yt_dlp/extractor/loom.py

   1 import json
   2 import textwrap
   3 import urllib.parse
   4 import uuid
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     ExtractorError,
   9     determine_ext,
  10     filter_dict,
  11     get_first,
  12     int_or_none,
  13     parse_iso8601,
  14     update_url,
  15     url_or_none,
  16     variadic,
  17 )
  18 from ..utils.traversal import traverse_obj
  19
  20
  21 class LoomIE(InfoExtractor):
  22     IE_NAME = 'loom'
  23     _VALID_URL = r'https?://(?:www\.)?loom\.com/(?:share|embed)/(?P<id>[\da-f]{32})'
  24     _EMBED_REGEX = [rf'<iframe[^>]+\bsrc=["\'](?P<url>{_VALID_URL})']
  25     _TESTS = [{
  26         # m3u8 raw-url, mp4 transcoded-url, cdn url == raw-url, json subs only
  27         'url': 'https://www.loom.com/share/43d05f362f734614a2e81b4694a3a523',
  28         'md5': 'bfc2d7e9c2e0eb4813212230794b6f42',
  29         'info_dict': {
  30             'id': '43d05f362f734614a2e81b4694a3a523',
  31             'ext': 'mp4',
  32             'title': 'A Ruler for Windows - 28 March 2022',
  33             'uploader': 'wILLIAM PIP',
  34             'upload_date': '20220328',
  35             'timestamp': 1648454238,
  36             'duration': 27,
  37         },
  38     }, {
  39         # webm raw-url, mp4 transcoded-url, cdn url == transcoded-url, no subs
  40         'url': 'https://www.loom.com/share/c43a642f815f4378b6f80a889bb73d8d',
  41         'md5': '70f529317be8cf880fcc2c649a531900',
  42         'info_dict': {
  43             'id': 'c43a642f815f4378b6f80a889bb73d8d',
  44             'ext': 'webm',
  45             'title': 'Lilah Nielsen Intro Video',
  46             'uploader': 'Lilah Nielsen',
  47             'upload_date': '20200826',
  48             'timestamp': 1598480716,
  49             'duration': 20,
  50         },
  51     }, {
  52         # m3u8 raw-url, mp4 transcoded-url, cdn url == raw-url, vtt sub and json subs
  53         'url': 'https://www.loom.com/share/9458bcbf79784162aa62ffb8dd66201b',
  54         'md5': '51737ec002969dd28344db4d60b9cbbb',
  55         'info_dict': {
  56             'id': '9458bcbf79784162aa62ffb8dd66201b',
  57             'ext': 'mp4',
  58             'title': 'Sharing screen with gpt-4',
  59             'description': 'Sharing screen with GPT 4 vision model and asking questions to guide through blender.',
  60             'uploader': 'Suneel Matham',
  61             'chapters': 'count:3',
  62             'upload_date': '20231109',
  63             'timestamp': 1699518978,
  64             'duration': 93,
  65         },
  66     }, {
  67         # mpd raw-url, mp4 transcoded-url, cdn url == raw-url, no subs
  68         'url': 'https://www.loom.com/share/24351eb8b317420289b158e4b7e96ff2',
  69         'info_dict': {
  70             'id': '24351eb8b317420289b158e4b7e96ff2',
  71             'ext': 'webm',
  72             'title': 'OMFG clown',
  73             'description': 'md5:285c5ee9d62aa087b7e3271b08796815',
  74             'uploader': 'MrPumkin B',
  75             'upload_date': '20210924',
  76             'timestamp': 1632519618,
  77             'duration': 210,
  78         },
  79         'params': {'skip_download': 'dash'},
  80     }, {
  81         # password-protected
  82         'url': 'https://www.loom.com/share/50e26e8aeb7940189dff5630f95ce1f4',
  83         'md5': '5cc7655e7d55d281d203f8ffd14771f7',
  84         'info_dict': {
  85             'id': '50e26e8aeb7940189dff5630f95ce1f4',
  86             'ext': 'mp4',
  87             'title': 'iOS Mobile Upload',
  88             'uploader': 'Simon Curran',
  89             'upload_date': '20200520',
  90             'timestamp': 1590000123,
  91             'duration': 35,
  92         },
  93         'params': {'videopassword': 'seniorinfants2'},
  94     }, {
  95         # embed, transcoded-url endpoint sends empty JSON response, split video and audio HLS formats
  96         'url': 'https://www.loom.com/embed/ddcf1c1ad21f451ea7468b1e33917e4e',
  97         'md5': 'b321d261656848c184a94e3b93eae28d',
  98         'info_dict': {
  99             'id': 'ddcf1c1ad21f451ea7468b1e33917e4e',
