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     }
  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': fmt.get('id'),
 151             'url': fmt.get('url'),
 152             'width': fmt.get('width'),
 153             'height': fmt.get('height'),
 154             'vcodec': vcodec,
 155         } for fmt 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
 159         thumbnails = [{
 160             'url': thumbnail.get('url'),
 161             'width': thumbnail.get('width'),
 162             'height': thumbnail.get('height'),
 163         } for thumbnail in traverse_obj(product_media, ('image_versions2', 'candidates')) or []]
 164         return {
 165             'id': media_id,
 166             'duration': float_or_none(product_media.get('video_duration')),
 167             'formats': formats,
 168             'thumbnails': thumbnails,
 169         }
 170
 171     def _extract_product(self, product_info):
 172         if isinstance(product_info, list):
 173             product_info = product_info[0]
 174
 175         user_info = product_info.get('user') or {}
 176         info_dict = {
 177             'id': _pk_to_id(traverse_obj(product_info, 'pk', 'id', expected_type=str_or_none)[:19]),
 178             'title': product_info.get('title') or f'Video by {user_info.get("username")}',
 179             'description': traverse_obj(product_info, ('caption', 'text'), expected_type=str_or_none),
 180             'timestamp': int_or_none(product_info.get('taken_at')),
 181             'channel': user_info.get('username'),
 182             'uploader': user_info.get('full_name'),
 183             'uploader_id': str_or_none(user_info.get('pk')),
 184             'view_count': int_or_none(product_info.get('view_count')),
 185             'like_count': int_or_none(product_info.get('like_count')),
 186             'comment_count': int_or_none(product_info.get('comment_count')),
 187             '__post_extractor': self.extract_comments(_pk_to_id(product_info.get('pk'))),
 188             'http_headers': {
 189                 'Referer': 'https://www.instagram.com/',
 190             },
 191         }
 192         carousel_media = product_info.get('carousel_media')
 193         if carousel_media:
 194             return {
 195                 '_type': 'playlist',
 196                 **info_dict,
 197                 'title': f'Post by {user_info.get("username")}',
 198                 'entries': [{
 199                     **info_dict,
 200                     **self._extract_product_media(product_media),
 201                 } for product_media in carousel_media],
 202             }
 203
 204         return {
 205             **info_dict,
 206             **self._extract_product_media(product_info),
 207         }
 208
 209     def _get_comments(self, video_id):
 210         comments_info = self._download_json(
 211             f'{self._API_BASE_URL}/media/{_id_to_pk(video_id)}/comments/?can_support_threading=true&permalink_enabled=false', video_id,
 212             fatal=False, errnote='Comments extraction failed', note='Downloading comments info', headers=self._API_HEADERS) or {}
 213
 214         comment_data = traverse_obj(comments_info, ('edge_media_to_parent_comment', 'edges'), 'comments')
 215         for comment_dict in comment_data or []:
 216             yield {
 217                 'author': traverse_obj(comment_dict, ('node', 'owner', 'username'), ('user', 'username')),
 218                 'author_id': traverse_obj(comment_dict, ('node', 'owner', 'id'), ('user', 'pk')),
 219                 'author_thumbnail': traverse_obj(comment_dict, ('node', 'owner', 'profile_pic_url'), ('user', 'profile_pic_url'), expected_type=url_or_none),
 220                 'id': traverse_obj(comment_dict, ('node', 'id'), 'pk'),
 221                 'text': traverse_obj(comment_dict, ('node', 'text'), 'text'),
 222                 'like_count': traverse_obj(comment_dict, ('node', 'edge_liked_by', 'count'), 'comment_like_count', expected_type=int_or_none),
 223                 'timestamp': traverse_obj(comment_dict, ('node', 'created_at'), 'created_at', expected_type=int_or_none),
 224             }
 225
 226
 227 class InstagramIOSIE(InfoExtractor):
 228     IE_DESC = 'IOS instagram:// URL'
 229     _VALID_URL = r'instagram://media\?id=(?P<id>[\d_]+)'
 230     _TESTS = [{
 231         'url': 'instagram://media?id=482584233761418119',
 232         'md5': '0d2da106a9d2631273e192b372806516',
 233         'info_dict': {
 234             'id': 'aye83DjauH',
 235             'ext': 'mp4',
 236             'title': 'Video by naomipq',
 237             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
 238             'thumbnail': r're:^https?://.*\.jpg',
 239             'duration': 0,
