Merge 180d2d1a9b into f2a4983df7

Closes #11526
Authored by: jshumphrey

yt_dlp/extractor/archiveorg.py

@@ -205,6 +205,26 @@ class ArchiveOrgIE(InfoExtractor):
                 },
             },
         ],
+    }, {
+        # The reviewbody is None for one of the reviews; just need to extract data without crashing
+        'url': 'https://archive.org/details/gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+        'info_dict': {
+            'id': 'gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+            'ext': 'mp3',
+            'title': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'creators': ['Grateful Dead'],
+            'duration': 338.31,
+            'track': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'description': 'md5:764348a470b986f1217ffd38d6ac7b72',
+            'display_id': 'gd95-04-02d1t04.shn',
+            'location': 'Pyramid Arena',
+            'uploader': 'jon@archive.org',
+            'album': '1995-04-02 - Pyramid Arena',
+            'upload_date': '20040519',
+            'track_number': 4,
+            'release_date': '19950402',
+            'timestamp': 1084927901,
+        },
     }]
 
     @staticmethod
@@ -335,7 +355,7 @@ class ArchiveOrgIE(InfoExtractor):
                 info['comments'].append({
                     'id': review.get('review_id'),
                     'author': review.get('reviewer'),
-                    'text': str_or_none(review.get('reviewtitle'), '') + '\n\n' + review.get('reviewbody'),
+                    'text': join_nonempty('reviewtitle', 'reviewbody', from_dict=review, delim='\n\n'),
                     'timestamp': unified_timestamp(review.get('createdate')),
                     'parent': 'root'})
 

yt_dlp/extractor/chaturbate.py

@@ -9,7 +9,7 @@ from ..utils import (
 
 
 class ChaturbateIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:[^/]+\.)?chaturbate\.com/(?:fullvideo/?\?.*?\bb=)?(?P<id>[^/?&#]+)'
+    _VALID_URL = r'https?://(?:[^/]+\.)?chaturbate\.(?P<tld>com|eu|global)/(?:fullvideo/?\?.*?\bb=)?(?P<id>[^/?&#]+)'
     _TESTS = [{
         'url': 'https://www.chaturbate.com/siswet19/',
         'info_dict': {
@@ -29,15 +29,24 @@ class ChaturbateIE(InfoExtractor):
     }, {
         'url': 'https://en.chaturbate.com/siswet19/',
         'only_matching': True,
+    }, {
+        'url': 'https://chaturbate.eu/siswet19/',
+        'only_matching': True,
+    }, {
+        'url': 'https://chaturbate.eu/fullvideo/?b=caylin',
+        'only_matching': True,
+    }, {
+        'url': 'https://chaturbate.global/siswet19/',
+        'only_matching': True,
     }]
 
     _ROOM_OFFLINE = 'Room is currently offline'
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
+        video_id, tld = self._match_valid_url(url).group('id', 'tld')
 
         webpage = self._download_webpage(
-            f'https://chaturbate.com/{video_id}/', video_id,
+            f'https://chaturbate.{tld}/{video_id}/', video_id,
             headers=self.geo_verification_headers())
 
         found_m3u8_urls = []

yt_dlp/extractor/facebook.py

@@ -563,13 +563,13 @@ class FacebookIE(InfoExtractor):
                 return extract_video_data(try_get(
                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 
-        def extract_dash_manifest(video, formats):
+        def extract_dash_manifest(vid_data, formats, mpd_url=None):
             dash_manifest = traverse_obj(
-                video, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', expected_type=str)
+                vid_data, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', 'manifest_xml', expected_type=str)
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
-                    mpd_url=url_or_none(video.get('dash_manifest_url'))))
+                    mpd_url=url_or_none(video.get('dash_manifest_url')) or mpd_url))
 
