Merge 180d2d1a9b into f919729538

yt_dlp/extractor/zdf.py

@@ -5,7 +5,6 @@
     NO_DEFAULT,
     ExtractorError,
     determine_ext,
-    extract_attributes,
     float_or_none,
     int_or_none,
     join_nonempty,
@@ -25,6 +24,11 @@ class ZDFBaseIE(InfoExtractor):
     _GEO_COUNTRIES = ['DE']
     _QUALITIES = ('auto', 'low', 'med', 'high', 'veryhigh', 'hd', 'fhd', 'uhd')
 
+    def _download_mediathekv2_document(self, document_id):
+        return self._download_json(
+            f'https://zdf-prod-futura.zdf.de/mediathekV2/document/{document_id}',
+            document_id)
+
     def _call_api(self, url, video_id, item, api_token=None, referrer=None):
         headers = {}
         if api_token:
@@ -320,9 +324,7 @@ def _extract_regular(self, url, player, video_id):
         return self._extract_entry(player['content'], player, content, video_id)
 
     def _extract_mobile(self, video_id):
-        video = self._download_json(
+        video = self._download_mediathekv2_document(video_id)
-            f'https://zdf-cdn.live.cellular.de/mediathekV2/document/{video_id}',
-            video_id)
 
         formats = []
         formitaeten = try_get(video, lambda x: x['document']['formitaeten'], list)
@@ -387,18 +389,19 @@ class ZDFChannelIE(ZDFBaseIE):
         'info_dict': {
             'id': 'planet-e',
             'title': 'planet e.',
+            'description': 'md5:87e3b9c66a63cf1407ee443d2c4eb88e',
         },
         'playlist_mincount': 50,
     }, {
         'url': 'https://www.zdf.de/gesellschaft/aktenzeichen-xy-ungeloest',
         'info_dict': {
             'id': 'aktenzeichen-xy-ungeloest',
-            'title': 'Aktenzeichen XY... ungelöst',
+            'title': 'Aktenzeichen XY... Ungelöst',
-            'entries': "lambda x: not any('xy580-fall1-kindermoerder-gesucht-100' in e['url'] for e in x)",
+            'description': 'md5:623ede5819c400c6d04943fa8100e6e7',
         },
         'playlist_mincount': 2,
     }, {
-        'url': 'https://www.zdf.de/filme/taunuskrimi/',
+        'url': 'https://www.zdf.de/serien/taunuskrimi/',
         'only_matching': True,
     }]
 
@@ -410,32 +413,72 @@ def _og_search_title(self, webpage, fatal=False):
         title = super()._og_search_title(webpage, fatal=fatal)
         return re.split(r'\s+[-|]\s+ZDF(?:mediathek)?$', title or '')[0] or None
 
+    def _extract_document_id(self, webpage):
+        matches = re.search(r'docId\s*:\s*[\'"](?P<docid>[^\'"]+)[\'"]', webpage)
+        return matches and matches.group('docid')
+
+    def _get_playlist_description(self, page_data):
+        headline = traverse_obj(page_data, ('shortText', 'headline'))
+        text = traverse_obj(page_data, ('shortText', 'text'))
+        if headline is not None and text is not None:
+            return f'{headline}\n\n{text}'
+        return headline or text
+
+    def _convert_thumbnails(self, thumbnails):
+        return [{
+            'id': key,
+            'url': thumbnail_info['url'],
+            'width': int_or_none(thumbnail_info.get('width')),
+            'height': int_or_none(thumbnail_info.get('height')),
+        } for key, thumbnail_info in thumbnails.items() if url_or_none(thumbnail_info.get('url'))]
+
+    def _teaser_to_url_result(self, teaser):
+        return self.url_result(
+            teaser['sharingUrl'], ie=ZDFIE.ie_key(),
+            id=teaser.get('id'), title=teaser.get('titel', ''),
+            thumbnails=self._convert_thumbnails(teaser.get('teaserBild', {})),
+            description=teaser.get('beschreibung'),
+            duration=float_or_none(teaser.get('length')),
+            media_type=teaser.get('currentVideoType') or teaser.get('contentType'),
+            season_number=int_or_none(teaser.get('seasonNumber')),
+            episode_number=int_or_none(teaser.get('episodeNumber')))
+
     def _real_extract(self, url):
         channel_id = self._match_id(url)
 
         webpage = self._download_webpage(url, channel_id)
 
-        matches = re.finditer(
+        main_video = None
-            rf'''<div\b[^>]*?\sdata-plusbar-id\s*=\s*(["'])(?P<p_id>[\w-]+)\1[^>]*?\sdata-plusbar-url=\1(?P<url>{ZDFIE._VALID_URL})\1''',
+        playlist_videos = []
-            webpage)
+
+        document_id = self._extract_document_id(webpage)
+        if document_id is not None:
+            data = self._download_mediathekv2_document(document_id)
+
+            for cluster in data['cluster']:
+                for teaser in cluster['teaser']:
+                    if cluster['type'] == 'teaserContent' and teaser['type'] == 'video':
+                        main_video = main_video or teaser
+                    elif cluster['type'] == 'teaser' and teaser['type'] == 'video':
+                        if teaser['brandId'] != document_id:
+                            # These are unrelated 'You might also like' videos, filter them out
+                            continue
+                        playlist_videos.append(teaser)
 
         if self._downloader.params.get('noplaylist', False):
-            entry = next(
+            return self._teaser_to_url_result(main_video) if main_video else None
-                (self.url_result(m.group('url'), ie=ZDFIE.ie_key()) for m in matches),
+
-                None)
-            self.to_screen('Downloading just the main video because of --no-playlist')
-            if entry:
-                return entry
-        else:
         self.to_screen(f'Downloading playlist {channel_id} - add --no-playlist to download just the main video')
 
-        def check_video(m):
+        thumbnails = (
-            v_ref = self._search_regex(
+            traverse_obj(data, ('document', 'image'))
-                r'''(<a\b[^>]*?\shref\s*=[^>]+?\sdata-target-id\s*=\s*(["']){}\2[^>]*>)'''.format(m.group('p_id')),
+            or traverse_obj(data, ('document', 'teaserBild'))
-                webpage, 'check id', default='')
+            or traverse_obj(data, ('stageHeader', 'image'))
-            v_ref = extract_attributes(v_ref)
+            or {})
-            return v_ref.get('data-target-video-type') != 'novideo'
 
-        return self.playlist_from_matches(
+        return self.playlist_result(
-            (m.group('url') for m in matches if check_video(m)),
+            (self._teaser_to_url_result(video) for video in playlist_videos),
-            channel_id, self._og_search_title(webpage, fatal=False))
+            playlist_id=channel_id,
+            playlist_title=self._og_search_title(webpage, fatal=False),
+            description=self._get_playlist_description(data),
+            thumbnails=self._convert_thumbnails(thumbnails))