_ Git - youtube-dl/blob - youtube_dl/extractor/instagram.py

   1 from __future__ import unicode_literals
   2
   3 import itertools
   4 import hashlib
   5 import json
   6 import re
   7
   8 from .common import InfoExtractor
   9 from ..compat import compat_str
  10 from ..utils import (
  11     get_element_by_attribute,
  12     int_or_none,
  13     lowercase_escape,
  14     try_get,
  15 )
  16
  17
  18 class InstagramIE(InfoExtractor):
  19     _VALID_URL = r'(?P<url>https?://(?:www\.)?instagram\.com/p/(?P<id>[^/?#&]+))'
  20     _TESTS = [{
  21         'url': 'https://instagram.com/p/aye83DjauH/?foo=bar#abc',
  22         'md5': '0d2da106a9d2631273e192b372806516',
  23         'info_dict': {
  24             'id': 'aye83DjauH',
  25             'ext': 'mp4',
  26             'title': 'Video by naomipq',
  27             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
  28             'thumbnail': r're:^https?://.*\.jpg',
  29             'timestamp': 1371748545,
  30             'upload_date': '20130620',
  31             'uploader_id': 'naomipq',
  32             'uploader': 'Naomi Leonor Phan-Quang',
  33             'like_count': int,
  34             'comment_count': int,
  35             'comments': list,
  36         },
  37     }, {
  38         # missing description
  39         'url': 'https://www.instagram.com/p/BA-pQFBG8HZ/?taken-by=britneyspears',
  40         'info_dict': {
  41             'id': 'BA-pQFBG8HZ',
  42             'ext': 'mp4',
  43             'title': 'Video by britneyspears',
  44             'thumbnail': r're:^https?://.*\.jpg',
  45             'timestamp': 1453760977,
  46             'upload_date': '20160125',
  47             'uploader_id': 'britneyspears',
  48             'uploader': 'Britney Spears',
  49             'like_count': int,
  50             'comment_count': int,
  51             'comments': list,
  52         },
  53         'params': {
  54             'skip_download': True,
  55         },
  56     }, {
  57         # multi video post
  58         'url': 'https://www.instagram.com/p/BQ0eAlwhDrw/',
  59         'playlist': [{
  60             'info_dict': {
  61                 'id': 'BQ0dSaohpPW',
  62                 'ext': 'mp4',
  63                 'title': 'Video 1',
  64             },
  65         }, {
  66             'info_dict': {
  67                 'id': 'BQ0dTpOhuHT',
  68                 'ext': 'mp4',
  69                 'title': 'Video 2',
  70             },
  71         }, {
  72             'info_dict': {
  73                 'id': 'BQ0dT7RBFeF',
  74                 'ext': 'mp4',
  75                 'title': 'Video 3',
  76             },
  77         }],
  78         'info_dict': {
  79             'id': 'BQ0eAlwhDrw',
  80             'title': 'Post by instagram',
  81             'description': 'md5:0f9203fc6a2ce4d228da5754bcf54957',
  82         },
  83     }, {
  84         'url': 'https://instagram.com/p/-Cmh1cukG2/',
  85         'only_matching': True,
  86     }, {
  87         'url': 'http://instagram.com/p/9o6LshA7zy/embed/',
  88         'only_matching': True,
  89     }]
  90
  91     @staticmethod
  92     def _extract_embed_url(webpage):
  93         mobj = re.search(
  94             r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?instagram\.com/p/[^/]+/embed.*?)\1',
  95             webpage)
  96         if mobj:
  97             return mobj.group('url')
  98
  99         blockquote_el = get_element_by_attribute(
 100             'class', 'instagram-media', webpage)
 101         if blockquote_el is None:
 102             return
 103
 104         mobj = re.search(
 105             r'<a[^>]+href=([\'"])(?P<link>[^\'"]+)\1', blockquote_el)
 106         if mobj:
 107             return mobj.group('link')
 108
 109     def _real_extract(self, url):
 110         mobj = re.match(self._VALID_URL, url)
 111         video_id = mobj.group('id')
 112         url = mobj.group('url')
 113
 114         webpage = self._download_webpage(url, video_id)
 115
 116         (video_url, description, thumbnail, timestamp, uploader,
 117          uploader_id, like_count, comment_count, comments, height,
 118          width) = [None] * 11
 119
 120         shared_data = self._parse_json(
 121             self._search_regex(
 122                 r'window\._sharedData\s*=\s*({.+?});',
 123                 webpage, 'shared data', default='{}'),
 124             video_id, fatal=False)
 125         if shared_data:
 126             media = try_get(
 127                 shared_data,
 128                 (lambda x: x['entry_data']['PostPage'][0]['graphql']['shortcode_media'],
