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