yt_dlp/extractor/instagram.py

   1 import hashlib
   2 import itertools
   3 import json
   4 import re
   5 import time
   6
   7 from .common import InfoExtractor
   8 from ..networking.exceptions import HTTPError
   9 from ..utils import (
  10     ExtractorError,
  11     decode_base_n,
  12     encode_base_n,
  13     filter_dict,
  14     float_or_none,
  15     format_field,
  16     get_element_by_attribute,
  17     int_or_none,
  18     lowercase_escape,
  19     str_or_none,
  20     str_to_int,
  21     traverse_obj,
  22     url_or_none,
  23     urlencode_postdata,
  24 )
  25
  26 _ENCODING_CHARS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_'
  27
  28
  29 def _pk_to_id(media_id):
  30     """Source: https://stackoverflow.com/questions/24437823/getting-instagram-post-url-from-media-id"""
  31     return encode_base_n(int(media_id.split('_')[0]), table=_ENCODING_CHARS)
  32
  33
  34 def _id_to_pk(shortcode):
  35     """Covert a shortcode to a numeric value"""
  36     return decode_base_n(shortcode[:11], table=_ENCODING_CHARS)
  37
  38
  39 class InstagramBaseIE(InfoExtractor):
  40     _NETRC_MACHINE = 'instagram'
  41     _IS_LOGGED_IN = False
  42
  43     _API_BASE_URL = 'https://i.instagram.com/api/v1'
  44     _LOGIN_URL = 'https://www.instagram.com/accounts/login'
  45     _API_HEADERS = {
  46         'X-IG-App-ID': '936619743392459',
  47         'X-ASBD-ID': '198387',
  48         'X-IG-WWW-Claim': '0',
  49         'Origin': 'https://www.instagram.com',
  50         'Accept': '*/*',
  51         'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/104.0.0.0 Safari/537.36',
  52     }
  53
  54     def _perform_login(self, username, password):
  55         if self._IS_LOGGED_IN:
  56             return
  57
  58         login_webpage = self._download_webpage(
  59             self._LOGIN_URL, None, note='Downloading login webpage', errnote='Failed to download login webpage')
  60
  61         shared_data = self._parse_json(self._search_regex(
  62             r'window\._sharedData\s*=\s*({.+?});', login_webpage, 'shared data', default='{}'), None)
  63
  64         login = self._download_json(
  65             f'{self._LOGIN_URL}/ajax/', None, note='Logging in', headers={
  66                 **self._API_HEADERS,
  67                 'X-Requested-With': 'XMLHttpRequest',
  68                 'X-CSRFToken': shared_data['config']['csrf_token'],
  69                 'X-Instagram-AJAX': shared_data['rollout_hash'],
  70                 'Referer': 'https://www.instagram.com/',
  71             }, data=urlencode_postdata({
  72                 'enc_password': f'#PWD_INSTAGRAM_BROWSER:0:{int(time.time())}:{password}',
  73                 'username': username,
  74                 'queryParams': '{}',
  75                 'optIntoOneTap': 'false',
  76                 'stopDeletionNonce': '',
  77                 'trustedDeviceRecords': '{}',
  78             }))
  79
  80         if not login.get('authenticated'):
  81             if login.get('message'):
  82                 raise ExtractorError(f'Unable to login: {login["message"]}')
  83             elif login.get('user'):
  84                 raise ExtractorError('Unable to login: Sorry, your password was incorrect. Please double-check your password.', expected=True)
  85             elif login.get('user') is False:
  86                 raise ExtractorError('Unable to login: The username you entered doesn\'t belong to an account. Please check your username and try again.', expected=True)
  87             raise ExtractorError('Unable to login')
  88         InstagramBaseIE._IS_LOGGED_IN = True
  89
  90     def _get_count(self, media, kind, *keys):
  91         return traverse_obj(
  92             media, (kind, 'count'), *((f'edge_media_{key}', 'count') for key in keys),
  93             expected_type=int_or_none)
  94
  95     def _get_dimension(self, name, media, webpage=None):
  96         return (
  97             traverse_obj(media, ('dimensions', name), expected_type=int_or_none)
  98             or int_or_none(self._html_search_meta(
  99                 (f'og:video:{name}', f'video:{name}'), webpage or '', default=None)))
 100
