yt_dlp/extractor/patreon.py

   1 import itertools
   2
   3 from .common import InfoExtractor
   4 from .vimeo import VimeoIE
   5 from ..compat import compat_urllib_parse_unquote
   6 from ..networking.exceptions import HTTPError
   7 from ..utils import (
   8     KNOWN_EXTENSIONS,
   9     ExtractorError,
  10     clean_html,
  11     determine_ext,
  12     int_or_none,
  13     mimetype2ext,
  14     parse_iso8601,
  15     str_or_none,
  16     traverse_obj,
  17     try_get,
  18     url_or_none,
  19     urljoin,
  20 )
  21
  22
  23 class PatreonBaseIE(InfoExtractor):
  24     USER_AGENT = 'Patreon/7.6.28 (Android; Android 11; Scale/2.10)'
  25
  26     def _call_api(self, ep, item_id, query=None, headers=None, fatal=True, note=None):
  27         if headers is None:
  28             headers = {}
  29         if 'User-Agent' not in headers:
  30             headers['User-Agent'] = self.USER_AGENT
  31         if query:
  32             query.update({'json-api-version': 1.0})
  33
  34         try:
  35             return self._download_json(
  36                 f'https://www.patreon.com/api/{ep}',
  37                 item_id, note='Downloading API JSON' if not note else note,
  38                 query=query, fatal=fatal, headers=headers)
  39         except ExtractorError as e:
  40             if not isinstance(e.cause, HTTPError) or mimetype2ext(e.cause.response.headers.get('Content-Type')) != 'json':
  41                 raise
  42             err_json = self._parse_json(self._webpage_read_content(e.cause.response, None, item_id), item_id, fatal=False)
  43             err_message = traverse_obj(err_json, ('errors', ..., 'detail'), get_all=False)
  44             if err_message:
  45                 raise ExtractorError(f'Patreon said: {err_message}', expected=True)
  46             raise
  47
  48
  49 class PatreonIE(PatreonBaseIE):
  50     _VALID_URL = r'https?://(?:www\.)?patreon\.com/(?:creation\?hid=|posts/(?:[\w-]+-)?)(?P<id>\d+)'
  51     _TESTS = [{
  52         'url': 'http://www.patreon.com/creation?hid=743933',
  53         'md5': 'e25505eec1053a6e6813b8ed369875cc',
  54         'info_dict': {
  55             'id': '743933',
  56             'ext': 'mp3',
  57             'title': 'Episode 166: David Smalley of Dogma Debate',
  58             'description': 'md5:34d207dd29aa90e24f1b3f58841b81c7',
  59             'uploader': 'Cognitive Dissonance Podcast',
  60             'thumbnail': 're:^https?://.*$',
  61             'timestamp': 1406473987,
  62             'upload_date': '20140727',
  63             'uploader_id': '87145',
  64             'like_count': int,
  65             'comment_count': int,
  66             'uploader_url': 'https://www.patreon.com/dissonancepod',
  67             'channel_id': '80642',
  68             'channel_url': 'https://www.patreon.com/dissonancepod',
  69             'channel_follower_count': int,
  70         },
  71     }, {
  72         'url': 'http://www.patreon.com/creation?hid=754133',
  73         'md5': '3eb09345bf44bf60451b8b0b81759d0a',
  74         'info_dict': {
  75             'id': '754133',
  76             'ext': 'mp3',
  77             'title': 'CD 167 Extra',
  78             'uploader': 'Cognitive Dissonance Podcast',
  79             'thumbnail': 're:^https?://.*$',
  80             'like_count': int,
  81             'comment_count': int,
  82             'uploader_url': 'https://www.patreon.com/dissonancepod',
  83         },
  84         'skip': 'Patron-only content',
  85     }, {
  86         'url': 'https://www.patreon.com/creation?hid=1682498',
  87         'info_dict': {
  88             'id': 'SU4fj_aEMVw',
  89             'ext': 'mp4',
  90             'title': 'I\'m on Patreon!',
  91             'uploader': 'TraciJHines',
  92             'thumbnail': 're:^https?://.*$',
  93             'upload_date': '20150211',
  94             'description': 'md5:8af6425f50bd46fbf29f3db0fc3a8364',
  95             'uploader_id': '@TraciHinesMusic',
  96             'categories': ['Entertainment'],
  97             'duration': 282,
  98             'view_count': int,
  99             'tags': 'count:39',
 100             'age_limit': 0,
 101             'channel': 'TraciJHines',
 102             'channel_url': 'https://www.youtube.com/channel/UCGLim4T2loE5rwCMdpCIPVg',
 103             'live_status': 'not_live',
 104             'like_count': int,
 105             'channel_id': 'UCGLim4T2loE5rwCMdpCIPVg',
 106             'availability': 'public',
 107             'channel_follower_count': int,
