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

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