 240             'timestamp': 1371748545,
 241             'upload_date': '20130620',
 242             'uploader_id': 'naomipq',
 243             'uploader': 'B E A U T Y  F O R  A S H E S',
 244             'like_count': int,
 245             'comment_count': int,
 246             'comments': list,
 247         },
 248         'add_ie': ['Instagram'],
 249     }]
 250
 251     def _real_extract(self, url):
 252         video_id = _pk_to_id(self._match_id(url))
 253         return self.url_result(f'http://instagram.com/tv/{video_id}', InstagramIE, video_id)
 254
 255
 256 class InstagramIE(InstagramBaseIE):
 257     _VALID_URL = r'(?P<url>https?://(?:www\.)?instagram\.com(?:/[^/]+)?/(?:p|tv|reels?(?!/audio/))/(?P<id>[^/?#&]+))'
 258     _EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?instagram\.com/p/[^/]+/embed.*?)\1']
 259     _TESTS = [{
 260         'url': 'https://instagram.com/p/aye83DjauH/?foo=bar#abc',
 261         'md5': '0d2da106a9d2631273e192b372806516',
 262         'info_dict': {
 263             'id': 'aye83DjauH',
 264             'ext': 'mp4',
 265             'title': 'Video by naomipq',
 266             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
 267             'thumbnail': r're:^https?://.*\.jpg',
 268             'duration': 8.747,
 269             'timestamp': 1371748545,
 270             'upload_date': '20130620',
 271             'uploader_id': '2815873',
 272             'uploader': 'B E A U T Y  F O R  A S H E S',
 273             'channel': 'naomipq',
 274             'like_count': int,
 275             'comment_count': int,
 276             'comments': list,
 277         },
 278         'expected_warnings': [
 279             'General metadata extraction failed',
 280             'Main webpage is locked behind the login page',
 281         ],
 282     }, {
 283         # reel
 284         'url': 'https://www.instagram.com/reel/Chunk8-jurw/',
 285         'md5': 'f6d8277f74515fa3ff9f5791426e42b1',
 286         'info_dict': {
 287             'id': 'Chunk8-jurw',
 288             'ext': 'mp4',
 289             'title': 'Video by instagram',
 290             'description': 'md5:c9cde483606ed6f80fbe9283a6a2b290',
 291             'thumbnail': r're:^https?://.*\.jpg',
 292             'duration': 5.016,
 293             'timestamp': 1661529231,
 294             'upload_date': '20220826',
 295             'uploader_id': '25025320',
 296             'uploader': 'Instagram',
 297             'channel': 'instagram',
 298             'like_count': int,
 299             'comment_count': int,
 300             'comments': list,
 301         },
 302         'expected_warnings': [
 303             'General metadata extraction failed',
 304             'Main webpage is locked behind the login page',
 305         ],
 306     }, {
 307         # multi video post
 308         'url': 'https://www.instagram.com/p/BQ0eAlwhDrw/',
 309         'playlist': [{
 310             'info_dict': {
 311                 'id': 'BQ0dSaohpPW',
 312                 'ext': 'mp4',
 313                 'title': 'Video 1',
 314                 'thumbnail': r're:^https?://.*\.jpg',
 315                 'view_count': int,
 316             },
 317         }, {
 318             'info_dict': {
 319                 'id': 'BQ0dTpOhuHT',
 320                 'ext': 'mp4',
 321                 'title': 'Video 2',
 322                 'thumbnail': r're:^https?://.*\.jpg',
 323                 'view_count': int,
 324             },
 325         }, {
 326             'info_dict': {
 327                 'id': 'BQ0dT7RBFeF',
 328                 'ext': 'mp4',
 329                 'title': 'Video 3',
 330                 'thumbnail': r're:^https?://.*\.jpg',
 331                 'view_count': int,
 332             },
 333         }],
 334         'info_dict': {
 335             'id': 'BQ0eAlwhDrw',
 336             'title': 'Post by instagram',
 337             'description': 'md5:0f9203fc6a2ce4d228da5754bcf54957',
 338         },
 339         'expected_warnings': [
 340             'General metadata extraction failed',
 341             'Main webpage is locked behind the login page',
 342         ],
 343     }, {
 344         # IGTV
 345         'url': 'https://www.instagram.com/tv/BkfuX9UB-eK/',
 346         'info_dict': {
 347             'id': 'BkfuX9UB-eK',
 348             'ext': 'mp4',
 349             'title': 'Fingerboarding Tricks with @cass.fb',
 350             'thumbnail': r're:^https?://.*\.jpg',
 351             'duration': 53.83,
 352             'timestamp': 1530032919,
 353             'upload_date': '20180626',
 354             'uploader_id': '25025320',
 355             'uploader': 'Instagram',
 356             'channel': 'instagram',