 108             'playable_in_embed': True,
 109             'uploader_url': 'https://www.youtube.com/@TraciHinesMusic',
 110             'comment_count': int,
 111             'channel_is_verified': True,
 112             'chapters': 'count:4',
 113         },
 114         'params': {
 115             'noplaylist': True,
 116             'skip_download': True,
 117         }
 118     }, {
 119         'url': 'https://www.patreon.com/posts/episode-166-of-743933',
 120         'only_matching': True,
 121     }, {
 122         'url': 'https://www.patreon.com/posts/743933',
 123         'only_matching': True,
 124     }, {
 125         'url': 'https://www.patreon.com/posts/kitchen-as-seen-51706779',
 126         'md5': '96656690071f6d64895866008484251b',
 127         'info_dict': {
 128             'id': '555089736',
 129             'ext': 'mp4',
 130             'title': 'KITCHEN AS SEEN ON DEEZ NUTS EXTENDED!',
 131             'uploader': 'Cold Ones',
 132             'thumbnail': 're:^https?://.*$',
 133             'upload_date': '20210526',
 134             'description': 'md5:557a409bd79d3898689419094934ba79',
 135             'uploader_id': '14936315',
 136         },
 137         'skip': 'Patron-only content'
 138     }, {
 139         # m3u8 video (https://github.com/yt-dlp/yt-dlp/issues/2277)
 140         'url': 'https://www.patreon.com/posts/video-sketchbook-32452882',
 141         'info_dict': {
 142             'id': '32452882',
 143             'ext': 'mp4',
 144             'comment_count': int,
 145             'uploader_id': '4301314',
 146             'like_count': int,
 147             'timestamp': 1576696962,
 148             'upload_date': '20191218',
 149             'thumbnail': r're:^https?://.*$',
 150             'uploader_url': 'https://www.patreon.com/loish',
 151             'description': 'md5:e2693e97ee299c8ece47ffdb67e7d9d2',
 152             'title': 'VIDEO // sketchbook flipthrough',
 153             'uploader': 'Loish ',
 154             'tags': ['sketchbook', 'video'],
 155             'channel_id': '1641751',
 156             'channel_url': 'https://www.patreon.com/loish',
 157             'channel_follower_count': int,
 158         }
 159     }, {
 160         # bad videos under media (if media is included). Real one is under post_file
 161         'url': 'https://www.patreon.com/posts/premium-access-70282931',
 162         'info_dict': {
 163             'id': '70282931',
 164             'ext': 'mp4',
 165             'title': '[Premium Access + Uncut] The Office - 2x6 The Fight - Group Reaction',
 166             'channel_url': 'https://www.patreon.com/thenormies',
 167             'channel_id': '573397',
 168             'uploader_id': '2929435',
 169             'uploader': 'The Normies',
 170             'description': 'md5:79c9fd8778e2cef84049a94c058a5e23',
 171             'comment_count': int,
 172             'upload_date': '20220809',
 173             'thumbnail': r're:^https?://.*$',
 174             'channel_follower_count': int,
 175             'like_count': int,
 176             'timestamp': 1660052820,
 177             'tags': ['The Office', 'early access', 'uncut'],
 178             'uploader_url': 'https://www.patreon.com/thenormies',
 179         },
 180         'skip': 'Patron-only content',
 181     }, {
 182         # dead vimeo and embed URLs, need to extract post_file
 183         'url': 'https://www.patreon.com/posts/hunter-x-hunter-34007913',
 184         'info_dict': {
 185             'id': '34007913',
 186             'ext': 'mp4',
 187             'title': 'Hunter x Hunter | Kurapika DESTROYS Uvogin!!!',
 188             'like_count': int,
 189             'uploader': 'YaBoyRoshi',
 190             'timestamp': 1581636833,
 191             'channel_url': 'https://www.patreon.com/yaboyroshi',
 192             'thumbnail': r're:^https?://.*$',
 193             'tags': ['Hunter x Hunter'],
 194             'uploader_id': '14264111',
 195             'comment_count': int,
 196             'channel_follower_count': int,
 197             'description': 'Kurapika is a walking cheat code!',
 198             'upload_date': '20200213',
 199             'channel_id': '2147162',
 200             'uploader_url': 'https://www.patreon.com/yaboyroshi',
 201         },
 202     }]
 203
 204     def _real_extract(self, url):
 205         video_id = self._match_id(url)
 206         post = self._call_api(
 207             f'posts/{video_id}', video_id, query={
 208                 'fields[media]': 'download_url,mimetype,size_bytes',