 101     def _extract_nodes(self, nodes, is_direct=False):
 102         for idx, node in enumerate(nodes, start=1):
 103             if node.get('__typename') != 'GraphVideo' and node.get('is_video') is not True:
 104                 continue
 105
 106             video_id = node.get('shortcode')
 107
 108             if is_direct:
 109                 info = {
 110                     'id': video_id or node['id'],
 111                     'url': node.get('video_url'),
 112                     'width': self._get_dimension('width', node),
 113                     'height': self._get_dimension('height', node),
 114                     'http_headers': {
 115                         'Referer': 'https://www.instagram.com/',
 116                     },
 117                 }
 118             elif not video_id:
 119                 continue
 120             else:
 121                 info = {
 122                     '_type': 'url',
 123                     'ie_key': 'Instagram',
 124                     'id': video_id,
 125                     'url': f'https://instagram.com/p/{video_id}',
 126                 }
 127
 128             yield {
 129                 **info,
 130                 'title': node.get('title') or (f'Video {idx}' if is_direct else None),
 131                 'description': traverse_obj(
 132                     node, ('edge_media_to_caption', 'edges', 0, 'node', 'text'), expected_type=str),
 133                 'thumbnail': traverse_obj(
 134                     node, 'display_url', 'thumbnail_src', 'display_src', expected_type=url_or_none),
 135                 'duration': float_or_none(node.get('video_duration')),
 136                 'timestamp': int_or_none(node.get('taken_at_timestamp')),
 137                 'view_count': int_or_none(node.get('video_view_count')),
 138                 'comment_count': self._get_count(node, 'comments', 'preview_comment', 'to_comment', 'to_parent_comment'),
 139                 'like_count': self._get_count(node, 'likes', 'preview_like'),
 140             }
 141
 142     def _extract_product_media(self, product_media):
 143         media_id = product_media.get('code') or _pk_to_id(product_media.get('pk'))
 144         vcodec = product_media.get('video_codec')
 145         dash_manifest_raw = product_media.get('video_dash_manifest')
 146         videos_list = product_media.get('video_versions')
 147         if not (dash_manifest_raw or videos_list):
 148             return {}
 149
 150         formats = [{
 151             'format_id': fmt.get('id'),
 152             'url': fmt.get('url'),
 153             'width': fmt.get('width'),
 154             'height': fmt.get('height'),
 155             'vcodec': vcodec,
 156         } for fmt in videos_list or []]
 157         if dash_manifest_raw:
 158             formats.extend(self._parse_mpd_formats(self._parse_xml(dash_manifest_raw, media_id), mpd_id='dash'))
 159
 160         thumbnails = [{
 161             'url': thumbnail.get('url'),
 162             'width': thumbnail.get('width'),
 163             'height': thumbnail.get('height'),
 164         } for thumbnail in traverse_obj(product_media, ('image_versions2', 'candidates')) or []]
 165         return {
 166             'id': media_id,
 167             'duration': float_or_none(product_media.get('video_duration')),
 168             'formats': formats,
 169             'thumbnails': thumbnails,
 170         }
 171
 172     def _extract_product(self, product_info):
 173         if isinstance(product_info, list):
 174             product_info = product_info[0]
 175
 176         user_info = product_info.get('user') or {}
 177         info_dict = {
 178             'id': _pk_to_id(traverse_obj(product_info, 'pk', 'id', expected_type=str_or_none)[:19]),
 179             'title': product_info.get('title') or f'Video by {user_info.get("username")}',
 180             'description': traverse_obj(product_info, ('caption', 'text'), expected_type=str_or_none),
 181             'timestamp': int_or_none(product_info.get('taken_at')),
 182             'channel': user_info.get('username'),
 183             'uploader': user_info.get('full_name'),
 184             'uploader_id': str_or_none(user_info.get('pk')),
