yt_dlp/extractor/pinterest.py

   1 import json
   2
   3 from .common import InfoExtractor
   4 from ..compat import compat_str
   5 from ..utils import (
   6     determine_ext,
   7     float_or_none,
   8     int_or_none,
   9     try_get,
  10     unified_timestamp,
  11     url_or_none,
  12 )
  13
  14
  15 class PinterestBaseIE(InfoExtractor):
  16     _VALID_URL_BASE = r'https?://(?:[^/]+\.)?pinterest\.(?:com|fr|de|ch|jp|cl|ca|it|co\.uk|nz|ru|com\.au|at|pt|co\.kr|es|com\.mx|dk|ph|th|com\.uy|co|nl|info|kr|ie|vn|com\.vn|ec|mx|in|pe|co\.at|hu|co\.in|co\.nz|id|com\.ec|com\.py|tw|be|uk|com\.bo|com\.pe)'
  17
  18     def _call_api(self, resource, video_id, options):
  19         return self._download_json(
  20             'https://www.pinterest.com/resource/%sResource/get/' % resource,
  21             video_id, 'Download %s JSON metadata' % resource, query={
  22                 'data': json.dumps({'options': options})
  23             })['resource_response']
  24
  25     def _extract_video(self, data, extract_formats=True):
  26         video_id = data['id']
  27
  28         title = (data.get('title') or data.get('grid_title') or video_id).strip()
  29
  30         urls = []
  31         formats = []
  32         duration = None
  33         if extract_formats:
  34             for format_id, format_dict in data['videos']['video_list'].items():
  35                 if not isinstance(format_dict, dict):
  36                     continue
  37                 format_url = url_or_none(format_dict.get('url'))
  38                 if not format_url or format_url in urls:
  39                     continue
  40                 urls.append(format_url)
  41                 duration = float_or_none(format_dict.get('duration'), scale=1000)
  42                 ext = determine_ext(format_url)
  43                 if 'hls' in format_id.lower() or ext == 'm3u8':
  44                     formats.extend(self._extract_m3u8_formats(
  45                         format_url, video_id, 'mp4', entry_protocol='m3u8_native',
  46                         m3u8_id=format_id, fatal=False))
  47                 else:
  48                     formats.append({
  49                         'url': format_url,
  50                         'format_id': format_id,
  51                         'width': int_or_none(format_dict.get('width')),
  52                         'height': int_or_none(format_dict.get('height')),
  53                         'duration': duration,
  54                     })
  55             self._sort_formats(formats)
  56
  57         description = data.get('description') or data.get('description_html') or data.get('seo_description')
  58         timestamp = unified_timestamp(data.get('created_at'))
  59
  60         def _u(field):
  61             return try_get(data, lambda x: x['closeup_attribution'][field], compat_str)
  62
  63         uploader = _u('full_name')
  64         uploader_id = _u('id')
  65
  66         repost_count = int_or_none(data.get('repin_count'))
  67         comment_count = int_or_none(data.get('comment_count'))
  68         categories = try_get(data, lambda x: x['pin_join']['visual_annotation'], list)
  69         tags = data.get('hashtags')
  70
  71         thumbnails = []
  72         images = data.get('images')
  73         if isinstance(images, dict):
  74             for thumbnail_id, thumbnail in images.items():
  75                 if not isinstance(thumbnail, dict):
  76                     continue
  77                 thumbnail_url = url_or_none(thumbnail.get('url'))
  78                 if not thumbnail_url:
  79                     continue
  80                 thumbnails.append({
  81                     'url': thumbnail_url,
  82                     'width': int_or_none(thumbnail.get('width')),
  83                     'height': int_or_none(thumbnail.get('height')),
  84                 })
  85
  86         return {
  87             'id': video_id,
  88             'title': title,
  89             'description': description,
  90             'duration': duration,
  91             'timestamp': timestamp,
  92             'thumbnails': thumbnails,
  93             'uploader': uploader,
  94             'uploader_id': uploader_id,