 209                 'fields[post]': 'comment_count,content,embed,image,like_count,post_file,published_at,title,current_user_can_view',
 210                 'fields[user]': 'full_name,url',
 211                 'fields[post_tag]': 'value',
 212                 'fields[campaign]': 'url,name,patron_count',
 213                 'json-api-use-default-includes': 'false',
 214                 'include': 'audio,user,user_defined_tags,campaign,attachments_media',
 215             })
 216         attributes = post['data']['attributes']
 217         title = attributes['title'].strip()
 218         image = attributes.get('image') or {}
 219         info = {
 220             'id': video_id,
 221             'title': title,
 222             'description': clean_html(attributes.get('content')),
 223             'thumbnail': image.get('large_url') or image.get('url'),
 224             'timestamp': parse_iso8601(attributes.get('published_at')),
 225             'like_count': int_or_none(attributes.get('like_count')),
 226             'comment_count': int_or_none(attributes.get('comment_count')),
 227         }
 228         can_view_post = traverse_obj(attributes, 'current_user_can_view')
 229         if can_view_post and info['comment_count']:
 230             info['__post_extractor'] = self.extract_comments(video_id)
 231
 232         for i in post.get('included', []):
 233             i_type = i.get('type')
 234             if i_type == 'media':
 235                 media_attributes = i.get('attributes') or {}
 236                 download_url = media_attributes.get('download_url')
 237                 ext = mimetype2ext(media_attributes.get('mimetype'))
 238
 239                 # if size_bytes is None, this media file is likely unavailable
 240                 # See: https://github.com/yt-dlp/yt-dlp/issues/4608
 241                 size_bytes = int_or_none(media_attributes.get('size_bytes'))
 242                 if download_url and ext in KNOWN_EXTENSIONS and size_bytes is not None:
 243                     # XXX: what happens if there are multiple attachments?
 244                     return {
 245                         **info,
 246                         'ext': ext,
 247                         'filesize': size_bytes,
 248                         'url': download_url,
 249                     }
 250             elif i_type == 'user':
 251                 user_attributes = i.get('attributes')
 252                 if user_attributes:
 253                     info.update({
 254                         'uploader': user_attributes.get('full_name'),
 255                         'uploader_id': str_or_none(i.get('id')),
 256                         'uploader_url': user_attributes.get('url'),
 257                     })
 258
 259             elif i_type == 'post_tag':
 260                 info.setdefault('tags', []).append(traverse_obj(i, ('attributes', 'value')))
 261
 262             elif i_type == 'campaign':
 263                 info.update({
 264                     'channel': traverse_obj(i, ('attributes', 'title')),
 265                     'channel_id': str_or_none(i.get('id')),
 266                     'channel_url': traverse_obj(i, ('attributes', 'url')),
 267                     'channel_follower_count': int_or_none(traverse_obj(i, ('attributes', 'patron_count'))),
 268                 })
 269
 270         # handle Vimeo embeds
 271         if try_get(attributes, lambda x: x['embed']['provider']) == 'Vimeo':
 272             embed_html = try_get(attributes, lambda x: x['embed']['html'])
 273             v_url = url_or_none(compat_urllib_parse_unquote(
 274                 self._search_regex(r'(https(?:%3A%2F%2F|://)player\.vimeo\.com.+app_id(?:=|%3D)+\d+)', embed_html, 'vimeo url', fatal=False)))
 275             if v_url:
 276                 v_url = VimeoIE._smuggle_referrer(v_url, 'https://patreon.com')
 277                 if self._request_webpage(v_url, video_id, 'Checking Vimeo embed URL', fatal=False, errnote=False):
 278                     return self.url_result(v_url, VimeoIE, url_transparent=True, **info)
 279
 280         embed_url = try_get(attributes, lambda x: x['embed']['url'])
 281         if embed_url and self._request_webpage(embed_url, video_id, 'Checking embed URL', fatal=False, errnote=False):
 282             return self.url_result(embed_url, **info)
 283
 284         post_file = traverse_obj(attributes, 'post_file')
 285         if post_file:
 286             name = post_file.get('name')
 287             ext = determine_ext(name)
 288             if ext in KNOWN_EXTENSIONS:
