExtractorError as e:
                 if isinstance(e.cause, compat_HTTPError) and e.cause.code == 400:
                     response = self._parse_json(
                         e.cause.read().decode('utf-8'), None)
-                    fail(response.get('message') or response['errors'][0])
+                    fail(response.get('error_description') or response.get('error_code'))
                 raise
 
             if 'Authenticated successfully' in response.get('message', ''):
@@ -104,7 +110,7 @@ class TwitchBaseIE(InfoExtractor):
                 headers=headers)
 
         login_page, handle = self._download_webpage_handle(
-            self._LOGIN_URL, None, 'Downloading login page')
+            self._LOGIN_FORM_URL, None, 'Downloading login page')
 
         # Some TOR nodes and public proxies are blocked completely
         if 'blacklist_message' in login_page:
@@ -114,6 +120,7 @@ class TwitchBaseIE(InfoExtractor):
             login_page, handle, 'Logging in', {
                 'username': username,
                 'password': password,
+                'client_id': self._CLIENT_ID,
             })
 
         # Successful login
@@ -239,7 +246,7 @@ class TwitchVodIE(TwitchItemBaseIE):
     _VALID_URL = r'''(?x)
                     https?://
                         (?:
-                            (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v|videos)/|
+                            (?:(?:www|go|m)\.)?twitch\.tv/(?:[^/]+/v(?:ideo)?|videos)/|
                             player\.twitch\.tv/\?.*?\bvideo=v
                         )
                         (?P<id>\d+)
@@ -295,6 +302,9 @@ class TwitchVodIE(TwitchItemBaseIE):
     }, {
         'url': 'https://m.twitch.tv/beagsandjam/v/247478721',
         'only_matching': True,
+    }, {
+        'url': 'https://www.twitch.tv/northernlion/video/291940395',
+        'only_matching': True,
     }]
 
     def _real_extract(self, url):
@@ -663,8 +673,8 @@ class TwitchClipsIE(TwitchBaseIE):
         for option in status['quality_options']:
             if not isinstance(option, dict):
                 continue
-            source = option.get('source')
-            if not source or not isinstance(source, compat_str):
+            source = url_or_none(option.get('source'))
+            if not source:
                 continue
             formats.append({
                 'url': source,
@@ -673,6 +683,8 @@ class TwitchClipsIE(TwitchBaseIE):
                 'fps': int_or_none(option.get('frame_rate')),
             })
 
+        self._sort_formats(formats)
+
         info = {
             'formats': formats,
         }