[ie/umg:de] Rework extractor (#13373)

Authored by: doe1080

yt_dlp/extractor/umg.py

@@ -1,98 +1,53 @@
 from .common import InfoExtractor
-from ..utils import (
+from ..utils import clean_html
-    int_or_none,
+from ..utils.traversal import find_element, traverse_obj
-    parse_filesize,
-    parse_iso8601,
-)
 
 
 class UMGDeIE(InfoExtractor):
-    _WORKING = False
     IE_NAME = 'umg:de'
     IE_DESC = 'Universal Music Deutschland'
-    _VALID_URL = r'https?://(?:www\.)?universal-music\.de/[^/]+/videos/[^/?#]+-(?P<id>\d+)'
+    _VALID_URL = r'https?://(?:www\.)?universal-music\.de/[^/?#]+/videos/(?P<slug>[^/?#]+-(?P<id>\d+))'
-    _TEST = {
+    _TESTS = [{
         'url': 'https://www.universal-music.de/sido/videos/jedes-wort-ist-gold-wert-457803',
-        'md5': 'ebd90f48c80dcc82f77251eb1902634f',
         'info_dict': {
             'id': '457803',
             'ext': 'mp4',
             'title': 'Jedes Wort ist Gold wert',
+            'artists': ['Sido'],
+            'description': 'md5:df2dbffcff1a74e0a7c9bef4b497aeec',
+            'display_id': 'jedes-wort-ist-gold-wert-457803',
+            'duration': 210.0,
+            'thumbnail': r're:https?://images\.universal-music\.de/img/assets/.+\.jpg',
             'timestamp': 1513591800,
             'upload_date': '20171218',
+            'view_count': int,
         },
-    }
+    }, {
+        'url': 'https://www.universal-music.de/alexander-eder/videos/der-doktor-hat-gesagt-609533',
+        'info_dict': {
+            'id': '609533',
+            'ext': 'mp4',
+            'title': 'Der Doktor hat gesagt',
+            'artists': ['Alexander Eder'],
+            'display_id': 'der-doktor-hat-gesagt-609533',
+            'duration': 146.0,
+            'thumbnail': r're:https?://images\.universal-music\.de/img/assets/.+\.jpg',
+            'timestamp': 1742982100,
+            'upload_date': '20250326',
+        },
+    }]
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
+        display_id, video_id = self._match_valid_url(url).group('slug', 'id')
-        video_data = self._download_json(
+        webpage = self._download_webpage(url, display_id)
-            'https://graphql.universal-music.de/',
-            video_id, query={
-                'query': '''{
-  universalMusic(channel:16) {
-    video(id:%s) {
-      headline
-      formats {
-        formatId
-        url
-        type
-        width
-        height
-        mimeType
-        fileSize
-      }
-      duration
-      createdDate
-    }
-  }
-}''' % video_id})['data']['universalMusic']['video']  # noqa: UP031
-
-        title = video_data['headline']
-        hls_url_template = 'http://mediadelivery.universal-music-services.de/vod/mp4:autofill/storage/' + '/'.join(list(video_id)) + '/content/%s/file/playlist.m3u8'
-
-        thumbnails = []
-        formats = []
-
-        def add_m3u8_format(format_id):
-            formats.extend(self._extract_m3u8_formats(
-                hls_url_template % format_id, video_id, 'mp4',
-                'm3u8_native', m3u8_id='hls', fatal=False))
-
-        for f in video_data.get('formats', []):
-            f_url = f.get('url')
-            mime_type = f.get('mimeType')
-            if not f_url or mime_type == 'application/mxf':
-                continue
-            fmt = {
-                'url': f_url,
-                'width': int_or_none(f.get('width')),
-                'height': int_or_none(f.get('height')),
-                'filesize': parse_filesize(f.get('fileSize')),
-            }
-            f_type = f.get('type')
-            if f_type == 'Image':
-                thumbnails.append(fmt)
-            elif f_type == 'Video':
-                format_id = f.get('formatId')
-                if format_id:
-                    fmt['format_id'] = format_id
-                    if mime_type == 'video/mp4':
-                        add_m3u8_format(format_id)
-                urlh = self._request_webpage(f_url, video_id, fatal=False)
-                if urlh:
-                    first_byte = urlh.read(1)
-                    if first_byte not in (b'F', b'\x00'):
-                        continue
-                    formats.append(fmt)
-        if not formats:
-            for format_id in (867, 836, 940):
-                add_m3u8_format(format_id)
 
         return {
+            **self._search_json_ld(webpage, display_id),
             'id': video_id,
-            'title': title,
+            'artists': traverse_obj(self._html_search_meta('umg-artist-screenname', webpage), (filter, all)),
-            'duration': int_or_none(video_data.get('duration')),
+            # The JSON LD description duplicates the title
-            'timestamp': parse_iso8601(video_data.get('createdDate'), ' '),
+            'description': traverse_obj(webpage, ({find_element(cls='_3Y0Lj')}, {clean_html})),
-            'thumbnails': thumbnails,
+            'display_id': display_id,
-            'formats': formats,
+            'formats': self._extract_m3u8_formats(
+                'https://hls.universal-music.de/get', display_id, 'mp4', query={'id': video_id}),
         }