 129                  lambda x: x['entry_data']['PostPage'][0]['media']),
 130                 dict)
 131             if media:
 132                 video_url = media.get('video_url')
 133                 height = int_or_none(media.get('dimensions', {}).get('height'))
 134                 width = int_or_none(media.get('dimensions', {}).get('width'))
 135                 description = try_get(
 136                     media, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
 137                     compat_str) or media.get('caption')
 138                 thumbnail = media.get('display_src')
 139                 timestamp = int_or_none(media.get('taken_at_timestamp') or media.get('date'))
 140                 uploader = media.get('owner', {}).get('full_name')
 141                 uploader_id = media.get('owner', {}).get('username')
 142
 143                 def get_count(key, kind):
 144                     return int_or_none(try_get(
 145                         media, (lambda x: x['edge_media_%s' % key]['count'],
 146                                 lambda x: x['%ss' % kind]['count'])))
 147                 like_count = get_count('preview_like', 'like')
 148                 comment_count = get_count('to_comment', 'comment')
 149
 150                 comments = [{
 151                     'author': comment.get('user', {}).get('username'),
 152                     'author_id': comment.get('user', {}).get('id'),
 153                     'id': comment.get('id'),
 154                     'text': comment.get('text'),
 155                     'timestamp': int_or_none(comment.get('created_at')),
 156                 } for comment in media.get(
 157                     'comments', {}).get('nodes', []) if comment.get('text')]
 158                 if not video_url:
 159                     edges = try_get(
 160                         media, lambda x: x['edge_sidecar_to_children']['edges'],
 161                         list) or []
 162                     if edges:
 163                         entries = []
 164                         for edge_num, edge in enumerate(edges, start=1):
 165                             node = try_get(edge, lambda x: x['node'], dict)
 166                             if not node:
 167                                 continue
 168                             node_video_url = try_get(node, lambda x: x['video_url'], compat_str)
 169                             if not node_video_url:
 170                                 continue
 171                             entries.append({
 172                                 'id': node.get('shortcode') or node['id'],
 173                                 'title': 'Video %d' % edge_num,
 174                                 'url': node_video_url,
 175                                 'thumbnail': node.get('display_url'),
 176                                 'width': int_or_none(try_get(node, lambda x: x['dimensions']['width'])),
 177                                 'height': int_or_none(try_get(node, lambda x: x['dimensions']['height'])),
 178                                 'view_count': int_or_none(node.get('video_view_count')),
 179                             })
 180                         return self.playlist_result(
 181                             entries, video_id,
 182                             'Post by %s' % uploader_id if uploader_id else None,
 183                             description)
 184
 185         if not video_url:
 186             video_url = self._og_search_video_url(webpage, secure=False)
 187
 188         formats = [{
 189             'url': video_url,
 190             'width': width,
 191             'height': height,
 192         }]
 193
 194         if not uploader_id:
 195             uploader_id = self._search_regex(
 196                 r'"owner"\s*:\s*{\s*"username"\s*:\s*"(.+?)"',
 197                 webpage, 'uploader id', fatal=False)
 198
 199         if not description:
 200             description = self._search_regex(
 201                 r'"caption"\s*:\s*"(.+?)"', webpage, 'description', default=None)
 202             if description is not None:
 203                 description = lowercase_escape(description)
 204
 205         if not thumbnail:
 206             thumbnail = self._og_search_thumbnail(webpage)
 207
 208         return {
 209             'id': video_id,
 210             'formats': formats,
 211             'ext': 'mp4',
 212             'title': 'Video by %s' % uploader_id,
 213             'description': description,
 214             'thumbnail': thumbnail,
 215             'timestamp': timestamp,
 216             'uploader_id': uploader_id,
 217             'uploader': uploader,
 218             'like_count': like_count,