 185             'view_count': int_or_none(product_info.get('view_count')),
 186             'like_count': int_or_none(product_info.get('like_count')),
 187             'comment_count': int_or_none(product_info.get('comment_count')),
 188             '__post_extractor': self.extract_comments(_pk_to_id(product_info.get('pk'))),
 189             'http_headers': {
 190                 'Referer': 'https://www.instagram.com/',
 191             },
 192         }
 193         carousel_media = product_info.get('carousel_media')
 194         if carousel_media:
 195             return {
 196                 '_type': 'playlist',
 197                 **info_dict,
 198                 'title': f'Post by {user_info.get("username")}',
 199                 'entries': [{
 200                     **info_dict,
 201                     **self._extract_product_media(product_media),
 202                 } for product_media in carousel_media],
 203             }
 204
 205         return {
 206             **info_dict,
 207             **self._extract_product_media(product_info),
 208         }
 209
 210     def _get_comments(self, video_id):
 211         comments_info = self._download_json(
 212             f'{self._API_BASE_URL}/media/{_id_to_pk(video_id)}/comments/?can_support_threading=true&permalink_enabled=false', video_id,
 213             fatal=False, errnote='Comments extraction failed', note='Downloading comments info', headers=self._API_HEADERS) or {}
 214
 215         comment_data = traverse_obj(comments_info, ('edge_media_to_parent_comment', 'edges'), 'comments')
 216         for comment_dict in comment_data or []:
 217             yield {
 218                 'author': traverse_obj(comment_dict, ('node', 'owner', 'username'), ('user', 'username')),
 219                 'author_id': traverse_obj(comment_dict, ('node', 'owner', 'id'), ('user', 'pk')),
 220                 'author_thumbnail': traverse_obj(comment_dict, ('node', 'owner', 'profile_pic_url'), ('user', 'profile_pic_url'), expected_type=url_or_none),
 221                 'id': traverse_obj(comment_dict, ('node', 'id'), 'pk'),
 222                 'text': traverse_obj(comment_dict, ('node', 'text'), 'text'),
 223                 'like_count': traverse_obj(comment_dict, ('node', 'edge_liked_by', 'count'), 'comment_like_count', expected_type=int_or_none),
 224                 'timestamp': traverse_obj(comment_dict, ('node', 'created_at'), 'created_at', expected_type=int_or_none),
 225             }
 226
 227
 228 class InstagramIOSIE(InfoExtractor):
 229     IE_DESC = 'IOS instagram:// URL'
 230     _VALID_URL = r'instagram://media\?id=(?P<id>[\d_]+)'
 231     _TESTS = [{
 232         'url': 'instagram://media?id=482584233761418119',
 233         'md5': '0d2da106a9d2631273e192b372806516',
 234         'info_dict': {
 235             'id': 'aye83DjauH',
 236             'ext': 'mp4',
 237             'title': 'Video by naomipq',
 238             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
 239             'thumbnail': r're:^https?://.*\.jpg',
 240             'duration': 0,
 241             'timestamp': 1371748545,
 242             'upload_date': '20130620',
 243             'uploader_id': 'naomipq',
 244             'uploader': 'B E A U T Y  F O R  A S H E S',
 245             'like_count': int,
 246             'comment_count': int,
 247             'comments': list,
 248         },
 249         'add_ie': ['Instagram'],
 250     }]
 251
 252     def _real_extract(self, url):
 253         video_id = _pk_to_id(self._match_id(url))
 254         return self.url_result(f'http://instagram.com/tv/{video_id}', InstagramIE, video_id)
 255
 256
 257 class InstagramIE(InstagramBaseIE):
 258     _VALID_URL = r'(?P<url>https?://(?:www\.)?instagram\.com(?:/[^/]+)?/(?:p|tv|reels?(?!/audio/))/(?P<id>[^/?#&]+))'
 259     _EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?instagram\.com/p/[^/]+/embed.*?)\1']
 260     _TESTS = [{
 261         'url': 'https://instagram.com/p/aye83DjauH/?foo=bar#abc',
 262         'md5': '0d2da106a9d2631273e192b372806516',
 263         'info_dict': {
 264             'id': 'aye83DjauH',
 265             'ext': 'mp4',
 266             'title': 'Video by naomipq',