  95             'repost_count': repost_count,
  96             'comment_count': comment_count,
  97             'categories': categories,
  98             'tags': tags,
  99             'formats': formats,
 100             'extractor_key': PinterestIE.ie_key(),
 101         }
 102
 103
 104 class PinterestIE(PinterestBaseIE):
 105     _VALID_URL = r'%s/pin/(?P<id>\d+)' % PinterestBaseIE._VALID_URL_BASE
 106     _TESTS = [{
 107         'url': 'https://www.pinterest.com/pin/664281013778109217/',
 108         'md5': '6550c2af85d6d9f3fe3b88954d1577fc',
 109         'info_dict': {
 110             'id': '664281013778109217',
 111             'ext': 'mp4',
 112             'title': 'Origami',
 113             'description': 'md5:b9d90ddf7848e897882de9e73344f7dd',
 114             'duration': 57.7,
 115             'timestamp': 1593073622,
 116             'upload_date': '20200625',
 117             'uploader': 'Love origami -I am Dafei',
 118             'uploader_id': '586523688879454212',
 119             'repost_count': 50,
 120             'comment_count': 0,
 121             'categories': list,
 122             'tags': list,
 123         },
 124     }, {
 125         'url': 'https://co.pinterest.com/pin/824721750502199491/',
 126         'only_matching': True,
 127     }]
 128
 129     def _real_extract(self, url):
 130         video_id = self._match_id(url)
 131         data = self._call_api(
 132             'Pin', video_id, {
 133                 'field_set_key': 'unauth_react_main_pin',
 134                 'id': video_id,
 135             })['data']
 136         return self._extract_video(data)
 137
 138
 139 class PinterestCollectionIE(PinterestBaseIE):
 140     _VALID_URL = r'%s/(?P<username>[^/]+)/(?P<id>[^/?#&]+)' % PinterestBaseIE._VALID_URL_BASE
 141     _TESTS = [{
 142         'url': 'https://www.pinterest.ca/mashal0407/cool-diys/',
 143         'info_dict': {
 144             'id': '585890301462791043',
 145             'title': 'cool diys',
 146         },
 147         'playlist_count': 8,
 148     }, {
 149         'url': 'https://www.pinterest.ca/fudohub/videos/',
 150         'info_dict': {
 151             'id': '682858430939307450',
 152             'title': 'VIDEOS',
 153         },
 154         'playlist_mincount': 365,
 155         'skip': 'Test with extract_formats=False',
 156     }]
 157
 158     @classmethod
 159     def suitable(cls, url):
 160         return False if PinterestIE.suitable(url) else super(
 161             PinterestCollectionIE, cls).suitable(url)
 162
 163     def _real_extract(self, url):
 164         username, slug = self._match_valid_url(url).groups()
 165         board = self._call_api(
 166             'Board', slug, {
 167                 'slug': slug,
 168                 'username': username
 169             })['data']
 170         board_id = board['id']
 171         options = {
 172             'board_id': board_id,
 173             'page_size': 250,
 174         }
 175         bookmark = None
 176         entries = []
 177         while True:
 178             if bookmark:
 179                 options['bookmarks'] = [bookmark]
 180             board_feed = self._call_api('BoardFeed', board_id, options)
 181             for item in (board_feed.get('data') or []):
 182                 if not isinstance(item, dict) or item.get('type') != 'pin':
 183                     continue
 184                 video_id = item.get('id')
 185                 if video_id:
 186                     # Some pins may not be available anonymously via pin URL
 187                     # video = self._extract_video(item, extract_formats=False)
 188                     # video.update({
 189                     #     '_type': 'url_transparent',
 190                     #     'url': 'https://www.pinterest.com/pin/%s/' % video_id,
 191                     # })
 192                     # entries.append(video)
 193                     entries.append(self._extract_video(item))
 194             bookmark = board_feed.get('bookmark')
 195             if not bookmark:
 196                 break
 197         return self.playlist_result(
 198             entries, playlist_id=board_id, playlist_title=board.get('name'))