 357             'like_count': int,
 358             'comment_count': int,
 359             'comments': list,
 360             'description': 'Meet Cass Hirst (@cass.fb), a fingerboarding pro who can perform tiny ollies and kickflips while blindfolded.',
 361         },
 362         'expected_warnings': [
 363             'General metadata extraction failed',
 364             'Main webpage is locked behind the login page',
 365         ],
 366     }, {
 367         'url': 'https://instagram.com/p/-Cmh1cukG2/',
 368         'only_matching': True,
 369     }, {
 370         'url': 'http://instagram.com/p/9o6LshA7zy/embed/',
 371         'only_matching': True,
 372     }, {
 373         'url': 'https://www.instagram.com/tv/aye83DjauH/',
 374         'only_matching': True,
 375     }, {
 376         'url': 'https://www.instagram.com/reel/CDUMkliABpa/',
 377         'only_matching': True,
 378     }, {
 379         'url': 'https://www.instagram.com/marvelskies.fc/reel/CWqAgUZgCku/',
 380         'only_matching': True,
 381     }, {
 382         'url': 'https://www.instagram.com/reels/Cop84x6u7CP/',
 383         'only_matching': True,
 384     }]
 385
 386     @classmethod
 387     def _extract_embed_urls(cls, url, webpage):
 388         res = tuple(super()._extract_embed_urls(url, webpage))
 389         if res:
 390             return res
 391
 392         mobj = re.search(r'<a[^>]+href=([\'"])(?P<link>[^\'"]+)\1',
 393                          get_element_by_attribute('class', 'instagram-media', webpage) or '')
 394         if mobj:
 395             return [mobj.group('link')]
 396
 397     def _real_extract(self, url):
 398         video_id, url = self._match_valid_url(url).group('id', 'url')
 399         media, webpage = {}, ''
 400
 401         if self._get_cookies(url).get('sessionid'):
 402             info = traverse_obj(self._download_json(
 403                 f'{self._API_BASE_URL}/media/{_id_to_pk(video_id)}/info/', video_id,
 404                 fatal=False, errnote='Video info extraction failed',
 405                 note='Downloading video info', headers=self._API_HEADERS), ('items', 0))
 406             if info:
 407                 media.update(info)
 408                 return self._extract_product(media)
 409
 410         api_check = self._download_json(
 411             f'{self._API_BASE_URL}/web/get_ruling_for_content/?content_type=MEDIA&target_id={_id_to_pk(video_id)}',
 412             video_id, headers=self._API_HEADERS, fatal=False, note='Setting up session', errnote=False) or {}
 413         csrf_token = self._get_cookies('https://www.instagram.com').get('csrftoken')
 414
 415         if not csrf_token:
 416             self.report_warning('No csrf token set by Instagram API', video_id)
 417         else:
 418             csrf_token = csrf_token.value if api_check.get('status') == 'ok' else None
 419             if not csrf_token:
 420                 self.report_warning('Instagram API is not granting access', video_id)
 421
 422         variables = {
 423             'shortcode': video_id,
 424             'child_comment_count': 3,
 425             'fetch_comment_count': 40,
 426             'parent_comment_count': 24,
 427             'has_threaded_comments': True,
 428         }
 429         general_info = self._download_json(
 430             'https://www.instagram.com/graphql/query/', video_id, fatal=False, errnote=False,
 431             headers={
 432                 **self._API_HEADERS,
 433                 'X-CSRFToken': csrf_token or '',
 434                 'X-Requested-With': 'XMLHttpRequest',
 435                 'Referer': url,
 436             }, query={
 437                 'doc_id': '8845758582119845',
 438                 'variables': json.dumps(variables, separators=(',', ':')),
 439             })
 440         media.update(traverse_obj(general_info, ('data', 'xdt_shortcode_media')) or {})
 441
 442         if not general_info:
 443             self.report_warning('General metadata extraction failed (some metadata might be missing).', video_id)
 444             webpage, urlh = self._download_webpage_handle(url, video_id)
 445             shared_data = self._search_json(
 446                 r'window\._sharedData\s*=', webpage, 'shared data', video_id, fatal=False) or {}
 447
 448             if shared_data and self._LOGIN_URL not in urlh.url:
 449                 media.update(traverse_obj(
 450                     shared_data, ('entry_data', 'PostPage', 0, 'graphql', 'shortcode_media'),
 451                     ('entry_data', 'PostPage', 0, 'media'), expected_type=dict) or {})
 452             else:
 453                 self.report_warning('Main webpage is locked behind the login page. Retrying with embed webpage (some metadata might be missing).')
