yt_dlp/extractor/bluesky.py

   1 from .common import InfoExtractor
   2 from ..utils import (
   3     ExtractorError,
   4     format_field,
   5     int_or_none,
   6     mimetype2ext,
   7     orderedSet,
   8     parse_iso8601,
   9     truncate_string,
  10     update_url_query,
  11     url_basename,
  12     url_or_none,
  13     variadic,
  14 )
  15 from ..utils.traversal import traverse_obj
  16
  17
  18 class BlueskyIE(InfoExtractor):
  19     _VALID_URL = [
  20         r'https?://(?:www\.)?(?:bsky\.app|main\.bsky\.dev)/profile/(?P<handle>[\w.:%-]+)/post/(?P<id>\w+)',
  21         r'at://(?P<handle>[\w.:%-]+)/app\.bsky\.feed\.post/(?P<id>\w+)',
  22     ]
  23     _TESTS = [{
  24         'url': 'https://bsky.app/profile/blu3blue.bsky.social/post/3l4omssdl632g',
  25         'md5': '375539c1930ab05d15585ed772ab54fd',
  26         'info_dict': {
  27             'id': '3l4omssdl632g',
  28             'ext': 'mp4',
  29             'uploader': 'Blu3Blu3Lilith',
  30             'uploader_id': 'blu3blue.bsky.social',
  31             'uploader_url': 'https://bsky.app/profile/blu3blue.bsky.social',
  32             'channel_id': 'did:plc:pzdr5ylumf7vmvwasrpr5bf2',
  33             'channel_url': 'https://bsky.app/profile/did:plc:pzdr5ylumf7vmvwasrpr5bf2',
  34             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
  35             'title': 'OMG WE HAVE VIDEOS NOW',
  36             'description': 'OMG WE HAVE VIDEOS NOW',
  37             'upload_date': '20240921',
  38             'timestamp': 1726940605,
  39             'like_count': int,
  40             'repost_count': int,
  41             'comment_count': int,
  42             'tags': [],
  43         },
  44     }, {
  45         'url': 'https://bsky.app/profile/bsky.app/post/3l3vgf77uco2g',
  46         'md5': 'b9e344fdbce9f2852c668a97efefb105',
  47         'info_dict': {
  48             'id': '3l3vgf77uco2g',
  49             'ext': 'mp4',
  50             'uploader': 'Bluesky',
  51             'uploader_id': 'bsky.app',
  52             'uploader_url': 'https://bsky.app/profile/bsky.app',
  53             'channel_id': 'did:plc:z72i7hdynmk6r22z27h6tvur',
  54             'channel_url': 'https://bsky.app/profile/did:plc:z72i7hdynmk6r22z27h6tvur',
  55             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
  56             'title': 'Bluesky now has video! Update your app to versi...',
  57             'alt_title': 'Bluesky video feature announcement',
  58             'description': r're:(?s)Bluesky now has video! .{239}',
  59             'upload_date': '20240911',
  60             'timestamp': 1726074716,
  61             'like_count': int,
  62             'repost_count': int,
  63             'comment_count': int,
  64             'tags': [],
  65             'subtitles': {
  66                 'en': 'mincount:1',
  67             },
  68         },
  69     }, {
  70         'url': 'https://main.bsky.dev/profile/souris.moe/post/3l4qhp7bcs52c',
  71         'md5': '5f2df8c200b5633eb7fb2c984d29772f',
  72         'info_dict': {
  73             'id': '3l4qhp7bcs52c',
  74             'ext': 'mp4',
  75             'uploader': 'souris',
  76             'uploader_id': 'souris.moe',
  77             'uploader_url': 'https://bsky.app/profile/souris.moe',
  78             'channel_id': 'did:plc:tj7g244gl5v6ai6cm4f4wlqp',
  79             'channel_url': 'https://bsky.app/profile/did:plc:tj7g244gl5v6ai6cm4f4wlqp',
  80             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
  81             'title': 'Bluesky video #3l4qhp7bcs52c',
  82             'upload_date': '20240922',
  83             'timestamp': 1727003838,
  84             'like_count': int,
  85             'repost_count': int,