 100             'ext': 'mp4',
 101             'title': 'CF Reset User\'s Password',
 102             'uploader': 'Aimee Heintz',
 103             'upload_date': '20220707',
 104             'timestamp': 1657216459,
 105             'duration': 181,
 106         },
 107         'params': {'format': 'bestvideo'},  # Test video-only fixup
 108         'expected_warnings': ['Failed to parse JSON'],
 109     }]
 110     _WEBPAGE_TESTS = [{
 111         'url': 'https://www.loom.com/community/e1229802a8694a09909e8ba0fbb6d073-pg',
 112         'md5': 'ec838cd01b576cf0386f32e1ae424609',
 113         'info_dict': {
 114             'id': 'e1229802a8694a09909e8ba0fbb6d073',
 115             'ext': 'mp4',
 116             'title': 'Rexie Jane Cimafranca - Founder\'s Presentation',
 117             'uploader': 'Rexie Cimafranca',
 118             'upload_date': '20230213',
 119             'duration': 247,
 120             'timestamp': 1676274030,
 121         },
 122     }]
 123
 124     _GRAPHQL_VARIABLES = {
 125         'GetVideoSource': {
 126             'acceptableMimes': ['DASH', 'M3U8', 'MP4'],
 127         },
 128     }
 129     _GRAPHQL_QUERIES = {
 130         'GetVideoSSR': textwrap.dedent('''\
 131             query GetVideoSSR($videoId: ID!, $password: String) {
 132               getVideo(id: $videoId, password: $password) {
 133                 __typename
 134                 ... on PrivateVideo {
 135                   id
 136                   status
 137                   message
 138                   __typename
 139                 }
 140                 ... on VideoPasswordMissingOrIncorrect {
 141                   id
 142                   message
 143                   __typename
 144                 }
 145                 ... on RegularUserVideo {
 146                   id
 147                   __typename
 148                   createdAt
 149                   description
 150                   download_enabled
 151                   folder_id
 152                   is_protected
 153                   needs_password
 154                   owner {
 155                     display_name
 156                     __typename
 157                   }
 158                   privacy
 159                   s3_id
 160                   name
 161                   video_properties {
 162                     avgBitRate
 163                     client
 164                     camera_enabled
 165                     client_version
 166                     duration
 167                     durationMs
 168                     format
 169                     height
 170                     microphone_enabled
 171                     os
 172                     os_version
 173                     recordingClient
 174                     recording_type
 175                     recording_version
 176                     screen_type
 177                     tab_audio
 178                     trim_duration
 179                     width
 180                     __typename
 181                   }
 182                   playable_duration
 183                   source_duration
 184                   visibility
 185                 }
 186               }
 187             }\n'''),
 188         'GetVideoSource': textwrap.dedent('''\
 189             query GetVideoSource($videoId: ID!, $password: String, $acceptableMimes: [CloudfrontVideoAcceptableMime]) {
 190               getVideo(id: $videoId, password: $password) {
 191                 ... on RegularUserVideo {
 192                   id
 193                   nullableRawCdnUrl(acceptableMimes: $acceptableMimes, password: $password) {
 194                     url
 195                     __typename
 196                   }
 197                   __typename
 198                 }
 199                 __typename
 200               }
 201             }\n'''),
 202         'FetchVideoTranscript': textwrap.dedent('''\
 203             query FetchVideoTranscript($videoId: ID!, $password: String) {
 204               fetchVideoTranscript(videoId: $videoId, password: $password) {
 205                 ... on VideoTranscriptDetails {