 454                 webpage = self._download_webpage(
 455                     f'{url}/embed/', video_id, note='Downloading embed webpage', fatal=False) or ''
 456                 additional_data = self._search_json(
 457                     r'window\.__additionalDataLoaded\s*\(\s*[^,]+,', webpage, 'additional data', video_id, fatal=False)
 458                 if not additional_data and not media:
 459                     self.raise_login_required('Requested content is not available, rate-limit reached or login required')
 460
 461                 product_item = traverse_obj(additional_data, ('items', 0), expected_type=dict)
 462                 if product_item:
 463                     media.update(product_item)
 464                     return self._extract_product(media)
 465
 466                 media.update(traverse_obj(
 467                     additional_data, ('graphql', 'shortcode_media'), 'shortcode_media', expected_type=dict) or {})
 468
 469         username = traverse_obj(media, ('owner', 'username')) or self._search_regex(
 470             r'"owner"\s*:\s*{\s*"username"\s*:\s*"(.+?)"', webpage, 'username', fatal=False)
 471
 472         description = (
 473             traverse_obj(media, ('edge_media_to_caption', 'edges', 0, 'node', 'text'), expected_type=str)
 474             or media.get('caption'))
 475         if not description:
 476             description = self._search_regex(
 477                 r'"caption"\s*:\s*"(.+?)"', webpage, 'description', default=None)
 478             if description is not None:
 479                 description = lowercase_escape(description)
 480
 481         video_url = media.get('video_url')
 482         if not video_url:
 483             nodes = traverse_obj(media, ('edge_sidecar_to_children', 'edges', ..., 'node'), expected_type=dict) or []
 484             if nodes:
 485                 return self.playlist_result(
 486                     self._extract_nodes(nodes, True), video_id,
 487                     format_field(username, None, 'Post by %s'), description)
 488
 489             video_url = self._og_search_video_url(webpage, secure=False)
 490
 491         formats = [{
 492             'url': video_url,
 493             'width': self._get_dimension('width', media, webpage),
 494             'height': self._get_dimension('height', media, webpage),
 495         }]
 496         dash = traverse_obj(media, ('dash_info', 'video_dash_manifest'))
 497         if dash:
 498             formats.extend(self._parse_mpd_formats(self._parse_xml(dash, video_id), mpd_id='dash'))
 499
 500         comment_data = traverse_obj(media, ('edge_media_to_parent_comment', 'edges'))
 501         comments = [{
 502             'author': traverse_obj(comment_dict, ('node', 'owner', 'username')),
 503             'author_id': traverse_obj(comment_dict, ('node', 'owner', 'id')),
 504             'id': traverse_obj(comment_dict, ('node', 'id')),
 505             'text': traverse_obj(comment_dict, ('node', 'text')),
 506             'timestamp': traverse_obj(comment_dict, ('node', 'created_at'), expected_type=int_or_none),
 507         } for comment_dict in comment_data] if comment_data else None
 508
 509         display_resources = (
 510             media.get('display_resources')
 511             or [{'src': media.get(key)} for key in ('display_src', 'display_url')]
 512             or [{'src': self._og_search_thumbnail(webpage)}])
 513         thumbnails = [{
 514             'url': thumbnail['src'],
 515             'width': thumbnail.get('config_width'),
 516             'height': thumbnail.get('config_height'),
 517         } for thumbnail in display_resources if thumbnail.get('src')]
 518
 519         return {
 520             'id': video_id,
 521             'formats': formats,
 522             'title': media.get('title') or f'Video by {username}',
 523             'description': description,
 524             'duration': float_or_none(media.get('video_duration')),
 525             'timestamp': traverse_obj(media, 'taken_at_timestamp', 'date', expected_type=int_or_none),
 526             'uploader_id': traverse_obj(media, ('owner', 'id')),
 527             'uploader': traverse_obj(media, ('owner', 'full_name')),
 528             'channel': username,
 529             'like_count': self._get_count(media, 'likes', 'preview_like') or str_to_int(self._search_regex(
 530                 r'data-log-event="likeCountClick"[^>]*>[^\d]*([\d,\.]+)', webpage, 'like count', fatal=False)),
 531             'comment_count': self._get_count(media, 'comments', 'preview_comment', 'to_comment', 'to_parent_comment'),
 532             'comments': comments,
 533             'thumbnails': thumbnails,
 534             'http_headers': {
 535                 'Referer': 'https://www.instagram.com/',
 536             },
 537         }
 538
 539
 540 class InstagramPlaylistBaseIE(InstagramBaseIE):
 541     _gis_tmpl = None  # used to cache GIS request type
 542
 543     def _parse_graphql(self, webpage, item_id):
 544         # Reads a webpage and returns its GraphQL data.