 267             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
 268             'thumbnail': r're:^https?://.*\.jpg',
 269             'duration': 8.747,
 270             'timestamp': 1371748545,
 271             'upload_date': '20130620',
 272             'uploader_id': '2815873',
 273             'uploader': 'B E A U T Y  F O R  A S H E S',
 274             'channel': 'naomipq',
 275             'like_count': int,
 276             'comment_count': int,
 277             'comments': list,
 278         },
 279         'expected_warnings': [
 280             'General metadata extraction failed',
 281             'Main webpage is locked behind the login page',
 282         ],
 283     }, {
 284         # reel
 285         'url': 'https://www.instagram.com/reel/Chunk8-jurw/',
 286         'md5': 'f6d8277f74515fa3ff9f5791426e42b1',
 287         'info_dict': {
 288             'id': 'Chunk8-jurw',
 289             'ext': 'mp4',
 290             'title': 'Video by instagram',
 291             'description': 'md5:c9cde483606ed6f80fbe9283a6a2b290',
 292             'thumbnail': r're:^https?://.*\.jpg',
 293             'duration': 5.016,
 294             'timestamp': 1661529231,
 295             'upload_date': '20220826',
 296             'uploader_id': '25025320',
 297             'uploader': 'Instagram',
 298             'channel': 'instagram',
 299             'like_count': int,
 300             'comment_count': int,
 301             'comments': list,
 302         },
 303         'expected_warnings': [
 304             'General metadata extraction failed',
 305             'Main webpage is locked behind the login page',
 306         ],
 307     }, {
 308         # multi video post
 309         'url': 'https://www.instagram.com/p/BQ0eAlwhDrw/',
 310         'playlist': [{
 311             'info_dict': {
 312                 'id': 'BQ0dSaohpPW',
 313                 'ext': 'mp4',
 314                 'title': 'Video 1',
 315                 'thumbnail': r're:^https?://.*\.jpg',
 316                 'view_count': int,
 317             },
 318         }, {
 319             'info_dict': {
 320                 'id': 'BQ0dTpOhuHT',
 321                 'ext': 'mp4',
 322                 'title': 'Video 2',
 323                 'thumbnail': r're:^https?://.*\.jpg',
 324                 'view_count': int,
 325             },
 326         }, {
 327             'info_dict': {
 328                 'id': 'BQ0dT7RBFeF',
 329                 'ext': 'mp4',
 330                 'title': 'Video 3',
 331                 'thumbnail': r're:^https?://.*\.jpg',
 332                 'view_count': int,
 333             },
 334         }],
 335         'info_dict': {
 336             'id': 'BQ0eAlwhDrw',
 337             'title': 'Post by instagram',
 338             'description': 'md5:0f9203fc6a2ce4d228da5754bcf54957',
 339         },
 340         'expected_warnings': [
 341             'General metadata extraction failed',
 342             'Main webpage is locked behind the login page',
 343         ],
 344     }, {
 345         # IGTV
 346         'url': 'https://www.instagram.com/tv/BkfuX9UB-eK/',
 347         'info_dict': {
 348             'id': 'BkfuX9UB-eK',
 349             'ext': 'mp4',
 350             'title': 'Fingerboarding Tricks with @cass.fb',
 351             'thumbnail': r're:^https?://.*\.jpg',
 352             'duration': 53.83,
 353             'timestamp': 1530032919,
 354             'upload_date': '20180626',
 355             'uploader_id': '25025320',
 356             'uploader': 'Instagram',
 357             'channel': 'instagram',
 358             'like_count': int,
 359             'comment_count': int,
 360             'comments': list,
 361             'description': 'Meet Cass Hirst (@cass.fb), a fingerboarding pro who can perform tiny ollies and kickflips while blindfolded.',
 362         },
 363         'expected_warnings': [
 364             'General metadata extraction failed',
 365             'Main webpage is locked behind the login page',
 366         ],
 367     }, {
 368         'url': 'https://instagram.com/p/-Cmh1cukG2/',
 369         'only_matching': True,
 370     }, {