 206                   id
 207                   video_id
 208                   source_url
 209                   captions_source_url
 210                   __typename
 211                 }
 212                 ... on GenericError {
 213                   message
 214                   __typename
 215                 }
 216                 __typename
 217               }
 218             }\n'''),
 219         'FetchChapters': textwrap.dedent('''\
 220             query FetchChapters($videoId: ID!, $password: String) {
 221               fetchVideoChapters(videoId: $videoId, password: $password) {
 222                 ... on VideoChapters {
 223                   video_id
 224                   content
 225                   __typename
 226                 }
 227                 ... on EmptyChaptersPayload {
 228                   content
 229                   __typename
 230                 }
 231                 ... on InvalidRequestWarning {
 232                   message
 233                   __typename
 234                 }
 235                 ... on Error {
 236                   message
 237                   __typename
 238                 }
 239                 __typename
 240               }
 241             }\n'''),
 242     }
 243     _APOLLO_GRAPHQL_VERSION = '0a1856c'
 244
 245     def _call_graphql_api(self, operations, video_id, note=None, errnote=None):
 246         password = self.get_param('videopassword')
 247         return self._download_json(
 248             'https://www.loom.com/graphql', video_id, note or 'Downloading GraphQL JSON',
 249             errnote or 'Failed to download GraphQL JSON', headers={
 250                 'Accept': 'application/json',
 251                 'Content-Type': 'application/json',
 252                 'x-loom-request-source': f'loom_web_{self._APOLLO_GRAPHQL_VERSION}',
 253                 'apollographql-client-name': 'web',
 254                 'apollographql-client-version': self._APOLLO_GRAPHQL_VERSION,
 255             }, data=json.dumps([{
 256                 'operationName': operation_name,
 257                 'variables': {
 258                     'videoId': video_id,
 259                     'password': password,
 260                     **self._GRAPHQL_VARIABLES.get(operation_name, {}),
 261                 },
 262                 'query': self._GRAPHQL_QUERIES[operation_name],
 263             } for operation_name in variadic(operations)], separators=(',', ':')).encode())
 264
 265     def _call_url_api(self, endpoint, video_id):
 266         response = self._download_json(
 267             f'https://www.loom.com/api/campaigns/sessions/{video_id}/{endpoint}', video_id,
 268             f'Downloading {endpoint} JSON', f'Failed to download {endpoint} JSON', fatal=False,
 269             headers={'Accept': 'application/json', 'Content-Type': 'application/json'},
 270             data=json.dumps({
 271                 'anonID': str(uuid.uuid4()),
 272                 'deviceID': None,
 273                 'force_original': False,  # HTTP error 401 if True
 274                 'password': self.get_param('videopassword'),
 275             }, separators=(',', ':')).encode())
 276         return traverse_obj(response, ('url', {url_or_none}))
 277
 278     def _extract_formats(self, video_id, metadata, gql_data):
 279         formats = []
 280         video_properties = traverse_obj(metadata, ('video_properties', {
 281             'width': ('width', {int_or_none}),
 282             'height': ('height', {int_or_none}),
 283             'acodec': ('microphone_enabled', {lambda x: 'none' if x is False else None}),
 284         }))
 285
 286         def get_formats(format_url, format_id, quality):
 287             if not format_url:
 288                 return
 289             ext = determine_ext(format_url)
 290             query = urllib.parse.urlparse(format_url).query
 291
 292             if ext == 'm3u8':
 293                 # Extract pre-merged HLS formats to avoid buggy parsing of metadata in split playlists
 294                 format_url = format_url.replace('-split.m3u8', '.m3u8')