 545         return self._parse_json(
 546             self._search_regex(
 547                 r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
 548             item_id)
 549
 550     def _extract_graphql(self, data, url):
 551         # Parses GraphQL queries containing videos and generates a playlist.
 552         uploader_id = self._match_id(url)
 553         csrf_token = data['config']['csrf_token']
 554         rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
 555
 556         cursor = ''
 557         for page_num in itertools.count(1):
 558             variables = {
 559                 'first': 12,
 560                 'after': cursor,
 561             }
 562             variables.update(self._query_vars_for(data))
 563             variables = json.dumps(variables)
 564
 565             if self._gis_tmpl:
 566                 gis_tmpls = [self._gis_tmpl]
 567             else:
 568                 gis_tmpls = [
 569                     f'{rhx_gis}',
 570                     '',
 571                     f'{rhx_gis}:{csrf_token}',
 572                     '{}:{}:{}'.format(rhx_gis, csrf_token, self.get_param('http_headers')['User-Agent']),
 573                 ]
 574
 575             # try all of the ways to generate a GIS query, and not only use the
 576             # first one that works, but cache it for future requests
 577             for gis_tmpl in gis_tmpls:
 578                 try:
 579                     json_data = self._download_json(
 580                         'https://www.instagram.com/graphql/query/', uploader_id,
 581                         f'Downloading JSON page {page_num}', headers={
 582                             'X-Requested-With': 'XMLHttpRequest',
 583                             'X-Instagram-GIS': hashlib.md5(
 584                                 (f'{gis_tmpl}:{variables}').encode()).hexdigest(),
 585                         }, query={
 586                             'query_hash': self._QUERY_HASH,
 587                             'variables': variables,
 588                         })
 589                     media = self._parse_timeline_from(json_data)
 590                     self._gis_tmpl = gis_tmpl
 591                     break
 592                 except ExtractorError as e:
 593                     # if it's an error caused by a bad query, and there are
 594                     # more GIS templates to try, ignore it and keep trying
 595                     if isinstance(e.cause, HTTPError) and e.cause.status == 403:
 596                         if gis_tmpl != gis_tmpls[-1]:
 597                             continue
 598                     raise
 599
 600             nodes = traverse_obj(media, ('edges', ..., 'node'), expected_type=dict) or []
 601             if not nodes:
 602                 break
 603             yield from self._extract_nodes(nodes)
 604
 605             has_next_page = traverse_obj(media, ('page_info', 'has_next_page'))
 606             cursor = traverse_obj(media, ('page_info', 'end_cursor'), expected_type=str)
 607             if not has_next_page or not cursor:
 608                 break
 609
 610     def _real_extract(self, url):
 611         user_or_tag = self._match_id(url)
 612         webpage = self._download_webpage(url, user_or_tag)
 613         data = self._parse_graphql(webpage, user_or_tag)
 614
 615         self._set_cookie('instagram.com', 'ig_pr', '1')
 616
 617         return self.playlist_result(
 618             self._extract_graphql(data, url), user_or_tag, user_or_tag)
 619
 620
 621 class InstagramUserIE(InstagramPlaylistBaseIE):
 622     _WORKING = False
 623     _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
 624     IE_DESC = 'Instagram user profile'
 625     IE_NAME = 'instagram:user'
 626     _TESTS = [{
 627         'url': 'https://instagram.com/porsche',
 628         'info_dict': {
 629             'id': 'porsche',
 630             'title': 'porsche',
 631         },
 632         'playlist_count': 5,
 633         'params': {
 634             'extract_flat': True,
 635             'skip_download': True,
 636             'playlistend': 5,
 637         },
 638     }]
 639
 640     _QUERY_HASH = ('42323d64886122307be10013ad2dcc44',)
 641