 371         'url': 'http://instagram.com/p/9o6LshA7zy/embed/',
 372         'only_matching': True,
 373     }, {
 374         'url': 'https://www.instagram.com/tv/aye83DjauH/',
 375         'only_matching': True,
 376     }, {
 377         'url': 'https://www.instagram.com/reel/CDUMkliABpa/',
 378         'only_matching': True,
 379     }, {
 380         'url': 'https://www.instagram.com/marvelskies.fc/reel/CWqAgUZgCku/',
 381         'only_matching': True,
 382     }, {
 383         'url': 'https://www.instagram.com/reels/Cop84x6u7CP/',
 384         'only_matching': True,
 385     }]
 386
 387     @classmethod
 388     def _extract_embed_urls(cls, url, webpage):
 389         res = tuple(super()._extract_embed_urls(url, webpage))
 390         if res:
 391             return res
 392
 393         mobj = re.search(r'<a[^>]+href=([\'"])(?P<link>[^\'"]+)\1',
 394                          get_element_by_attribute('class', 'instagram-media', webpage) or '')
 395         if mobj:
 396             return [mobj.group('link')]
 397
 398     def _real_extract(self, url):
 399         video_id, url = self._match_valid_url(url).group('id', 'url')
 400         media, webpage = {}, ''
 401
 402         if self._get_cookies(url).get('sessionid'):
 403             info = traverse_obj(self._download_json(
 404                 f'{self._API_BASE_URL}/media/{_id_to_pk(video_id)}/info/', video_id,
 405                 fatal=False, errnote='Video info extraction failed',
 406                 note='Downloading video info', headers=self._API_HEADERS), ('items', 0))
 407             if info:
 408                 media.update(info)
 409                 return self._extract_product(media)
 410
 411         api_check = self._download_json(
 412             f'{self._API_BASE_URL}/web/get_ruling_for_content/?content_type=MEDIA&target_id={_id_to_pk(video_id)}',
 413             video_id, headers=self._API_HEADERS, fatal=False, note='Setting up session', errnote=False) or {}
 414         csrf_token = self._get_cookies('https://www.instagram.com').get('csrftoken')
 415
 416         if not csrf_token:
 417             self.report_warning('No csrf token set by Instagram API', video_id)
 418         else:
 419             csrf_token = csrf_token.value if api_check.get('status') == 'ok' else None
 420             if not csrf_token:
 421                 self.report_warning('Instagram API is not granting access', video_id)
 422
 423         variables = {
 424             'shortcode': video_id,
 425             'child_comment_count': 3,
 426             'fetch_comment_count': 40,
 427             'parent_comment_count': 24,
 428             'has_threaded_comments': True,
 429         }
 430         general_info = self._download_json(
 431             'https://www.instagram.com/graphql/query/', video_id, fatal=False, errnote=False,
 432             headers={
 433                 **self._API_HEADERS,
 434                 'X-CSRFToken': csrf_token or '',
 435                 'X-Requested-With': 'XMLHttpRequest',
 436                 'Referer': url,
 437             }, query={
 438                 'query_hash': '9f8827793ef34641b2fb195d4d41151c',
 439                 'variables': json.dumps(variables, separators=(',', ':')),
 440             })
 441         media.update(traverse_obj(general_info, ('data', 'shortcode_media')) or {})
 442
 443         if not general_info:
 444             self.report_warning('General metadata extraction failed (some metadata might be missing).', video_id)
 445             webpage, urlh = self._download_webpage_handle(url, video_id)
 446             shared_data = self._search_json(
 447                 r'window\._sharedData\s*=', webpage, 'shared data', video_id, fatal=False) or {}
 448
 449             if shared_data and self._LOGIN_URL not in urlh.url:
 450                 media.update(traverse_obj(
 451                     shared_data, ('entry_data', 'PostPage', 0, 'graphql', 'shortcode_media'),
 452                     ('entry_data', 'PostPage', 0, 'media'), expected_type=dict) or {})
 453             else:
 454                 self.report_warning('Main webpage is locked behind the login page. Retrying with embed webpage (some metadata might be missing).')