 295                 m3u8_formats = self._extract_m3u8_formats(
 296                     format_url, video_id, 'mp4', m3u8_id=f'hls-{format_id}', fatal=False, quality=quality)
 297                 # Sometimes only split video/audio formats are available, need to fixup video-only formats
 298                 is_not_premerged = 'none' in traverse_obj(m3u8_formats, (..., 'vcodec'))
 299                 for fmt in m3u8_formats:
 300                     if is_not_premerged and fmt.get('vcodec') != 'none':
 301                         fmt['acodec'] = 'none'
 302                     yield {
 303                         **fmt,
 304                         'url': update_url(fmt['url'], query=query),
 305                         'extra_param_to_segment_url': query,
 306                     }
 307
 308             elif ext == 'mpd':
 309                 dash_formats = self._extract_mpd_formats(
 310                     format_url, video_id, mpd_id=f'dash-{format_id}', fatal=False)
 311                 for fmt in dash_formats:
 312                     yield {
 313                         **fmt,
 314                         'extra_param_to_segment_url': query,
 315                         'quality': quality,
 316                     }
 317
 318             else:
 319                 yield {
 320                     'url': format_url,
 321                     'ext': ext,
 322                     'format_id': f'http-{format_id}',
 323                     'quality': quality,
 324                     **video_properties,
 325                 }
 326
 327         raw_url = self._call_url_api('raw-url', video_id)
 328         formats.extend(get_formats(raw_url, 'raw', quality=1))  # original quality
 329
 330         transcoded_url = self._call_url_api('transcoded-url', video_id)
 331         formats.extend(get_formats(transcoded_url, 'transcoded', quality=-1))  # transcoded quality
 332
 333         cdn_url = get_first(gql_data, ('data', 'getVideo', 'nullableRawCdnUrl', 'url', {url_or_none}))
 334         # cdn_url is usually a dupe, but the raw-url/transcoded-url endpoints could return errors
 335         valid_urls = [update_url(url, query=None) for url in (raw_url, transcoded_url) if url]
 336         if cdn_url and update_url(cdn_url, query=None) not in valid_urls:
 337             formats.extend(get_formats(cdn_url, 'cdn', quality=0))  # could be original or transcoded
 338
 339         return formats
 340
 341     def _real_extract(self, url):
 342         video_id = self._match_id(url)
 343         metadata = get_first(
 344             self._call_graphql_api('GetVideoSSR', video_id, 'Downloading GraphQL metadata JSON'),
 345             ('data', 'getVideo', {dict})) or {}
 346
 347         if metadata.get('__typename') == 'VideoPasswordMissingOrIncorrect':
 348             if not self.get_param('videopassword'):
 349                 raise ExtractorError(
 350                     'This video is password-protected, use the --video-password option', expected=True)
 351             raise ExtractorError('Invalid video password', expected=True)
 352
 353         gql_data = self._call_graphql_api(['FetchChapters', 'FetchVideoTranscript', 'GetVideoSource'], video_id)
 354         duration = traverse_obj(metadata, ('video_properties', 'duration', {int_or_none}))
 355
 356         return {
 357             'id': video_id,
 358             'duration': duration,
 359             'chapters': self._extract_chapters_from_description(
 360                 get_first(gql_data, ('data', 'fetchVideoChapters', 'content', {str})), duration) or None,
 361             'formats': self._extract_formats(video_id, metadata, gql_data),
 362             'subtitles': filter_dict({
 363                 'en': traverse_obj(gql_data, (
 364                     ..., 'data', 'fetchVideoTranscript',
 365                     ('source_url', 'captions_source_url'), {
 366                         'url': {url_or_none},
 367                     })) or None,
 368             }),
 369             **traverse_obj(metadata, {
 370                 'title': ('name', {str}),
 371                 'description': ('description', {str}),
 372                 'uploader': ('owner', 'display_name', {str}),