 289                 return {
 290                     **info,
 291                     'ext': ext,
 292                     'url': post_file['url'],
 293                 }
 294             elif name == 'video' or determine_ext(post_file.get('url')) == 'm3u8':
 295                 formats, subtitles = self._extract_m3u8_formats_and_subtitles(post_file['url'], video_id)
 296                 return {
 297                     **info,
 298                     'formats': formats,
 299                     'subtitles': subtitles,
 300                 }
 301
 302         if can_view_post is False:
 303             self.raise_no_formats('You do not have access to this post', video_id=video_id, expected=True)
 304         else:
 305             self.raise_no_formats('No supported media found in this post', video_id=video_id, expected=True)
 306         return info
 307
 308     def _get_comments(self, post_id):
 309         cursor = None
 310         count = 0
 311         params = {
 312             'page[count]': 50,
 313             'include': 'parent.commenter.campaign,parent.post.user,parent.post.campaign.creator,parent.replies.parent,parent.replies.commenter.campaign,parent.replies.post.user,parent.replies.post.campaign.creator,commenter.campaign,post.user,post.campaign.creator,replies.parent,replies.commenter.campaign,replies.post.user,replies.post.campaign.creator,on_behalf_of_campaign',
 314             'fields[comment]': 'body,created,is_by_creator',
 315             'fields[user]': 'image_url,full_name,url',
 316             'filter[flair]': 'image_tiny_url,name',
 317             'sort': '-created',
 318             'json-api-version': 1.0,
 319             'json-api-use-default-includes': 'false',
 320         }
 321
 322         for page in itertools.count(1):
 323
 324             params.update({'page[cursor]': cursor} if cursor else {})
 325             response = self._call_api(
 326                 f'posts/{post_id}/comments', post_id, query=params, note='Downloading comments page %d' % page)
 327
 328             cursor = None
 329             for comment in traverse_obj(response, (('data', ('included', lambda _, v: v['type'] == 'comment')), ...)):
 330                 count += 1
 331                 comment_id = comment.get('id')
 332                 attributes = comment.get('attributes') or {}
 333                 if comment_id is None:
 334                     continue
 335                 author_id = traverse_obj(comment, ('relationships', 'commenter', 'data', 'id'))
 336                 author_info = traverse_obj(
 337                     response, ('included', lambda _, v: v['id'] == author_id and v['type'] == 'user', 'attributes'),
 338                     get_all=False, expected_type=dict, default={})
 339
 340                 yield {
 341                     'id': comment_id,
 342                     'text': attributes.get('body'),
 343                     'timestamp': parse_iso8601(attributes.get('created')),
 344                     'parent': traverse_obj(comment, ('relationships', 'parent', 'data', 'id'), default='root'),
 345                     'author_is_uploader': attributes.get('is_by_creator'),
 346                     'author_id': author_id,
 347                     'author': author_info.get('full_name'),
 348                     'author_thumbnail': author_info.get('image_url'),
 349                 }
 350
 351             if count < traverse_obj(response, ('meta', 'count')):
 352                 cursor = traverse_obj(response, ('data', -1, 'id'))
 353
 354             if cursor is None:
 355                 break
 356
 357
 358 class PatreonCampaignIE(PatreonBaseIE):
 359
 360     _VALID_URL = r'https?://(?:www\.)?patreon\.com/(?!rss)(?:(?:m/(?P<campaign_id>\d+))|(?P<vanity>[-\w]+))'
 361     _TESTS = [{
 362         'url': 'https://www.patreon.com/dissonancepod/',
 363         'info_dict': {
 364             'title': 'Cognitive Dissonance Podcast',
 365             'channel_url': 'https://www.patreon.com/dissonancepod',
 366             'id': '80642',
 367             'description': 'md5:eb2fa8b83da7ab887adeac34da6b7af7',
 368             'channel_id': '80642',
 369             'channel': 'Cognitive Dissonance Podcast',
 370             'age_limit': 0,
 371             'channel_follower_count': int,
 372             'uploader_id': '87145',
 373             'uploader_url': 'https://www.patreon.com/dissonancepod',
 374             'uploader': 'Cognitive Dissonance Podcast',
 375             'thumbnail': r're:^https?://.*$',
 376         },
