youtube-dl

commit 84b91dd4e3cd45498fb6cd3c7aa611816f61a3d5
parent 92c9c2a88b8bc6e19b6fc9ac5e3814aa75c437e9
Author: Yen Chi Hsuan <yan12125@gmail.com>
Date:   Wed,  7 Sep 2016 23:07:50 +0800

[gamestar] Fix metadata extraction (closes #10479)

Diffstat:
M
ChangeLog
 | 
1
+
M
youtube_dl/extractor/gamestar.py
 | 
51
+++++++++++++++++----------------------------------

2 files changed, 18 insertions(+), 34 deletions(-)
diff --git a/ChangeLog b/ChangeLog
@@ -1,6 +1,7 @@
 version <unreleased>
 
 Extractors
+* [gamestar] Fix metadata extraction (#10479)
 + [bilibili] Support episodes (#10190)
 + [tvnoe] New extractor (#10524)
 
diff --git a/youtube_dl/extractor/gamestar.py b/youtube_dl/extractor/gamestar.py
@@ -1,14 +1,10 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
 from ..utils import (
     int_or_none,
-    parse_duration,
-    str_to_int,
-    unified_strdate,
+    remove_end,
 )
 
 
@@ -21,8 +17,9 @@ class GameStarIE(InfoExtractor):
             'id': '76110',
             'ext': 'mp4',
             'title': 'Hobbit 3: Die Schlacht der Fünf Heere - Teaser-Trailer zum dritten Teil',
-            'description': 'Der Teaser-Trailer zu Hobbit 3: Die Schlacht der Fünf Heere zeigt einige Szenen aus dem dritten Teil der Saga und kündigt den vollständigen Trailer an.',
-            'thumbnail': 'http://images.gamestar.de/images/idgwpgsgp/bdb/2494525/600x.jpg',
+            'description': 'Der Teaser-Trailer zu Hobbit 3: Die Schlacht der Fünf Heere zeigt einige Szenen aus dem dritten Teil der Saga und kündigt den...',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'timestamp': 1406542020,
             'upload_date': '20140728',
             'duration': 17
         }
@@ -32,41 +29,27 @@ class GameStarIE(InfoExtractor):
         video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
 
-        og_title = self._og_search_title(webpage)
-        title = re.sub(r'\s*- Video (bei|-) GameStar\.de$', '', og_title)
-
         url = 'http://gamestar.de/_misc/videos/portal/getVideoUrl.cfm?premium=0&videoId=' + video_id
 
-        description = self._og_search_description(webpage).strip()
-
-        thumbnail = self._proto_relative_url(
-            self._og_search_thumbnail(webpage), scheme='http:')
-
-        upload_date = unified_strdate(self._html_search_regex(
-            r'<span style="float:left;font-size:11px;">Datum: ([0-9]+\.[0-9]+\.[0-9]+)&nbsp;&nbsp;',
-            webpage, 'upload_date', fatal=False))
-
-        duration = parse_duration(self._html_search_regex(
-            r'&nbsp;&nbsp;Länge: ([0-9]+:[0-9]+)</span>', webpage, 'duration',
-            fatal=False))
-
-        view_count = str_to_int(self._html_search_regex(
-            r'&nbsp;&nbsp;Zuschauer: ([0-9\.]+)&nbsp;&nbsp;', webpage,
-            'view_count', fatal=False))
+        # TODO: there are multiple ld+json objects in the webpage,
+        # while _search_json_ld finds only the first one
+        json_ld = self._parse_json(self._search_regex(
+            r'(?s)<script[^>]+type=(["\'])application/ld\+json\1[^>]*>(?P<json_ld>[^<]+VideoObject[^<]+)</script>',
+            webpage, 'JSON-LD', group='json_ld'), video_id)
+        info_dict = self._json_ld(json_ld, video_id)
+        info_dict['title'] = remove_end(info_dict['title'], ' - GameStar')
 
+        view_count = json_ld.get('interactionCount')
         comment_count = int_or_none(self._html_search_regex(
-            r'>Kommentieren \(([0-9]+)\)</a>', webpage, 'comment_count',
+            r'([0-9]+) Kommentare</span>', webpage, 'comment_count',
             fatal=False))
 
-        return {
+        info_dict.update({
             'id': video_id,
-            'title': title,
             'url': url,
             'ext': 'mp4',
-            'thumbnail': thumbnail,
-            'description': description,
-            'upload_date': upload_date,
-            'duration': duration,
             'view_count': view_count,
             'comment_count': comment_count
-        }
+        })
+
+        return info_dict