 455                 webpage = self._download_webpage(
 456                     f'{url}/embed/', video_id, note='Downloading embed webpage', fatal=False)
 457                 additional_data = self._search_json(
 458                     r'window\.__additionalDataLoaded\s*\(\s*[^,]+,', webpage, 'additional data', video_id, fatal=False)
 459                 if not additional_data and not media:
 460                     self.raise_login_required('Requested content is not available, rate-limit reached or login required')
 461
 462                 product_item = traverse_obj(additional_data, ('items', 0), expected_type=dict)
 463                 if product_item:
 464                     media.update(product_item)
 465                     return self._extract_product(media)
 466
 467                 media.update(traverse_obj(
 468                     additional_data, ('graphql', 'shortcode_media'), 'shortcode_media', expected_type=dict) or {})
 469
 470         username = traverse_obj(media, ('owner', 'username')) or self._search_regex(
 471             r'"owner"\s*:\s*{\s*"username"\s*:\s*"(.+?)"', webpage, 'username', fatal=False)
 472
 473         description = (
 474             traverse_obj(media, ('edge_media_to_caption', 'edges', 0, 'node', 'text'), expected_type=str)
 475             or media.get('caption'))
 476         if not description:
 477             description = self._search_regex(
 478                 r'"caption"\s*:\s*"(.+?)"', webpage, 'description', default=None)
 479             if description is not None:
 480                 description = lowercase_escape(description)
 481
 482         video_url = media.get('video_url')
 483         if not video_url:
 484             nodes = traverse_obj(media, ('edge_sidecar_to_children', 'edges', ..., 'node'), expected_type=dict) or []
 485             if nodes:
 486                 return self.playlist_result(
 487                     self._extract_nodes(nodes, True), video_id,
 488                     format_field(username, None, 'Post by %s'), description)
 489
 490             video_url = self._og_search_video_url(webpage, secure=False)
 491
 492         formats = [{
 493             'url': video_url,
 494             'width': self._get_dimension('width', media, webpage),
 495             'height': self._get_dimension('height', media, webpage),
 496         }]
 497         dash = traverse_obj(media, ('dash_info', 'video_dash_manifest'))
 498         if dash:
 499             formats.extend(self._parse_mpd_formats(self._parse_xml(dash, video_id), mpd_id='dash'))
 500
 501         comment_data = traverse_obj(media, ('edge_media_to_parent_comment', 'edges'))
 502         comments = [{
 503             'author': traverse_obj(comment_dict, ('node', 'owner', 'username')),
 504             'author_id': traverse_obj(comment_dict, ('node', 'owner', 'id')),
 505             'id': traverse_obj(comment_dict, ('node', 'id')),
 506             'text': traverse_obj(comment_dict, ('node', 'text')),
 507             'timestamp': traverse_obj(comment_dict, ('node', 'created_at'), expected_type=int_or_none),
 508         } for comment_dict in comment_data] if comment_data else None
 509
 510         display_resources = (
 511             media.get('display_resources')
 512             or [{'src': media.get(key)} for key in ('display_src', 'display_url')]
 513             or [{'src': self._og_search_thumbnail(webpage)}])
 514         thumbnails = [{
 515             'url': thumbnail['src'],
 516             'width': thumbnail.get('config_width'),
 517             'height': thumbnail.get('config_height'),
 518         } for thumbnail in display_resources if thumbnail.get('src')]
 519
 520         return {
 521             'id': video_id,
 522             'formats': formats,
 523             'title': media.get('title') or f'Video by {username}',
 524             'description': description,
 525             'duration': float_or_none(media.get('video_duration')),
 526             'timestamp': traverse_obj(media, 'taken_at_timestamp', 'date', expected_type=int_or_none),
 527             'uploader_id': traverse_obj(media, ('owner', 'id')),
 528             'uploader': traverse_obj(media, ('owner', 'full_name')),
 529             'channel': username,
 530             'like_count': self._get_count(media, 'likes', 'preview_like') or str_to_int(self._search_regex(
 531                 r'data-log-event="likeCountClick"[^>]*>[^\d]*([\d,\.]+)', webpage, 'like count', fatal=False)),
 532             'comment_count': self._get_count(media, 'comments', 'preview_comment', 'to_comment', 'to_parent_comment'),
 533             'comments': comments,
 534             'thumbnails': thumbnails,
 535             'http_headers': {
 536                 'Referer': 'https://www.instagram.com/',
 537             },
 538         }
 539
 540
 541 class InstagramPlaylistBaseIE(InstagramBaseIE):
 542     _gis_tmpl = None  # used to cache GIS request type
 543
 544     def _parse_graphql(self, webpage, item_id):
 545         # Reads a webpage and returns its GraphQL data.