  86             'comment_count': int,
  87             'tags': [],
  88         },
  89     }, {
  90         'url': 'https://bsky.app/profile/de1.pds.tentacle.expert/post/3l3w4tnezek2e',
  91         'md5': '1af9c7fda061cf7593bbffca89e43d1c',
  92         'info_dict': {
  93             'id': '3l3w4tnezek2e',
  94             'ext': 'mp4',
  95             'uploader': 'clean',
  96             'uploader_id': 'de1.pds.tentacle.expert',
  97             'uploader_url': 'https://bsky.app/profile/de1.pds.tentacle.expert',
  98             'channel_id': 'did:web:de1.tentacle.expert',
  99             'channel_url': 'https://bsky.app/profile/did:web:de1.tentacle.expert',
 100             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 101             'title': 'Bluesky video #3l3w4tnezek2e',
 102             'upload_date': '20240911',
 103             'timestamp': 1726098823,
 104             'like_count': int,
 105             'repost_count': int,
 106             'comment_count': int,
 107             'tags': [],
 108         },
 109     }, {
 110         'url': 'https://bsky.app/profile/yunayuispink.bsky.social/post/3l7gqcfes742o',
 111         'info_dict': {
 112             'id': 'XxK3t_5V3ao',
 113             'ext': 'mp4',
 114             'uploader': 'yunayu',
 115             'uploader_id': '@yunayuispink',
 116             'uploader_url': 'https://www.youtube.com/@yunayuispink',
 117             'channel': 'yunayu',
 118             'channel_id': 'UCPLvXnHa7lTyNoR_dGsU14w',
 119             'channel_url': 'https://www.youtube.com/channel/UCPLvXnHa7lTyNoR_dGsU14w',
 120             'thumbnail': 'https://i.ytimg.com/vi_webp/XxK3t_5V3ao/maxresdefault.webp',
 121             'description': r're:Have a good goodx10000day',
 122             'title': '5min vs 5hours drawing',
 123             'availability': 'public',
 124             'live_status': 'not_live',
 125             'playable_in_embed': True,
 126             'upload_date': '20241026',
 127             'timestamp': 1729967784,
 128             'duration': 321,
 129             'age_limit': 0,
 130             'like_count': int,
 131             'view_count': int,
 132             'comment_count': int,
 133             'channel_follower_count': int,
 134             'categories': ['Entertainment'],
 135             'tags': [],
 136         },
 137         'add_ie': ['Youtube'],
 138     }, {
 139         'url': 'https://bsky.app/profile/endshark.bsky.social/post/3jzxjkcemae2m',
 140         'info_dict': {
 141             'id': '222792849',
 142             'ext': 'mp3',
 143             'uploader': 'LASERBAT',
 144             'uploader_id': 'laserbatx',
 145             'uploader_url': 'https://laserbatx.bandcamp.com',
 146             'artists': ['LASERBAT'],
 147             'album_artists': ['LASERBAT'],
 148             'album': 'Hari Nezumi [EP]',
 149             'track': 'Forward to the End',
 150             'title': 'LASERBAT - Forward to the End',
 151             'thumbnail': 'https://f4.bcbits.com/img/a2507705510_5.jpg',
 152             'duration': 228.571,
 153             'track_id': '222792849',
 154             'release_date': '20230423',
 155             'upload_date': '20230423',
 156             'timestamp': 1682276040.0,
 157             'release_timestamp': 1682276040.0,
 158             'track_number': 1,
 159         },
 160         'add_ie': ['Bandcamp'],
 161     }, {
 162         'url': 'https://bsky.app/profile/dannybhoix.bsky.social/post/3l6oe5mtr2c2j',
 163         'md5': 'b9e344fdbce9f2852c668a97efefb105',
 164         'info_dict': {
 165             'id': '3l3vgf77uco2g',
 166             'ext': 'mp4',
 167             'uploader': 'Bluesky',
 168             'uploader_id': 'bsky.app',
 169             'uploader_url': 'https://bsky.app/profile/bsky.app',