         def process_formats(info):
             # Downloads with browser's User-Agent are rate limited. Working around
@@ -619,9 +619,12 @@ class FacebookIE(InfoExtractor):
                         video = video['creation_story']
                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
                         video.update(reel_info)
-                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
+
                     formats = []
                     q = qualities(['sd', 'hd'])
+
+                    # Legacy formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
                                            ('browser_native_sd_url', 'sd')):
@@ -629,7 +632,7 @@ class FacebookIE(InfoExtractor):
                         if not playable_url:
                             continue
                         if determine_ext(playable_url) == 'mpd':
-                            formats.extend(self._extract_mpd_formats(playable_url, video_id))
+                            formats.extend(self._extract_mpd_formats(playable_url, video_id, fatal=False))
                         else:
                             formats.append({
                                 'format_id': format_id,
@@ -638,6 +641,28 @@ class FacebookIE(InfoExtractor):
                                 'url': playable_url,
                             })
                     extract_dash_manifest(fmt_data, formats)
+
+                    # New videoDeliveryResponse formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryResponseFragment', 'videoDeliveryResponseResult'))
+                    mpd_urls = traverse_obj(fmt_data, ('dash_manifest_urls', ..., 'manifest_url', {url_or_none}))
+                    dash_manifests = traverse_obj(fmt_data, ('dash_manifests', lambda _, v: v['manifest_xml']))
+                    for idx, dash_manifest in enumerate(dash_manifests):
+                        extract_dash_manifest(dash_manifest, formats, mpd_url=traverse_obj(mpd_urls, idx))
+                    if not dash_manifests:
+                        # Only extract from MPD URLs if the manifests are not already provided
+                        for mpd_url in mpd_urls:
+                            formats.extend(self._extract_mpd_formats(mpd_url, video_id, fatal=False))
+                    for prog_fmt in traverse_obj(fmt_data, ('progressive_urls', lambda _, v: v['progressive_url'])):
+                        format_id = traverse_obj(prog_fmt, ('metadata', 'quality', {str.lower}))
+                        formats.append({
+                            'format_id': format_id,
+                            # sd, hd formats w/o resolution info should be deprioritized below DASH
+                            'quality': q(format_id) - 3,
+                            'url': prog_fmt['progressive_url'],
+                        })
+                    for m3u8_url in traverse_obj(fmt_data, ('hls_playlist_urls', ..., 'hls_playlist_url', {url_or_none})):
+                        formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', fatal=False, m3u8_id='hls'))
+
                     if not formats:
                         # Do not append false positive entry w/o any formats
                         return

yt_dlp/extractor/zdf.py

@@ -5,7 +5,6 @@ from ..utils import (
     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 @@ class ZDFIE(ZDFBaseIE):
         return self._extract_entry(player['content'], player, content, video_id)
 
     def _extract_mobile(self, video_id):
-        video = self._download_json(
-            f'https://zdf-cdn.live.cellular.de/mediathekV2/document/{video_id}',
-            video_id)
+        video = self._download_mediathekv2_document(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',
-            'entries': "lambda x: not any('xy580-fall1-kindermoerder-gesucht-100' in e['url'] for e in x)",
+            'title': 'Aktenzeichen XY... Ungelöst',
+            '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 @@ class ZDFChannelIE(ZDFBaseIE):
         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(
-            rf'''<div\b[^>]*?\sdata-plusbar-id\s*=\s*(["'])(?P<p_id>[\w-]+)\1[^>]*?\sdata-plusbar-url=\1(?P<url>{ZDFIE._VALID_URL})\1''',
-            webpage)
+        main_video = None
+        playlist_videos = []
+
+        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(
-                (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:
+            return self._teaser_to_url_result(main_video) if main_video else None
+
         self.to_screen(f'Downloading playlist {channel_id} - add --no-playlist to download just the main video')
 
-        def check_video(m):
-            v_ref = self._search_regex(
-                r'''(<a\b[^>]*?\shref\s*=[^>]+?\sdata-target-id\s*=\s*(["']){}\2[^>]*>)'''.format(m.group('p_id')),
-                webpage, 'check id', default='')
-            v_ref = extract_attributes(v_ref)
-            return v_ref.get('data-target-video-type') != 'novideo'
+        thumbnails = (
+            traverse_obj(data, ('document', 'image'))
+            or traverse_obj(data, ('document', 'teaserBild'))
+            or traverse_obj(data, ('stageHeader', 'image'))
+            or {})
 
-        return self.playlist_from_matches(
-            (m.group('url') for m in matches if check_video(m)),
-            channel_id, self._og_search_title(webpage, fatal=False))
+        return self.playlist_result(
+            (self._teaser_to_url_result(video) for video in playlist_videos),
+            playlist_id=channel_id,
+            playlist_title=self._og_search_title(webpage, fatal=False),
+            description=self._get_playlist_description(data),
+            thumbnails=self._convert_thumbnails(thumbnails))