yt_dlp/extractor/instagram.py

   1 import hashlib
   2 import itertools
   3 import json
   4 import re
   5 import time
   6 import urllib.error
   7
   8 from .common import InfoExtractor
   9 from ..utils import (
  10     ExtractorError,
  11     decode_base_n,
  12     encode_base_n,
  13     float_or_none,
  14     format_field,
  15     get_element_by_attribute,
  16     int_or_none,
  17     lowercase_escape,
  18     str_or_none,
  19     str_to_int,
  20     traverse_obj,
  21     url_or_none,
  22     urlencode_postdata,
  23 )
  24
  25 _ENCODING_CHARS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_'
  26
  27
  28 def _pk_to_id(id):
  29     """Source: https://stackoverflow.com/questions/24437823/getting-instagram-post-url-from-media-id"""
  30     return encode_base_n(int(id.split('_')[0]), table=_ENCODING_CHARS)
  31
  32
  33 def _id_to_pk(shortcode):
  34     """Covert a shortcode to a numeric value"""
  35     return decode_base_n(shortcode[:11], table=_ENCODING_CHARS)
  36
  37
  38 class InstagramBaseIE(InfoExtractor):
  39     _NETRC_MACHINE = 'instagram'
  40     _IS_LOGGED_IN = False
  41
  42     _API_BASE_URL = 'https://i.instagram.com/api/v1'
  43     _LOGIN_URL = 'https://www.instagram.com/accounts/login'
  44     _API_HEADERS = {
  45         'X-IG-App-ID': '936619743392459',
  46         'X-ASBD-ID': '198387',
  47         'X-IG-WWW-Claim': '0',
  48         'Origin': 'https://www.instagram.com',
  49         'Accept': '*/*',
  50         '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',
  51     }
  52
  53     def _perform_login(self, username, password):
  54         if self._IS_LOGGED_IN:
  55             return
  56
  57         login_webpage = self._download_webpage(
  58             self._LOGIN_URL, None, note='Downloading login webpage', errnote='Failed to download login webpage')
  59
  60         shared_data = self._parse_json(self._search_regex(
  61             r'window\._sharedData\s*=\s*({.+?});', login_webpage, 'shared data', default='{}'), None)
  62
  63         login = self._download_json(
  64             f'{self._LOGIN_URL}/ajax/', None, note='Logging in', headers={
  65                 **self._API_HEADERS,
  66                 'X-Requested-With': 'XMLHttpRequest',
  67                 'X-CSRFToken': shared_data['config']['csrf_token'],
  68                 'X-Instagram-AJAX': shared_data['rollout_hash'],
  69                 'Referer': 'https://www.instagram.com/',
  70             }, data=urlencode_postdata({
  71                 'enc_password': f'#PWD_INSTAGRAM_BROWSER:0:{int(time.time())}:{password}',
  72                 'username': username,
  73                 'queryParams': '{}',
  74                 'optIntoOneTap': 'false',
  75                 'stopDeletionNonce': '',
  76                 'trustedDeviceRecords': '{}',
  77             }))
  78
  79         if not login.get('authenticated'):
  80             if login.get('message'):
  81                 raise ExtractorError(f'Unable to login: {login["message"]}')
  82             elif login.get('user'):
  83                 raise ExtractorError('Unable to login: Sorry, your password was incorrect. Please double-check your password.', expected=True)
  84             elif login.get('user') is False:
  85                 raise ExtractorError('Unable to login: The username you entered doesn\'t belong to an account. Please check your username and try again.', expected=True)
  86             raise ExtractorError('Unable to login')
  87         InstagramBaseIE._IS_LOGGED_IN = True
  88
  89     def _get_count(self, media, kind, *keys):
  90         return traverse_obj(
  91             media, (kind, 'count'), *((f'edge_media_{key}', 'count') for key in keys),
  92             expected_type=int_or_none)
  93
  94     def _get_dimension(self, name, media, webpage=None):
  95         return (
  96             traverse_obj(media, ('dimensions', name), expected_type=int_or_none)
  97             or int_or_none(self._html_search_meta(
  98                 (f'og:video:{name}', f'video:{name}'), webpage or '', default=None)))
  99
 100     def _extract_nodes(self, nodes, is_direct=False):
 101         for idx, node in enumerate(nodes, start=1):
 102             if node.get('__typename') != 'GraphVideo' and node.get('is_video') is not True:
 103                 continue
 104
 105             video_id = node.get('shortcode')
 106
 107             if is_direct:
 108                 info = {
 109                     'id': video_id or node['id'],
 110                     'url': node.get('video_url'),
 111                     'width': self._get_dimension('width', node),
 112                     'height': self._get_dimension('height', node),
 113                     'http_headers': {
 114                         'Referer': 'https://www.instagram.com/',
 115                     }
 116                 }
 117             elif not video_id:
 118                 continue
 119             else:
 120                 info = {
 121                     '_type': 'url',
 122                     'ie_key': 'Instagram',
 123                     'id': video_id,
 124                     'url': f'https://instagram.com/p/{video_id}',
 125                 }
 126
 127             yield {
 128                 **info,
 129                 'title': node.get('title') or (f'Video {idx}' if is_direct else None),
 130                 'description': traverse_obj(