 373                 'timestamp': ('createdAt', {parse_iso8601}),
 374             }),
 375         }
 376
 377
 378 class LoomFolderIE(InfoExtractor):
 379     IE_NAME = 'loom:folder'
 380     _VALID_URL = r'https?://(?:www\.)?loom\.com/share/folder/(?P<id>[\da-f]{32})'
 381     _TESTS = [{
 382         # 2 subfolders, no videos in root
 383         'url': 'https://www.loom.com/share/folder/997db4db046f43e5912f10dc5f817b5c',
 384         'playlist_mincount': 16,
 385         'info_dict': {
 386             'id': '997db4db046f43e5912f10dc5f817b5c',
 387             'title': 'Blending Lessons',
 388         },
 389     }, {
 390         # only videos, no subfolders
 391         'url': 'https://www.loom.com/share/folder/9a8a87f6b6f546d9a400c8e7575ff7f2',
 392         'playlist_mincount': 12,
 393         'info_dict': {
 394             'id': '9a8a87f6b6f546d9a400c8e7575ff7f2',
 395             'title': 'List A- a, i, o',
 396         },
 397     }, {
 398         # videos in root and empty subfolder
 399         'url': 'https://www.loom.com/share/folder/886e534218c24fd292e97e9563078cc4',
 400         'playlist_mincount': 21,
 401         'info_dict': {
 402             'id': '886e534218c24fd292e97e9563078cc4',
 403             'title': 'Medicare Agent Training videos',
 404         },
 405     }, {
 406         # videos in root and videos in subfolders
 407         'url': 'https://www.loom.com/share/folder/b72c4ecdf04745da9403926d80a40c38',
 408         'playlist_mincount': 21,
 409         'info_dict': {
 410             'id': 'b72c4ecdf04745da9403926d80a40c38',
 411             'title': 'Quick Altos Q & A Tutorials',
 412         },
 413     }, {
 414         # recursive folder extraction
 415         'url': 'https://www.loom.com/share/folder/8b458a94e0e4449b8df9ea7a68fafc4e',
 416         'playlist_count': 23,
 417         'info_dict': {
 418             'id': '8b458a94e0e4449b8df9ea7a68fafc4e',
 419             'title': 'Sezer Texting Guide',
 420         },
 421     }, {
 422         # more than 50 videos in 1 folder
 423         'url': 'https://www.loom.com/share/folder/e056a91d290d47ca9b00c9d1df56c463',
 424         'playlist_mincount': 61,
 425         'info_dict': {
 426             'id': 'e056a91d290d47ca9b00c9d1df56c463',
 427             'title': 'User Videos',
 428         },
 429     }, {
 430         # many subfolders
 431         'url': 'https://www.loom.com/share/folder/c2dde8cc67454f0e99031677279d8954',
 432         'playlist_mincount': 75,
 433         'info_dict': {
 434             'id': 'c2dde8cc67454f0e99031677279d8954',
 435             'title': 'Honors 1',
 436         },
 437     }, {
 438         'url': 'https://www.loom.com/share/folder/bae17109a68146c7803454f2893c8cf8/Edpuzzle',
 439         'only_matching': True,
 440     }]
 441
 442     def _extract_folder_data(self, folder_id):
 443         return self._download_json(
 444             f'https://www.loom.com/v1/folders/{folder_id}', folder_id,
 445             'Downloading folder info JSON', query={'limit': '10000'})
 446
 447     def _extract_folder_entries(self, folder_id, initial_folder_data=None):
 448         folder_data = initial_folder_data or self._extract_folder_data(folder_id)
 449
 450         for video in traverse_obj(folder_data, ('videos', lambda _, v: v['id'])):
 451             video_id = video['id']
 452             yield self.url_result(
 453                 f'https://www.loom.com/share/{video_id}', LoomIE, video_id, video.get('name'))
 454
 455         # Recurse into subfolders
 456         for subfolder_id in traverse_obj(folder_data, (
 457                 'folders', lambda _, v: v['id'] != folder_id, 'id', {str})):
 458             yield from self._extract_folder_entries(subfolder_id)
 459
 460     def _real_extract(self, url):
 461         playlist_id = self._match_id(url)
 462         playlist_data = self._extract_folder_data(playlist_id)
 463
 464         return self.playlist_result(
 465             self._extract_folder_entries(playlist_id, playlist_data), playlist_id,
 466             traverse_obj(playlist_data, ('folder', 'name', {str.strip})))