 546         return self._parse_json(
 547             self._search_regex(
 548                 r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
 549             item_id)
 550
 551     def _extract_graphql(self, data, url):
 552         # Parses GraphQL queries containing videos and generates a playlist.
 553         uploader_id = self._match_id(url)
 554         csrf_token = data['config']['csrf_token']
 555         rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
 556
 557         cursor = ''
 558         for page_num in itertools.count(1):
 559             variables = {
 560                 'first': 12,
 561                 'after': cursor,
 562             }
 563             variables.update(self._query_vars_for(data))
 564             variables = json.dumps(variables)
 565
 566             if self._gis_tmpl:
 567                 gis_tmpls = [self._gis_tmpl]
 568             else:
 569                 gis_tmpls = [
 570                     f'{rhx_gis}',
 571                     '',
 572                     f'{rhx_gis}:{csrf_token}',
 573                     '{}:{}:{}'.format(rhx_gis, csrf_token, self.get_param('http_headers')['User-Agent']),
 574                 ]
 575
 576             # try all of the ways to generate a GIS query, and not only use the
 577             # first one that works, but cache it for future requests
 578             for gis_tmpl in gis_tmpls:
 579                 try:
 580                     json_data = self._download_json(
 581                         'https://www.instagram.com/graphql/query/', uploader_id,
 582                         f'Downloading JSON page {page_num}', headers={
 583                             'X-Requested-With': 'XMLHttpRequest',
 584                             'X-Instagram-GIS': hashlib.md5(
 585                                 (f'{gis_tmpl}:{variables}').encode()).hexdigest(),
 586                         }, query={
 587                             'query_hash': self._QUERY_HASH,
 588                             'variables': variables,
 589                         })
 590                     media = self._parse_timeline_from(json_data)
 591                     self._gis_tmpl = gis_tmpl
 592                     break
 593                 except ExtractorError as e:
 594                     # if it's an error caused by a bad query, and there are
 595                     # more GIS templates to try, ignore it and keep trying
 596                     if isinstance(e.cause, HTTPError) and e.cause.status == 403:
 597                         if gis_tmpl != gis_tmpls[-1]:
 598                             continue
 599                     raise
 600
 601             nodes = traverse_obj(media, ('edges', ..., 'node'), expected_type=dict) or []
 602             if not nodes:
 603                 break
 604             yield from self._extract_nodes(nodes)
 605
 606             has_next_page = traverse_obj(media, ('page_info', 'has_next_page'))
 607             cursor = traverse_obj(media, ('page_info', 'end_cursor'), expected_type=str)
 608             if not has_next_page or not cursor:
 609                 break
 610
 611     def _real_extract(self, url):
 612         user_or_tag = self._match_id(url)
 613         webpage = self._download_webpage(url, user_or_tag)
 614         data = self._parse_graphql(webpage, user_or_tag)
 615
 616         self._set_cookie('instagram.com', 'ig_pr', '1')
 617
 618         return self.playlist_result(
 619             self._extract_graphql(data, url), user_or_tag, user_or_tag)
 620
 621
 622 class InstagramUserIE(InstagramPlaylistBaseIE):
 623     _WORKING = False
 624     _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
 625     IE_DESC = 'Instagram user profile'
 626     IE_NAME = 'instagram:user'
 627     _TESTS = [{
 628         'url': 'https://instagram.com/porsche',
 629         'info_dict': {
 630             'id': 'porsche',
 631             'title': 'porsche',
 632         },
 633         'playlist_count': 5,
 634         'params': {
 635             'extract_flat': True,
 636             'skip_download': True,
 637             'playlistend': 5,
 638         },
 639     }]
 640
 641     _QUERY_HASH = ('42323d64886122307be10013ad2dcc44',)
 642