 131                     node, ('edge_media_to_caption', 'edges', 0, 'node', 'text'), expected_type=str),
 132                 'thumbnail': traverse_obj(
 133                     node, 'display_url', 'thumbnail_src', 'display_src', expected_type=url_or_none),
 134                 'duration': float_or_none(node.get('video_duration')),
 135                 'timestamp': int_or_none(node.get('taken_at_timestamp')),
 136                 'view_count': int_or_none(node.get('video_view_count')),
 137                 'comment_count': self._get_count(node, 'comments', 'preview_comment', 'to_comment', 'to_parent_comment'),
 138                 'like_count': self._get_count(node, 'likes', 'preview_like'),
 139             }
 140
 141     def _extract_product_media(self, product_media):
 142         media_id = product_media.get('code') or _pk_to_id(product_media.get('pk'))
 143         vcodec = product_media.get('video_codec')
 144         dash_manifest_raw = product_media.get('video_dash_manifest')
 145         videos_list = product_media.get('video_versions')
 146         if not (dash_manifest_raw or videos_list):
 147             return {}
 148
 149         formats = [{
 150             'format_id': format.get('id'),
 151             'url': format.get('url'),
 152             'width': format.get('width'),
 153             'height': format.get('height'),
 154             'vcodec': vcodec,
 155         } for format in videos_list or []]
 156         if dash_manifest_raw:
 157             formats.extend(self._parse_mpd_formats(self._parse_xml(dash_manifest_raw, media_id), mpd_id='dash'))
 158         self._sort_formats(formats)
 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|reel)/(?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
 384     @classmethod
 385     def _extract_embed_urls(cls, url, webpage):
 386         res = tuple(super()._extract_embed_urls(url, webpage))
 387         if res:
 388             return res
 389
 390         mobj = re.search(r'<a[^>]+href=([\'"])(?P<link>[^\'"]+)\1',
 391                          get_element_by_attribute('class', 'instagram-media', webpage) or '')
 392         if mobj:
 393             return [mobj.group('link')]
 394
 395     def _real_extract(self, url):
 396         video_id, url = self._match_valid_url(url).group('id', 'url')
 397         media, webpage = {}, ''
 398
 399         if self._get_cookies(url).get('sessionid'):
 400             info = traverse_obj(self._download_json(
 401                 f'{self._API_BASE_URL}/media/{_id_to_pk(video_id)}/info/', video_id,
 402                 fatal=False, errnote='Video info extraction failed',
 403                 note='Downloading video info', headers=self._API_HEADERS), ('items', 0))
 404             if info:
 405                 media.update(info)
 406                 return self._extract_product(media)
 407
 408         api_check = self._download_json(
 409             f'{self._API_BASE_URL}/web/get_ruling_for_content/?content_type=MEDIA&target_id={_id_to_pk(video_id)}',
 410             video_id, headers=self._API_HEADERS, fatal=False, note='Setting up session', errnote=False) or {}
 411         csrf_token = self._get_cookies('https://www.instagram.com').get('csrftoken')
 412
 413         if not csrf_token:
 414             self.report_warning('No csrf token set by Instagram API', video_id)
 415         else:
 416             csrf_token = csrf_token.value if api_check.get('status') == 'ok' else None
 417             if not csrf_token:
 418                 self.report_warning('Instagram API is not granting access', video_id)
 419
 420         variables = {
 421             'shortcode': video_id,
 422             'child_comment_count': 3,
 423             'fetch_comment_count': 40,
 424             'parent_comment_count': 24,
 425             'has_threaded_comments': True,
 426         }
 427         general_info = self._download_json(
 428             'https://www.instagram.com/graphql/query/', video_id, fatal=False, errnote=False,
 429             headers={
 430                 **self._API_HEADERS,
 431                 'X-CSRFToken': csrf_token or '',
 432                 'X-Requested-With': 'XMLHttpRequest',
 433                 'Referer': url,
 434             }, query={
 435                 'query_hash': '9f8827793ef34641b2fb195d4d41151c',
 436                 'variables': json.dumps(variables, separators=(',', ':')),
 437             })
 438         media.update(traverse_obj(general_info, ('data', 'shortcode_media')) or {})
 439
 440         if not general_info:
 441             self.report_warning('General metadata extraction failed (some metadata might be missing).', video_id)
 442             webpage, urlh = self._download_webpage_handle(url, video_id)
 443             shared_data = self._search_json(
 444                 r'window\._sharedData\s*=', webpage, 'shared data', video_id, fatal=False) or {}
 445
 446             if shared_data and self._LOGIN_URL not in urlh.geturl():
 447                 media.update(traverse_obj(
 448                     shared_data, ('entry_data', 'PostPage', 0, 'graphql', 'shortcode_media'),
 449                     ('entry_data', 'PostPage', 0, 'media'), expected_type=dict) or {})
 450             else:
 451                 self.report_warning('Main webpage is locked behind the login page. Retrying with embed webpage (some metadata might be missing).')