 642     @staticmethod
 643     def _parse_timeline_from(data):
 644         # extracts the media timeline data from a GraphQL result
 645         return data['data']['user']['edge_owner_to_timeline_media']
 646
 647     @staticmethod
 648     def _query_vars_for(data):
 649         # returns a dictionary of variables to add to the timeline query based
 650         # on the GraphQL of the original page
 651         return {
 652             'id': data['entry_data']['ProfilePage'][0]['graphql']['user']['id'],
 653         }
 654
 655
 656 class InstagramTagIE(InstagramPlaylistBaseIE):
 657     _VALID_URL = r'https?://(?:www\.)?instagram\.com/explore/tags/(?P<id>[^/]+)'
 658     IE_DESC = 'Instagram hashtag search URLs'
 659     IE_NAME = 'instagram:tag'
 660     _TESTS = [{
 661         'url': 'https://instagram.com/explore/tags/lolcats',
 662         'info_dict': {
 663             'id': 'lolcats',
 664             'title': 'lolcats',
 665         },
 666         'playlist_count': 50,
 667         'params': {
 668             'extract_flat': True,
 669             'skip_download': True,
 670             'playlistend': 50,
 671         },
 672     }]
 673
 674     _QUERY_HASH = ('f92f56d47dc7a55b606908374b43a314',)
 675
 676     @staticmethod
 677     def _parse_timeline_from(data):
 678         # extracts the media timeline data from a GraphQL result
 679         return data['data']['hashtag']['edge_hashtag_to_media']
 680
 681     @staticmethod
 682     def _query_vars_for(data):
 683         # returns a dictionary of variables to add to the timeline query based
 684         # on the GraphQL of the original page
 685         return {
 686             'tag_name':
 687                 data['entry_data']['TagPage'][0]['graphql']['hashtag']['name'],
 688         }
 689
 690
 691 class InstagramStoryIE(InstagramBaseIE):
 692     _VALID_URL = r'https?://(?:www\.)?instagram\.com/stories/(?P<user>[^/]+)/(?P<id>\d+)'
 693     IE_NAME = 'instagram:story'
 694
 695     _TESTS = [{
 696         'url': 'https://www.instagram.com/stories/highlights/18090946048123978/',
 697         'info_dict': {
 698             'id': '18090946048123978',
 699             'title': 'Rare',
 700         },
 701         'playlist_mincount': 50,
 702     }]
 703
 704     def _real_extract(self, url):
 705         username, story_id = self._match_valid_url(url).groups()
 706         story_info = self._download_webpage(url, story_id)
 707         user_info = self._search_json(r'"user":', story_info, 'user info', story_id, fatal=False)
 708         if not user_info:
 709             self.raise_login_required('This content is unreachable')
 710
 711         user_id = traverse_obj(user_info, 'pk', 'id', expected_type=str)
 712         story_info_url = user_id if username != 'highlights' else f'highlight:{story_id}'
 713         if not story_info_url:  # user id is only mandatory for non-highlights
 714             raise ExtractorError('Unable to extract user id')
 715
 716         videos = traverse_obj(self._download_json(
 717             f'{self._API_BASE_URL}/feed/reels_media/?reel_ids={story_info_url}',
 718             story_id, errnote=False, fatal=False, headers=self._API_HEADERS), 'reels')
 719         if not videos:
 720             self.raise_login_required('You need to log in to access this content')
 721
 722         full_name = traverse_obj(videos, (f'highlight:{story_id}', 'user', 'full_name'), (user_id, 'user', 'full_name'))
 723         story_title = traverse_obj(videos, (f'highlight:{story_id}', 'title'))
 724         if not story_title:
 725             story_title = f'Story by {username}'
 726
 727         highlights = traverse_obj(videos, (f'highlight:{story_id}', 'items'), (user_id, 'items'))
 728         info_data = []
 729         for highlight in highlights:
 730             highlight_data = self._extract_product(highlight)
 731             if highlight_data.get('formats'):
 732                 info_data.append({
 733                     'uploader': full_name,
 734                     'uploader_id': user_id,
 735                     **filter_dict(highlight_data),
 736                 })
 737         return self.playlist_result(info_data, playlist_id=story_id, playlist_title=story_title)