 219             'comment_count': comment_count,
 220             'comments': comments,
 221         }
 222
 223
 224 class InstagramUserIE(InfoExtractor):
 225     _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
 226     IE_DESC = 'Instagram user profile'
 227     IE_NAME = 'instagram:user'
 228     _TEST = {
 229         'url': 'https://instagram.com/porsche',
 230         'info_dict': {
 231             'id': 'porsche',
 232             'title': 'porsche',
 233         },
 234         'playlist_count': 5,
 235         'params': {
 236             'extract_flat': True,
 237             'skip_download': True,
 238             'playlistend': 5,
 239         }
 240     }
 241
 242     def _entries(self, data):
 243         def get_count(suffix):
 244             return int_or_none(try_get(
 245                 node, lambda x: x['edge_media_' + suffix]['count']))
 246
 247         uploader_id = data['entry_data']['ProfilePage'][0]['graphql']['user']['id']
 248         csrf_token = data['config']['csrf_token']
 249         rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
 250
 251         self._set_cookie('instagram.com', 'ig_pr', '1')
 252
 253         cursor = ''
 254         for page_num in itertools.count(1):
 255             variables = json.dumps({
 256                 'id': uploader_id,
 257                 'first': 100,
 258                 'after': cursor,
 259             })
 260             s = '%s:%s:%s' % (rhx_gis, csrf_token, variables)
 261             media = self._download_json(
 262                 'https://www.instagram.com/graphql/query/', uploader_id,
 263                 'Downloading JSON page %d' % page_num, headers={
 264                     'X-Requested-With': 'XMLHttpRequest',
 265                     'X-Instagram-GIS': hashlib.md5(s.encode('utf-8')).hexdigest(),
 266                 }, query={
 267                     'query_hash': '472f257a40c653c64c666ce877d59d2b',
 268                     'variables': variables,
 269                 })['data']['user']['edge_owner_to_timeline_media']
 270
 271             edges = media.get('edges')
 272             if not edges or not isinstance(edges, list):
 273                 break
 274
 275             for edge in edges:
 276                 node = edge.get('node')
 277                 if not node or not isinstance(node, dict):
 278                     continue
 279                 if node.get('__typename') != 'GraphVideo' and node.get('is_video') is not True:
 280                     continue
 281                 video_id = node.get('shortcode')
 282                 if not video_id:
 283                     continue
 284
 285                 info = self.url_result(
 286                     'https://instagram.com/p/%s/' % video_id,
 287                     ie=InstagramIE.ie_key(), video_id=video_id)
 288
 289                 description = try_get(
 290                     node, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
 291                     compat_str)
 292                 thumbnail = node.get('thumbnail_src') or node.get('display_src')
 293                 timestamp = int_or_none(node.get('taken_at_timestamp'))
 294
 295                 comment_count = get_count('to_comment')
 296                 like_count = get_count('preview_like')
 297                 view_count = int_or_none(node.get('video_view_count'))
 298
 299                 info.update({
 300                     'description': description,
 301                     'thumbnail': thumbnail,
 302                     'timestamp': timestamp,
 303                     'comment_count': comment_count,
 304                     'like_count': like_count,
 305                     'view_count': view_count,
 306                 })
 307
 308                 yield info
 309
 310             page_info = media.get('page_info')
 311             if not page_info or not isinstance(page_info, dict):
 312                 break
 313
 314             has_next_page = page_info.get('has_next_page')
 315             if not has_next_page:
 316                 break
 317
 318             cursor = page_info.get('end_cursor')
 319             if not cursor or not isinstance(cursor, compat_str):
 320                 break
 321
 322     def _real_extract(self, url):
 323         username = self._match_id(url)
 324
 325         webpage = self._download_webpage(url, username)
 326
 327         data = self._parse_json(
 328             self._search_regex(
 329                 r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
 330             username)
 331
 332         return self.playlist_result(
 333             self._entries(data), username, username)