 452                 webpage = self._download_webpage(
 453                     f'{url}/embed/', video_id, note='Downloading embed webpage', fatal=False)
 454                 additional_data = self._search_json(
 455                     r'window\.__additionalDataLoaded\s*\(\s*[^,]+,', webpage, 'additional data', video_id, fatal=False)
 456                 if not additional_data and not media:
 457                     self.raise_login_required('Requested content is not available, rate-limit reached or login required')
 458
 459                 product_item = traverse_obj(additional_data, ('items', 0), expected_type=dict)
 460                 if product_item:
 461                     media.update(product_item)
 462                     return self._extract_product(media)
 463
 464                 media.update(traverse_obj(
 465                     additional_data, ('graphql', 'shortcode_media'), 'shortcode_media', expected_type=dict) or {})
 466
 467         username = traverse_obj(media, ('owner', 'username')) or self._search_regex(
 468             r'"owner"\s*:\s*{\s*"username"\s*:\s*"(.+?)"', webpage, 'username', fatal=False)
 469
 470         description = (
 471             traverse_obj(media, ('edge_media_to_caption', 'edges', 0, 'node', 'text'), expected_type=str)
 472             or media.get('caption'))
 473         if not description:
 474             description = self._search_regex(
 475                 r'"caption"\s*:\s*"(.+?)"', webpage, 'description', default=None)
 476             if description is not None:
 477                 description = lowercase_escape(description)
 478
 479         video_url = media.get('video_url')
 480         if not video_url:
 481             nodes = traverse_obj(media, ('edge_sidecar_to_children', 'edges', ..., 'node'), expected_type=dict) or []
 482             if nodes:
 483                 return self.playlist_result(
 484                     self._extract_nodes(nodes, True), video_id,
 485                     format_field(username, None, 'Post by %s'), description)
 486
 487             video_url = self._og_search_video_url(webpage, secure=False)
 488
 489         formats = [{
 490             'url': video_url,
 491             'width': self._get_dimension('width', media, webpage),
 492             'height': self._get_dimension('height', media, webpage),
 493         }]
 494         dash = traverse_obj(media, ('dash_info', 'video_dash_manifest'))
 495         if dash:
 496             formats.extend(self._parse_mpd_formats(self._parse_xml(dash, video_id), mpd_id='dash'))
 497         self._sort_formats(formats)
 498
 499         comment_data = traverse_obj(media, ('edge_media_to_parent_comment', 'edges'))
 500         comments = [{
 501             'author': traverse_obj(comment_dict, ('node', 'owner', 'username')),
 502             'author_id': traverse_obj(comment_dict, ('node', 'owner', 'id')),
 503             'id': traverse_obj(comment_dict, ('node', 'id')),
 504             'text': traverse_obj(comment_dict, ('node', 'text')),
 505             'timestamp': traverse_obj(comment_dict, ('node', 'created_at'), expected_type=int_or_none),
 506         } for comment_dict in comment_data] if comment_data else None
 507
 508         display_resources = (
 509             media.get('display_resources')
 510             or [{'src': media.get(key)} for key in ('display_src', 'display_url')]
 511             or [{'src': self._og_search_thumbnail(webpage)}])
 512         thumbnails = [{
 513             'url': thumbnail['src'],
 514             'width': thumbnail.get('config_width'),
 515             'height': thumbnail.get('config_height'),
 516         } for thumbnail in display_resources if thumbnail.get('src')]
 517
 518         return {
 519             'id': video_id,
 520             'formats': formats,
 521             'title': media.get('title') or 'Video by %s' % username,
 522             'description': description,
 523             'duration': float_or_none(media.get('video_duration')),
 524             'timestamp': traverse_obj(media, 'taken_at_timestamp', 'date', expected_type=int_or_none),
 525             'uploader_id': traverse_obj(media, ('owner', 'id')),
 526             'uploader': traverse_obj(media, ('owner', 'full_name')),
 527             'channel': username,
 528             'like_count': self._get_count(media, 'likes', 'preview_like') or str_to_int(self._search_regex(
 529                 r'data-log-event="likeCountClick"[^>]*>[^\d]*([\d,\.]+)', webpage, 'like count', fatal=False)),
 530             'comment_count': self._get_count(media, 'comments', 'preview_comment', 'to_comment', 'to_parent_comment'),
 531             'comments': comments,
 532             'thumbnails': thumbnails,
 533             'http_headers': {
 534                 'Referer': 'https://www.instagram.com/',
 535             }
 536         }
 537
 538
 539 class InstagramPlaylistBaseIE(InstagramBaseIE):
 540     _gis_tmpl = None  # used to cache GIS request type
 541
 542     def _parse_graphql(self, webpage, item_id):
 543         # Reads a webpage and returns its GraphQL data.
