Merge 2be0c24897 into f2a4983df7

Closes #11526
Authored by: jshumphrey

yt_dlp/extractor/_extractors.py

@@ -1983,6 +1983,7 @@ from .stretchinternet import StretchInternetIE
 from .stripchat import StripchatIE
 from .stv import STVPlayerIE
 from .substack import SubstackIE
+from .suno import SunoIE, SunoPlaylistIE
 from .sunporno import SunPornoIE
 from .sverigesradio import (
     SverigesRadioEpisodeIE,

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/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/suno.py

@@ -0,0 +1,152 @@
+import re
+
+from .common import InfoExtractor
+from ..utils import ExtractorError, unescapeHTML
+
+
+class SunoBaseIE(InfoExtractor):
+    def _get_title(self, webpage):
+        return self._html_search_meta(
+            ['og:title', 'twitter:title'], webpage, 'title',
+            default=None) or self._html_extract_title(webpage)
+
+    def _get_description(self, webpage):
+        return self._html_search_meta(
+            ['og:description', 'description', 'twitter:description'],
+            webpage, 'description', default=None)
+
+    def _get_thumbnail(self, webpage):
+        return self._html_search_meta(
+            ['og:image', 'twitter:image'], webpage, 'thumbnail', default=None)
+
+
+class SunoIE(SunoBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?suno\.com/song/(?P<id>[-a-f0-9]+)'
+    _TESTS = [
+        {
+            'url': 'https://suno.com/song/ab39a04d-b2e6-463b-9b8e-ddea725422f5',
+            'md5': 'ef850763b175d8a3c7fba5e2dbdc6bc5',
+            'info_dict': {
+                'id': 'ab39a04d-b2e6-463b-9b8e-ddea725422f5',
+                'title': 'Life\'s a Soundtrack · AI Funk Factory @ YT by @funk | Suno',
+                'description': 'groovy funk, melodic song. Listen and make your own with Suno.',
+                'thumbnail': r're:https?://.*903f2bd7-ccc0-4029-a76a-887f07ebc2df.*\.jpeg$',
+                'ext': 'mp3',
+            },
+        },
+        {
+            'url': 'https://suno.com/song/9cbcb5f4-f367-4f1c-8a32-23ec62bdc47e',
+            'md5': '2f038badef88d189891d5f8cd8d8804d',
+            'info_dict': {
+                'id': '9cbcb5f4-f367-4f1c-8a32-23ec62bdc47e',
+                'title': 'Pequenos Prazeres da Vida by @groovebot | Suno',
+                'description': 'pop bossa nova song. Listen and make your own with Suno.',
+                'thumbnail': r're:https?://.*9cbcb5f4-f367-4f1c-8a32-23ec62bdc47e.*\.jpeg$',
+                'ext': 'mp3',
+            },
+        },
+    ]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+
+        url = self._html_search_meta(
+            ['og:audio', 'twitter:player:stream'], webpage, 'url', default=None)
+
+        return {
+            'id': video_id,
+            'title': self._get_title(webpage),
+            'description': self._get_description(webpage),
+            'thumbnail': self._get_thumbnail(webpage),
+            'url': url,
+        }
+
+
+class SunoPlaylistIE(SunoBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?suno\.com/playlist/(?P<id>[-a-f0-9]+)'
+    _TESTS = [
+        {
+            'url': 'https://suno.com/playlist/01f2ac32-c32e-4d26-b10c-221107c02946',
+            'info_dict': {
+                'id': '01f2ac32-c32e-4d26-b10c-221107c02946',
+                'title': 'Main 0 by @contemplativetranspositions367 | Suno',
+                'description': 'Hopefully the test case passed',
+                'thumbnail': r're:https?://.*19d6d518-1b87-43b3-90b9-2a476ca5824a.*\.jpeg$',
+            },
+            'playlist': [{
+                'info_dict': {
+                    'id': '19d6d518-1b87-43b3-90b9-2a476ca5824a',
+                    'title': 'Ceaseless <Echoes>',
+                    'ext': 'mp3',
+                },
+            }],
+            'playlist_count': 1,
+        },
+        {
+            'url': 'https://www.suno.com/playlist/568eeaab-dfbf-4da6-aa0a-0fb1a32330de',
+            'info_dict': {
+                'id': '568eeaab-dfbf-4da6-aa0a-0fb1a32330de',
+                'title': 'Piano by @kunal | Suno',
+                'description': 'Here are some good piano',
+                'thumbnail': r're:https?://.*0ecc0956-3b17-4d4b-8504-55849dd75e22.*\.jpeg$',
+            },
+            'playlist': [
+                {
+                    'info_dict': {
+                        'id': '0ecc0956-3b17-4d4b-8504-55849dd75e22',
+                        'title': 'ST',
+                        'ext': 'mp3',
+                    },
+                },
+                {
+                    'info_dict': {
+                        'id': '3fef7d44-c5a3-4181-9de3-d81542af23ef',
+                        'title': 'ST',
+                        'ext': 'mp3',
+                    },
+                },
+                {
+                    'info_dict': {
+                        'id': '15e797fa-06c0-4e11-8cc0-3b2580476039',
+                        'title': 'ST - 2',
+                        'ext': 'mp3',
+                    },
+                },
+            ],
+            'playlist_count': 3,
+        },
+    ]
+
+    def _real_extract(self, url):
+        playlist_id = self._match_id(url)
+        webpage = self._download_webpage(url, playlist_id)
+
+        # There are <a>s whose href is a song/ID path. The <span>s directly
+        # within them have the respective song title as their innerHTML.
+        # Alternatively, this info can be extracted through parsing an escaped
+        # JSON object inside a <script> array, though that seems even less stable
+        # than this HTML.
+        songs_regex = r'/song/(?P<id>[-a-f0-9]+)["\'][^>]*>\s*<span[^>]*>\s*(?P<title>[^<]+)</span>'
+        songs = re.findall(songs_regex, webpage)
+
+        og_audio_regex = self._og_regexes('audio')[0]
+        audio_urls = [matches[0] for matches in re.findall(og_audio_regex, webpage)]
+
+        if len(songs) != len(audio_urls):
+            raise ExtractorError('Unexpected mismatch between song HTML list and og audio URLs')
+
+        return {
+            '_type': 'playlist',
+            'id': playlist_id,
+            'title': self._get_title(webpage),
+            'description': self._get_description(webpage),
+            'thumbnail': self._get_thumbnail(webpage),
+
+            'entries': [{
+                'id': song_tuple[0],
+                'title': unescapeHTML(song_tuple[1]),
+                'url': audio_urls[i],
+
+            } for i, song_tuple in enumerate(songs)],
+        }