 377         'playlist_mincount': 68,
 378     }, {
 379         'url': 'https://www.patreon.com/m/4767637/posts',
 380         'info_dict': {
 381             'title': 'Not Just Bikes',
 382             'channel_follower_count': int,
 383             'id': '4767637',
 384             'channel_id': '4767637',
 385             'channel_url': 'https://www.patreon.com/notjustbikes',
 386             'description': 'md5:595c6e7dca76ae615b1d38c298a287a1',
 387             'age_limit': 0,
 388             'channel': 'Not Just Bikes',
 389             'uploader_url': 'https://www.patreon.com/notjustbikes',
 390             'uploader': 'Not Just Bikes',
 391             'uploader_id': '37306634',
 392             'thumbnail': r're:^https?://.*$',
 393         },
 394         'playlist_mincount': 71
 395     }, {
 396         'url': 'https://www.patreon.com/dissonancepod/posts',
 397         'only_matching': True
 398     }, {
 399         'url': 'https://www.patreon.com/m/5932659',
 400         'only_matching': True
 401     }]
 402
 403     @classmethod
 404     def suitable(cls, url):
 405         return False if PatreonIE.suitable(url) else super(PatreonCampaignIE, cls).suitable(url)
 406
 407     def _entries(self, campaign_id):
 408         cursor = None
 409         params = {
 410             'fields[post]': 'patreon_url,url',
 411             'filter[campaign_id]': campaign_id,
 412             'filter[is_draft]': 'false',
 413             'sort': '-published_at',
 414             'json-api-use-default-includes': 'false',
 415         }
 416
 417         for page in itertools.count(1):
 418
 419             params.update({'page[cursor]': cursor} if cursor else {})
 420             posts_json = self._call_api('posts', campaign_id, query=params, note='Downloading posts page %d' % page)
 421
 422             cursor = traverse_obj(posts_json, ('meta', 'pagination', 'cursors', 'next'))
 423             for post_url in traverse_obj(posts_json, ('data', ..., 'attributes', 'patreon_url')):
 424                 yield self.url_result(urljoin('https://www.patreon.com/', post_url), PatreonIE)
 425
 426             if cursor is None:
 427                 break
 428
 429     def _real_extract(self, url):
 430
 431         campaign_id, vanity = self._match_valid_url(url).group('campaign_id', 'vanity')
 432         if campaign_id is None:
 433             webpage = self._download_webpage(url, vanity, headers={'User-Agent': self.USER_AGENT})
 434             campaign_id = self._search_regex(r'https://www.patreon.com/api/campaigns/(\d+)/?', webpage, 'Campaign ID')
 435
 436         params = {
 437             'json-api-use-default-includes': 'false',
 438             'fields[user]': 'full_name,url',
 439             'fields[campaign]': 'name,summary,url,patron_count,creation_count,is_nsfw,avatar_photo_url',
 440             'include': 'creator'
 441         }
 442
 443         campaign_response = self._call_api(
 444             f'campaigns/{campaign_id}', campaign_id,
 445             note='Downloading campaign info', fatal=False,
 446             query=params) or {}
 447
 448         campaign_info = campaign_response.get('data') or {}
 449         channel_name = traverse_obj(campaign_info, ('attributes', 'name'))
 450         user_info = traverse_obj(
 451             campaign_response, ('included', lambda _, v: v['type'] == 'user'),
 452             default={}, expected_type=dict, get_all=False)
 453
 454         return {
 455             '_type': 'playlist',
 456             'id': campaign_id,
 457             'title': channel_name,
 458             'entries': self._entries(campaign_id),
 459             'description': clean_html(traverse_obj(campaign_info, ('attributes', 'summary'))),
 460             'channel_url': traverse_obj(campaign_info, ('attributes', 'url')),
 461             'channel_follower_count': int_or_none(traverse_obj(campaign_info, ('attributes', 'patron_count'))),
 462             'channel_id': campaign_id,
 463             'channel': channel_name,
 464             'uploader_url': traverse_obj(user_info, ('attributes', 'url')),
 465             'uploader_id': str_or_none(user_info.get('id')),
 466             'uploader': traverse_obj(user_info, ('attributes', 'full_name')),
 467             'playlist_count': traverse_obj(campaign_info, ('attributes', 'creation_count')),
 468             'age_limit': 18 if traverse_obj(campaign_info, ('attributes', 'is_nsfw')) else 0,
 469             'thumbnail': url_or_none(traverse_obj(campaign_info, ('attributes', 'avatar_photo_url'))),
 470         }