Merge branch 'shahid' of https://github.com/remitamine/youtube-dl into remitamine...

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

diff --git a/youtube_dl/extractor/twitch.py b/youtube_dl/extractor/twitch.py
index 49535cd80c8183be556aed196094e322ceb41574..a2b6a35aa3c89c9ff7c40c5c3385a285abc9a0e5 100644 (file)
--- a/youtube_dl/extractor/twitch.py
+++ b/youtube_dl/extractor/twitch.py
@@ -7,12 +7,15 @@ import random
 
 from .common import InfoExtractor
 from ..compat import (
+    compat_parse_qs,
     compat_str,
     compat_urllib_parse,
+    compat_urllib_parse_urlparse,
     compat_urllib_request,
 )
 from ..utils import (
     ExtractorError,
+    parse_duration,
     parse_iso8601,
 )
 
@@ -59,9 +62,7 @@ class TwitchBaseIE(InfoExtractor):
         login_page = self._download_webpage(
             self._LOGIN_URL, None, 'Downloading login page')
 
-        login_form = dict(re.findall(
-            r'<input\s+type="hidden"\s+name="([^"]+)"\s+(?:id="[^"]+"\s+)?value="([^"]*)"',
-            login_page))
+        login_form = self._hidden_inputs(login_page)
 
         login_form.update({
             'login': username.encode('utf-8'),
@@ -187,7 +188,7 @@ class TwitchVodIE(TwitchItemBaseIE):
     _ITEM_SHORTCUT = 'v'
 
     _TEST = {
-        'url': 'http://www.twitch.tv/riotgames/v/6528877',
+        'url': 'http://www.twitch.tv/riotgames/v/6528877?t=5m10s',
         'info_dict': {
             'id': 'v6528877',
             'ext': 'mp4',
@@ -199,6 +200,7 @@ class TwitchVodIE(TwitchItemBaseIE):
             'uploader': 'Riot Games',
             'uploader_id': 'riotgames',
             'view_count': int,
+            'start_time': 310,
         },
         'params': {
             # m3u8 download
@@ -218,6 +220,12 @@ class TwitchVodIE(TwitchItemBaseIE):
             item_id, 'mp4')
         self._prefer_source(formats)
         info['formats'] = formats
+
+        parsed_url = compat_urllib_parse_urlparse(url)
+        query = compat_parse_qs(parsed_url.query)
+        if 't' in query:
+            info['start_time'] = parse_duration(query['t'][0])
+
         return info
 
 
@@ -312,9 +320,9 @@ class TwitchBookmarksIE(TwitchPlaylistBaseIE):
 
 class TwitchStreamIE(TwitchBaseIE):
     IE_NAME = 'twitch:stream'
-    _VALID_URL = r'%s/(?P<id>[^/]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
+    _VALID_URL = r'%s/(?P<id>[^/#?]+)/?(?:\#.*)?$' % TwitchBaseIE._VALID_URL_BASE
 
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.twitch.tv/shroomztv',
         'info_dict': {
             'id': '12772022048',
@@ -333,7 +341,10 @@ class TwitchStreamIE(TwitchBaseIE):
             # m3u8 download
             'skip_download': True,
         },
-    }
+    }, {
+        'url': 'http://www.twitch.tv/miracle_doto#profile-0',
+        'only_matching': True,
+    }]
 
     def _real_extract(self, url):
         channel_id = self._match_id(url)
@@ -348,6 +359,12 @@ class TwitchStreamIE(TwitchBaseIE):
                 'http://www.twitch.tv/%s/profile' % channel_id,
                 'TwitchProfile', channel_id)
 
+        # Channel name may be typed if different case than the original channel name
+        # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
+        # an invalid m3u8 URL. Working around by use of original channel name from stream
+        # JSON and fallback to lowercase if it's not available.
+        channel_id = stream.get('channel', {}).get('name') or channel_id.lower()
+
         access_token = self._download_json(
             '%s/api/channels/%s/access_token' % (self._API_BASE, channel_id), channel_id,
             'Downloading channel access token')