 170             'channel_id': 'did:plc:z72i7hdynmk6r22z27h6tvur',
 171             'channel_url': 'https://bsky.app/profile/did:plc:z72i7hdynmk6r22z27h6tvur',
 172             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 173             'title': 'Bluesky now has video! Update your app to versi...',
 174             'alt_title': 'Bluesky video feature announcement',
 175             'description': r're:(?s)Bluesky now has video! .{239}',
 176             'upload_date': '20240911',
 177             'timestamp': 1726074716,
 178             'like_count': int,
 179             'repost_count': int,
 180             'comment_count': int,
 181             'tags': [],
 182             'subtitles': {
 183                 'en': 'mincount:1',
 184             },
 185         },
 186     }, {
 187         'url': 'https://bsky.app/profile/alt.bun.how/post/3l7rdfxhyds2f',
 188         'md5': '8775118b235cf9fa6b5ad30f95cda75c',
 189         'info_dict': {
 190             'id': '3l7rdfxhyds2f',
 191             'ext': 'mp4',
 192             'uploader': 'cinnamon',
 193             'uploader_id': 'alt.bun.how',
 194             'uploader_url': 'https://bsky.app/profile/alt.bun.how',
 195             'channel_id': 'did:plc:7x6rtuenkuvxq3zsvffp2ide',
 196             'channel_url': 'https://bsky.app/profile/did:plc:7x6rtuenkuvxq3zsvffp2ide',
 197             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 198             'title': 'crazy that i look like this tbh',
 199             'description': 'crazy that i look like this tbh',
 200             'upload_date': '20241030',
 201             'timestamp': 1730332128,
 202             'like_count': int,
 203             'repost_count': int,
 204             'comment_count': int,
 205             'tags': ['sexual'],
 206             'age_limit': 18,
 207         },
 208     }, {
 209         'url': 'at://did:plc:ia76kvnndjutgedggx2ibrem/app.bsky.feed.post/3l6zrz6zyl2dr',
 210         'md5': '71b0eb6d85d03145e6af6642c7fc6d78',
 211         'info_dict': {
 212             'id': '3l6zrz6zyl2dr',
 213             'ext': 'mp4',
 214             'uploader': 'mary🐇',
 215             'uploader_id': 'mary.my.id',
 216             'uploader_url': 'https://bsky.app/profile/mary.my.id',
 217             'channel_id': 'did:plc:ia76kvnndjutgedggx2ibrem',
 218             'channel_url': 'https://bsky.app/profile/did:plc:ia76kvnndjutgedggx2ibrem',
 219             'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 220             'title': 'Bluesky video #3l6zrz6zyl2dr',
 221             'upload_date': '20241021',
 222             'timestamp': 1729523172,
 223             'like_count': int,
 224             'repost_count': int,
 225             'comment_count': int,
 226             'tags': [],
 227         },
 228     }, {
 229         'url': 'https://bsky.app/profile/purpleicetea.bsky.social/post/3l7gv55dc2o2w',
 230         'info_dict': {
 231             'id': '3l7gv55dc2o2w',
 232         },
 233         'playlist': [{
 234             'info_dict': {
 235                 'id': '3l7gv55dc2o2w',
 236                 'ext': 'mp4',
 237                 'upload_date': '20241026',
 238                 'description': 'One of my favorite videos',
 239                 'comment_count': int,
 240                 'uploader_url': 'https://bsky.app/profile/purpleicetea.bsky.social',
 241                 'uploader': 'Purple.Ice.Tea',
 242                 'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 243                 'channel_url': 'https://bsky.app/profile/did:plc:bjh5ffwya5f53dfy47dezuwx',
 244                 'like_count': int,
 245                 'channel_id': 'did:plc:bjh5ffwya5f53dfy47dezuwx',
 246                 'repost_count': int,