 544         return self._parse_json(
 545             self._search_regex(
 546                 r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
 547             item_id)
 548
 549     def _extract_graphql(self, data, url):
 550         # Parses GraphQL queries containing videos and generates a playlist.
 551         uploader_id = self._match_id(url)
 552         csrf_token = data['config']['csrf_token']
 553         rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
 554
 555         cursor = ''
 556         for page_num in itertools.count(1):
 557             variables = {
 558                 'first': 12,
 559                 'after': cursor,
 560             }
 561             variables.update(self._query_vars_for(data))
 562             variables = json.dumps(variables)
 563
 564             if self._gis_tmpl:
 565                 gis_tmpls = [self._gis_tmpl]
 566             else:
 567                 gis_tmpls = [
 568                     '%s' % rhx_gis,
 569                     '',
 570                     '%s:%s' % (rhx_gis, csrf_token),
 571                     '%s:%s:%s' % (rhx_gis, csrf_token, self.get_param('http_headers')['User-Agent']),
 572                 ]
 573
 574             # try all of the ways to generate a GIS query, and not only use the
 575             # first one that works, but cache it for future requests
 576             for gis_tmpl in gis_tmpls:
 577                 try:
 578                     json_data = self._download_json(
 579                         'https://www.instagram.com/graphql/query/', uploader_id,
 580                         'Downloading JSON page %d' % page_num, headers={
 581                             'X-Requested-With': 'XMLHttpRequest',
 582                             'X-Instagram-GIS': hashlib.md5(
 583                                 ('%s:%s' % (gis_tmpl, variables)).encode('utf-8')).hexdigest(),
 584                         }, query={
 585                             'query_hash': self._QUERY_HASH,
 586                             'variables': variables,
 587                         })
 588                     media = self._parse_timeline_from(json_data)
 589                     self._gis_tmpl = gis_tmpl
 590                     break
 591                 except ExtractorError as e:
 592                     # if it's an error caused by a bad query, and there are
 593                     # more GIS templates to try, ignore it and keep trying
 594                     if isinstance(e.cause, urllib.error.HTTPError) and e.cause.code == 403:
 595                         if gis_tmpl != gis_tmpls[-1]:
 596                             continue
 597                     raise
 598
 599             nodes = traverse_obj(media, ('edges', ..., 'node'), expected_type=dict) or []
 600             if not nodes:
 601                 break
 602             yield from self._extract_nodes(nodes)
 603
 604             has_next_page = traverse_obj(media, ('page_info', 'has_next_page'))
 605             cursor = traverse_obj(media, ('page_info', 'end_cursor'), expected_type=str)
 606             if not has_next_page or not cursor:
 607                 break
 608
 609     def _real_extract(self, url):
 610         user_or_tag = self._match_id(url)
 611         webpage = self._download_webpage(url, user_or_tag)
 612         data = self._parse_graphql(webpage, user_or_tag)
 613
 614         self._set_cookie('instagram.com', 'ig_pr', '1')
 615
 616         return self.playlist_result(
 617             self._extract_graphql(data, url), user_or_tag, user_or_tag)
 618
 619
 620 class InstagramUserIE(InstagramPlaylistBaseIE):
 621     _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
 622     IE_DESC = 'Instagram user profile'
 623     IE_NAME = 'instagram:user'
 624     _TESTS = [{
 625         'url': 'https://instagram.com/porsche',
 626         'info_dict': {
 627             'id': 'porsche',
 628             'title': 'porsche',
 629         },
 630         'playlist_count': 5,
 631         'params': {
 632             'extract_flat': True,
 633             'skip_download': True,
 634             'playlistend': 5,