 643     @staticmethod
 644     def _parse_timeline_from(data):
 645         # extracts the media timeline data from a GraphQL result
 646         return data['data']['user']['edge_owner_to_timeline_media']
 647
 648     @staticmethod
 649     def _query_vars_for(data):
 650         # returns a dictionary of variables to add to the timeline query based
 651         # on the GraphQL of the original page
 652         return {
 653             'id': data['entry_data']['ProfilePage'][0]['graphql']['user']['id'],
 654         }
 655
 656
 657 class InstagramTagIE(InstagramPlaylistBaseIE):
 658     _VALID_URL = r'https?://(?:www\.)?instagram\.com/explore/tags/(?P<id>[^/]+)'
 659     IE_DESC = 'Instagram hashtag search URLs'
 660     IE_NAME = 'instagram:tag'
 661     _TESTS = [{
 662         'url': 'https://instagram.com/explore/tags/lolcats',
 663         'info_dict': {
 664             'id': 'lolcats',
 665             'title': 'lolcats',
 666         },
 667         'playlist_count': 50,
 668         'params': {
 669             'extract_flat': True,
 670             'skip_download': True,
 671             'playlistend': 50,
 672         },
 673     }]
 674
 675     _QUERY_HASH = ('f92f56d47dc7a55b606908374b43a314',)
 676
 677     @staticmethod
 678     def _parse_timeline_from(data):
 679         # extracts the media timeline data from a GraphQL result
 680         return data['data']['hashtag']['edge_hashtag_to_media']
 681
 682     @staticmethod
 683     def _query_vars_for(data):
 684         # returns a dictionary of variables to add to the timeline query based
 685         # on the GraphQL of the original page
 686         return {
 687             'tag_name':
 688                 data['entry_data']['TagPage'][0]['graphql']['hashtag']['name'],
 689         }
 690
 691
 692 class InstagramStoryIE(InstagramBaseIE):
 693     _VALID_URL = r'https?://(?:www\.)?instagram\.com/stories/(?P<user>[^/]+)/(?P<id>\d+)'
 694     IE_NAME = 'instagram:story'
 695
 696     _TESTS = [{
 697         'url': 'https://www.instagram.com/stories/highlights/18090946048123978/',
 698         'info_dict': {
 699             'id': '18090946048123978',
 700             'title': 'Rare',
 701         },
 702         'playlist_mincount': 50,
 703     }]
 704
 705     def _real_extract(self, url):
 706         username, story_id = self._match_valid_url(url).groups()
 707         story_info = self._download_webpage(url, story_id)
 708         user_info = self._search_json(r'"user":', story_info, 'user info', story_id, fatal=False)
 709         if not user_info:
 710             self.raise_login_required('This content is unreachable')
 711
 712         user_id = traverse_obj(user_info, 'pk', 'id', expected_type=str)
 713         story_info_url = user_id if username != 'highlights' else f'highlight:{story_id}'
 714         if not story_info_url:  # user id is only mandatory for non-highlights
 715             raise ExtractorError('Unable to extract user id')
 716
 717         videos = traverse_obj(self._download_json(
 718             f'{self._API_BASE_URL}/feed/reels_media/?reel_ids={story_info_url}',
 719             story_id, errnote=False, fatal=False, headers=self._API_HEADERS), 'reels')
 720         if not videos:
 721             self.raise_login_required('You need to log in to access this content')
 722
 723         full_name = traverse_obj(videos, (f'highlight:{story_id}', 'user', 'full_name'), (user_id, 'user', 'full_name'))
 724         story_title = traverse_obj(videos, (f'highlight:{story_id}', 'title'))
 725         if not story_title:
 726             story_title = f'Story by {username}'
 727
 728         highlights = traverse_obj(videos, (f'highlight:{story_id}', 'items'), (user_id, 'items'))
 729         info_data = []
 730         for highlight in highlights:
 731             highlight_data = self._extract_product(highlight)
 732             if highlight_data.get('formats'):
 733                 info_data.append({
 734                     'uploader': full_name,
 735                     'uploader_id': user_id,
 736                     **filter_dict(highlight_data),
 737                 })
 738         return self.playlist_result(info_data, playlist_id=story_id, playlist_title=story_title)