 247                 'timestamp': 1729973202,
 248                 'tags': [],
 249                 'uploader_id': 'purpleicetea.bsky.social',
 250                 'title': 'One of my favorite videos',
 251             },
 252         }, {
 253             'info_dict': {
 254                 'id': '3l77u64l7le2e',
 255                 'ext': 'mp4',
 256                 'title': 'hearing people on twitter say that bluesky isn\'...',
 257                 'like_count': int,
 258                 'uploader_id': 'thafnine.net',
 259                 'uploader_url': 'https://bsky.app/profile/thafnine.net',
 260                 'upload_date': '20241024',
 261                 'channel_url': 'https://bsky.app/profile/did:plc:6ttyq36rhiyed7wu3ws7dmqj',
 262                 'description': r're:(?s)hearing people on twitter say that bluesky .{93}',
 263                 'tags': [],
 264                 'alt_title': 'md5:9b1ee1937fb3d1a81e932f9ec14d560e',
 265                 'uploader': 'T9',
 266                 'channel_id': 'did:plc:6ttyq36rhiyed7wu3ws7dmqj',
 267                 'thumbnail': r're:https://video.bsky.app/watch/.*\.jpg$',
 268                 'timestamp': 1729731642,
 269                 'comment_count': int,
 270                 'repost_count': int,
 271             },
 272         }],
 273     }]
 274     _BLOB_URL_TMPL = '{}/xrpc/com.atproto.sync.getBlob'
 275
 276     def _get_service_endpoint(self, did, video_id):
 277         if did.startswith('did:web:'):
 278             url = f'https://{did[8:]}/.well-known/did.json'
 279         else:
 280             url = f'https://plc.directory/{did}'
 281         services = self._download_json(
 282             url, video_id, 'Fetching service endpoint', 'Falling back to bsky.social', fatal=False)
 283         return traverse_obj(
 284             services, ('service', lambda _, x: x['type'] == 'AtprotoPersonalDataServer',
 285                        'serviceEndpoint', {url_or_none}, any)) or 'https://bsky.social'
 286
 287     def _real_extract(self, url):
 288         handle, video_id = self._match_valid_url(url).group('handle', 'id')
 289
 290         post = self._download_json(
 291             'https://public.api.bsky.app/xrpc/app.bsky.feed.getPostThread',
 292             video_id, query={
 293                 'uri': f'at://{handle}/app.bsky.feed.post/{video_id}',
 294                 'depth': 0,
 295                 'parentHeight': 0,
 296             })['thread']['post']
 297
 298         entries = []
 299         # app.bsky.embed.video.view/app.bsky.embed.external.view
 300         entries.extend(self._extract_videos(post, video_id))
 301         # app.bsky.embed.recordWithMedia.view
 302         entries.extend(self._extract_videos(
 303             post, video_id, embed_path=('embed', 'media'), record_subpath=('embed', 'media')))
 304         # app.bsky.embed.record.view
 305         if nested_post := traverse_obj(post, ('embed', 'record', ('record', None), {dict}, any)):
 306             entries.extend(self._extract_videos(
 307                 nested_post, video_id, embed_path=('embeds', 0), record_path='value'))
 308
 309         if not entries:
 310             raise ExtractorError('No video could be found in this post', expected=True)
 311         if len(entries) == 1:
 312             return entries[0]
 313         return self.playlist_result(entries, video_id)
 314
 315     @staticmethod
 316     def _build_profile_url(path):
 317         return format_field(path, None, 'https://bsky.app/profile/%s', default=None)
 318
 319     def _extract_videos(self, root, video_id, embed_path='embed', record_path='record', record_subpath='embed'):
 320         embed_path = variadic(embed_path, (str, bytes, dict, set))
 321         record_path = variadic(record_path, (str, bytes, dict, set))