 635         }
 636     }]
 637
 638     _QUERY_HASH = '42323d64886122307be10013ad2dcc44',
 639
 640     @staticmethod
 641     def _parse_timeline_from(data):
 642         # extracts the media timeline data from a GraphQL result
 643         return data['data']['user']['edge_owner_to_timeline_media']
 644
 645     @staticmethod
 646     def _query_vars_for(data):
 647         # returns a dictionary of variables to add to the timeline query based
 648         # on the GraphQL of the original page
 649         return {
 650             'id': data['entry_data']['ProfilePage'][0]['graphql']['user']['id']
 651         }
 652
 653
 654 class InstagramTagIE(InstagramPlaylistBaseIE):
 655     _VALID_URL = r'https?://(?:www\.)?instagram\.com/explore/tags/(?P<id>[^/]+)'
 656     IE_DESC = 'Instagram hashtag search URLs'
 657     IE_NAME = 'instagram:tag'
 658     _TESTS = [{
 659         'url': 'https://instagram.com/explore/tags/lolcats',
 660         'info_dict': {
 661             'id': 'lolcats',
 662             'title': 'lolcats',
 663         },
 664         'playlist_count': 50,
 665         'params': {
 666             'extract_flat': True,
 667             'skip_download': True,
 668             'playlistend': 50,
 669         }
 670     }]
 671
 672     _QUERY_HASH = 'f92f56d47dc7a55b606908374b43a314',
 673
 674     @staticmethod
 675     def _parse_timeline_from(data):
 676         # extracts the media timeline data from a GraphQL result
 677         return data['data']['hashtag']['edge_hashtag_to_media']
 678
 679     @staticmethod
 680     def _query_vars_for(data):
 681         # returns a dictionary of variables to add to the timeline query based
 682         # on the GraphQL of the original page
 683         return {
 684             'tag_name':
 685                 data['entry_data']['TagPage'][0]['graphql']['hashtag']['name']
 686         }
 687
 688
 689 class InstagramStoryIE(InstagramBaseIE):
 690     _VALID_URL = r'https?://(?:www\.)?instagram\.com/stories/(?P<user>[^/]+)/(?P<id>\d+)'
 691     IE_NAME = 'instagram:story'
 692
 693     _TESTS = [{
 694         'url': 'https://www.instagram.com/stories/highlights/18090946048123978/',
 695         'info_dict': {
 696             'id': '18090946048123978',
 697             'title': 'Rare',
 698         },
 699         'playlist_mincount': 50
 700     }]
 701
 702     def _real_extract(self, url):
 703         username, story_id = self._match_valid_url(url).groups()
 704         story_info = self._download_webpage(url, story_id)
 705         user_info = self._search_json(r'"user":', story_info, 'user info', story_id, fatal=False)
 706         if not user_info:
 707             self.raise_login_required('This content is unreachable')
 708         user_id = user_info.get('id')
 709
 710         story_info_url = user_id if username != 'highlights' else f'highlight:{story_id}'
 711         videos = traverse_obj(self._download_json(
 712             f'{self._API_BASE_URL}/feed/reels_media/?reel_ids={story_info_url}',
 713             story_id, errnote=False, fatal=False, headers=self._API_HEADERS), 'reels')
 714         if not videos:
 715             self.raise_login_required('You need to log in to access this content')
 716
 717         full_name = traverse_obj(videos, (f'highlight:{story_id}', 'user', 'full_name'), (str(user_id), 'user', 'full_name'))
 718         story_title = traverse_obj(videos, (f'highlight:{story_id}', 'title'))
 719         if not story_title:
 720             story_title = f'Story by {username}'
 721
 722         highlights = traverse_obj(videos, (f'highlight:{story_id}', 'items'), (str(user_id), 'items'))
 723         info_data = []
 724         for highlight in highlights:
 725             highlight_data = self._extract_product(highlight)
 726             if highlight_data.get('formats'):
 727                 info_data.append({
 728                     **highlight_data,
 729                     'uploader': full_name,
 730                     'uploader_id': user_id,
 731                 })
 732         return self.playlist_result(info_data, playlist_id=story_id, playlist_title=story_title)