_ 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 (
  10     compat_str,
  11     compat_HTTPError,
  12 )
  13 from ..utils import (
  14     ExtractorError,
  15     get_element_by_attribute,
  16     int_or_none,
  17     lowercase_escape,
  18     std_headers,
  19     try_get,
  20 )
  21
  22
  23 class InstagramIE(InfoExtractor):
  24     _VALID_URL = r'(?P<url>https?://(?:www\.)?instagram\.com/p/(?P<id>[^/?#&]+))'
  25     _TESTS = [{
  26         'url': 'https://instagram.com/p/aye83DjauH/?foo=bar#abc',
  27         'md5': '0d2da106a9d2631273e192b372806516',
  28         'info_dict': {
  29             'id': 'aye83DjauH',
  30             'ext': 'mp4',
  31             'title': 'Video by naomipq',
  32             'description': 'md5:1f17f0ab29bd6fe2bfad705f58de3cb8',
  33             'thumbnail': r're:^https?://.*\.jpg',
  34             'timestamp': 1371748545,
  35             'upload_date': '20130620',
  36             'uploader_id': 'naomipq',
  37             'uploader': 'Naomi Leonor Phan-Quang',
  38             'like_count': int,
  39             'comment_count': int,
  40             'comments': list,
  41         },
  42     }, {
  43         # missing description
  44         'url': 'https://www.instagram.com/p/BA-pQFBG8HZ/?taken-by=britneyspears',
  45         'info_dict': {
  46             'id': 'BA-pQFBG8HZ',
  47             'ext': 'mp4',
  48             'title': 'Video by britneyspears',
  49             'thumbnail': r're:^https?://.*\.jpg',
  50             'timestamp': 1453760977,
  51             'upload_date': '20160125',
  52             'uploader_id': 'britneyspears',
  53             'uploader': 'Britney Spears',
  54             'like_count': int,
  55             'comment_count': int,
  56             'comments': list,
  57         },
  58         'params': {
  59             'skip_download': True,
  60         },
  61     }, {
  62         # multi video post
  63         'url': 'https://www.instagram.com/p/BQ0eAlwhDrw/',
  64         'playlist': [{
  65             'info_dict': {
  66                 'id': 'BQ0dSaohpPW',
  67                 'ext': 'mp4',
  68                 'title': 'Video 1',
  69             },
  70         }, {
  71             'info_dict': {
  72                 'id': 'BQ0dTpOhuHT',
  73                 'ext': 'mp4',
  74                 'title': 'Video 2',
  75             },
  76         }, {
  77             'info_dict': {
  78                 'id': 'BQ0dT7RBFeF',
  79                 'ext': 'mp4',
  80                 'title': 'Video 3',
  81             },
  82         }],
  83         'info_dict': {
  84             'id': 'BQ0eAlwhDrw',
  85             'title': 'Post by instagram',
  86             'description': 'md5:0f9203fc6a2ce4d228da5754bcf54957',
  87         },
  88     }, {
  89         'url': 'https://instagram.com/p/-Cmh1cukG2/',
  90         'only_matching': True,
  91     }, {
  92         'url': 'http://instagram.com/p/9o6LshA7zy/embed/',
  93         'only_matching': True,
  94     }]
  95
  96     @staticmethod
  97     def _extract_embed_url(webpage):
  98         mobj = re.search(
  99             r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?instagram\.com/p/[^/]+/embed.*?)\1',
 100             webpage)
 101         if mobj:
 102             return mobj.group('url')
 103
 104         blockquote_el = get_element_by_attribute(
 105             'class', 'instagram-media', webpage)
 106         if blockquote_el is None:
 107             return
 108
 109         mobj = re.search(
 110             r'<a[^>]+href=([\'"])(?P<link>[^\'"]+)\1', blockquote_el)
 111         if mobj:
 112             return mobj.group('link')
 113
 114     def _real_extract(self, url):
 115         mobj = re.match(self._VALID_URL, url)
 116         video_id = mobj.group('id')
 117         url = mobj.group('url')
 118
 119         webpage = self._download_webpage(url, video_id)
 120
 121         (video_url, description, thumbnail, timestamp, uploader,
 122          uploader_id, like_count, comment_count, comments, height,
 123          width) = [None] * 11
 124
 125         shared_data = self._parse_json(
 126             self._search_regex(
 127                 r'window\._sharedData\s*=\s*({.+?});',
 128                 webpage, 'shared data', default='{}'),
 129             video_id, fatal=False)
 130         if shared_data:
 131             media = try_get(
 132                 shared_data,
 133                 (lambda x: x['entry_data']['PostPage'][0]['graphql']['shortcode_media'],
 134                  lambda x: x['entry_data']['PostPage'][0]['media']),
 135                 dict)
 136             if media:
 137                 video_url = media.get('video_url')
 138                 height = int_or_none(media.get('dimensions', {}).get('height'))
 139                 width = int_or_none(media.get('dimensions', {}).get('width'))
 140                 description = try_get(