 322         record_subpath = variadic(record_subpath, (str, bytes, dict, set))
 323
 324         entries = []
 325         if external_uri := traverse_obj(root, (
 326                 ((*record_path, *record_subpath), embed_path), 'external', 'uri', {url_or_none}, any)):
 327             entries.append(self.url_result(external_uri))
 328         if playlist := traverse_obj(root, (*embed_path, 'playlist', {url_or_none})):
 329             formats, subtitles = self._extract_m3u8_formats_and_subtitles(
 330                 playlist, video_id, 'mp4', m3u8_id='hls', fatal=False)
 331         else:
 332             return entries
 333
 334         video_cid = traverse_obj(
 335             root, (*embed_path, 'cid', {str}),
 336             (*record_path, *record_subpath, 'video', 'ref', '$link', {str}))
 337         did = traverse_obj(root, ('author', 'did', {str}))
 338
 339         if did and video_cid:
 340             endpoint = self._get_service_endpoint(did, video_id)
 341
 342             formats.append({
 343                 'format_id': 'blob',
 344                 'url': update_url_query(
 345                     self._BLOB_URL_TMPL.format(endpoint), {'did': did, 'cid': video_cid}),
 346                 **traverse_obj(root, (*embed_path, 'aspectRatio', {
 347                     'width': ('width', {int_or_none}),
 348                     'height': ('height', {int_or_none}),
 349                 })),
 350                 **traverse_obj(root, (*record_path, *record_subpath, 'video', {
 351                     'filesize': ('size', {int_or_none}),
 352                     'ext': ('mimeType', {mimetype2ext}),
 353                 })),
 354             })
 355
 356             for sub_data in traverse_obj(root, (
 357                     *record_path, *record_subpath, 'captions', lambda _, v: v['file']['ref']['$link'])):
 358                 subtitles.setdefault(sub_data.get('lang') or 'und', []).append({
 359                     'url': update_url_query(
 360                         self._BLOB_URL_TMPL.format(endpoint), {'did': did, 'cid': sub_data['file']['ref']['$link']}),
 361                     'ext': traverse_obj(sub_data, ('file', 'mimeType', {mimetype2ext})),
 362                 })
 363
 364         entries.append({
 365             'id': video_id,
 366             'formats': formats,
 367             'subtitles': subtitles,
 368             **traverse_obj(root, {
 369                 'id': ('uri', {url_basename}),
 370                 'thumbnail': (*embed_path, 'thumbnail', {url_or_none}),
 371                 'alt_title': (*embed_path, 'alt', {str}, filter),
 372                 'uploader': ('author', 'displayName', {str}),
 373                 'uploader_id': ('author', 'handle', {str}),
 374                 'uploader_url': ('author', 'handle', {self._build_profile_url}),
 375                 'channel_id': ('author', 'did', {str}),
 376                 'channel_url': ('author', 'did', {self._build_profile_url}),
 377                 'like_count': ('likeCount', {int_or_none}),
 378                 'repost_count': ('repostCount', {int_or_none}),
 379                 'comment_count': ('replyCount', {int_or_none}),
 380                 'timestamp': ('indexedAt', {parse_iso8601}),
 381                 'tags': ('labels', ..., 'val', {str}, all, {orderedSet}),
 382                 'age_limit': (
 383                     'labels', ..., 'val', {lambda x: 18 if x in ('sexual', 'porn', 'graphic-media') else None}, any),
 384                 'description': (*record_path, 'text', {str}, filter),
 385                 'title': (*record_path, 'text', {lambda x: x.replace('\n', ' ')}, {truncate_string(left=50)}),
 386             }),
 387         })
 388         return entries