 141                     media, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
 142                     compat_str) or media.get('caption')
 143                 thumbnail = media.get('display_src')
 144                 timestamp = int_or_none(media.get('taken_at_timestamp') or media.get('date'))
 145                 uploader = media.get('owner', {}).get('full_name')
 146                 uploader_id = media.get('owner', {}).get('username')
 147
 148                 def get_count(key, kind):
 149                     return int_or_none(try_get(
 150                         media, (lambda x: x['edge_media_%s' % key]['count'],
 151                                 lambda x: x['%ss' % kind]['count'])))
 152                 like_count = get_count('preview_like', 'like')
 153                 comment_count = get_count('to_comment', 'comment')
 154
 155                 comments = [{
 156                     'author': comment.get('user', {}).get('username'),
 157                     'author_id': comment.get('user', {}).get('id'),
 158                     'id': comment.get('id'),
 159                     'text': comment.get('text'),
 160                     'timestamp': int_or_none(comment.get('created_at')),
 161                 } for comment in media.get(
 162                     'comments', {}).get('nodes', []) if comment.get('text')]
 163                 if not video_url:
 164                     edges = try_get(
 165                         media, lambda x: x['edge_sidecar_to_children']['edges'],
 166                         list) or []
 167                     if edges:
 168                         entries = []
 169                         for edge_num, edge in enumerate(edges, start=1):
 170                             node = try_get(edge, lambda x: x['node'], dict)
 171                             if not node:
 172                                 continue
 173                             node_video_url = try_get(node, lambda x: x['video_url'], compat_str)
 174                             if not node_video_url:
 175                                 continue
 176                             entries.append({
 177                                 'id': node.get('shortcode') or node['id'],
 178                                 'title': 'Video %d' % edge_num,
 179                                 'url': node_video_url,
 180                                 'thumbnail': node.get('display_url'),
 181                                 'width': int_or_none(try_get(node, lambda x: x['dimensions']['width'])),
 182                                 'height': int_or_none(try_get(node, lambda x: x['dimensions']['height'])),
 183                                 'view_count': int_or_none(node.get('video_view_count')),
 184                             })
 185                         return self.playlist_result(
 186                             entries, video_id,
 187                             'Post by %s' % uploader_id if uploader_id else None,
 188                             description)
 189
 190         if not video_url:
 191             video_url = self._og_search_video_url(webpage, secure=False)
 192
 193         formats = [{
 194             'url': video_url,
 195             'width': width,
 196             'height': height,
 197         }]
 198
 199         if not uploader_id:
 200             uploader_id = self._search_regex(
 201                 r'"owner"\s*:\s*{\s*"username"\s*:\s*"(.+?)"',
 202                 webpage, 'uploader id', fatal=False)
 203
 204         if not description:
 205             description = self._search_regex(
 206                 r'"caption"\s*:\s*"(.+?)"', webpage, 'description', default=None)
 207             if description is not None:
 208                 description = lowercase_escape(description)
 209
 210         if not thumbnail:
 211             thumbnail = self._og_search_thumbnail(webpage)
 212
 213         return {
 214             'id': video_id,
 215             'formats': formats,
 216             'ext': 'mp4',
 217             'title': 'Video by %s' % uploader_id,
 218             'description': description,
 219             'thumbnail': thumbnail,
 220             'timestamp': timestamp,
 221             'uploader_id': uploader_id,
 222             'uploader': uploader,
 223             'like_count': like_count,
 224             'comment_count': comment_count,
 225             'comments': comments,
 226         }
 227
 228
 229 class InstagramUserIE(InfoExtractor):
 230     _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
 231     IE_DESC = 'Instagram user profile'
 232     IE_NAME = 'instagram:user'
 233     _TEST = {
 234         'url': 'https://instagram.com/porsche',
 235         'info_dict': {
 236             'id': 'porsche',
 237             'title': 'porsche',
 238         },
 239         'playlist_count': 5,
 240         'params': {
 241             'extract_flat': True,
 242             'skip_download': True,
 243             'playlistend': 5,
 244         }
 245     }
 246
 247     _gis_tmpl = None
 248
 249     def _entries(self, data):
 250         def get_count(suffix):
 251             return int_or_none(try_get(
 252                 node, lambda x: x['edge_media_' + suffix]['count']))
 253
 254         uploader_id = data['entry_data']['ProfilePage'][0]['graphql']['user']['id']
 255         csrf_token = data['config']['csrf_token']
 256         rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
 257
 258         self._set_cookie('instagram.com', 'ig_pr', '1')
 259
 260         cursor = ''
 261         for page_num in itertools.count(1):
 262             variables = json.dumps({
 263                 'id': uploader_id,
 264                 'first': 12,
 265                 'after': cursor,
 266             })
 267
 268             if self._gis_tmpl:
 269                 gis_tmpls = [self._gis_tmpl]
 270             else:
 271                 gis_tmpls = [
 272                     '%s' % rhx_gis,
 273                     '',
 274                     '%s:%s' % (rhx_gis, csrf_token),
 275                     '%s:%s:%s' % (rhx_gis, csrf_token, std_headers['User-Agent']),
 276                 ]
 277
 278             for gis_tmpl in gis_tmpls:
 279                 try:
 280                     media = self._download_json(
 281                         'https://www.instagram.com/graphql/query/', uploader_id,
 282                         'Downloading JSON page %d' % page_num, headers={
 283                             'X-Requested-With': 'XMLHttpRequest',
 284                             'X-Instagram-GIS': hashlib.md5(
 285                                 ('%s:%s' % (gis_tmpl, variables)).encode('utf-8')).hexdigest(),
 286                         }, query={
 287                             'query_hash': '42323d64886122307be10013ad2dcc44',
 288                             'variables': variables,
 289                         })['data']['user']['edge_owner_to_timeline_media']
 290                     self._gis_tmpl = gis_tmpl
 291                     break
 292                 except ExtractorError as e:
 293                     if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 294                         if gis_tmpl != gis_tmpls[-1]:
 295                             continue
 296                     raise
 297
 298             edges = media.get('edges')
 299             if not edges or not isinstance(edges, list):
 300                 break
 301
 302             for edge in edges:
 303                 node = edge.get('node')
 304                 if not node or not isinstance(node, dict):
 305                     continue
 306                 if node.get('__typename') != 'GraphVideo' and node.get('is_video') is not True:
 307                     continue
 308                 video_id = node.get('shortcode')
 309                 if not video_id:
 310                     continue
 311
 312                 info = self.url_result(
 313                     'https://instagram.com/p/%s/' % video_id,
 314                     ie=InstagramIE.ie_key(), video_id=video_id)
 315
 316                 description = try_get(
 317                     node, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
 318                     compat_str)
 319                 thumbnail = node.get('thumbnail_src') or node.get('display_src')
 320                 timestamp = int_or_none(node.get('taken_at_timestamp'))
 321
 322                 comment_count = get_count('to_comment')
 323                 like_count = get_count('preview_like')
 324                 view_count = int_or_none(node.get('video_view_count'))
 325
 326                 info.update({
 327                     'description': description,
 328                     'thumbnail': thumbnail,
 329                     'timestamp': timestamp,
 330                     'comment_count': comment_count,
 331                     'like_count': like_count,
 332                     'view_count': view_count,
 333                 })
 334
 335                 yield info
 336
 337             page_info = media.get('page_info')
 338             if not page_info or not isinstance(page_info, dict):
 339                 break
 340
 341             has_next_page = page_info.get('has_next_page')
 342             if not has_next_page:
 343                 break
 344
 345             cursor = page_info.get('end_cursor')
 346             if not cursor or not isinstance(cursor, compat_str):
 347                 break
 348
 349     def _real_extract(self, url):
 350         username = self._match_id(url)
 351
 352         webpage = self._download_webpage(url, username)
 353
 354         data = self._parse_json(
 355             self._search_regex(
 356                 r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
 357             username)
 358
 359         return self.playlist_result(
 360             self._entries(data), username, username)