yt_dlp/extractor/nytimes.py

   1 import hmac
   2 import hashlib
   3 import base64
   4
   5 from .common import InfoExtractor
   6 from ..utils import (
   7     determine_ext,
   8     float_or_none,
   9     int_or_none,
  10     js_to_json,
  11     mimetype2ext,
  12     parse_iso8601,
  13     remove_start,
  14 )
  15
  16
  17 class NYTimesBaseIE(InfoExtractor):
  18     _SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v'
  19
  20     def _extract_video_from_id(self, video_id):
  21         # Authorization generation algorithm is reverse engineered from `signer` in
  22         # http://graphics8.nytimes.com/video/vhs/vhs-2.x.min.js
  23         path = '/svc/video/api/v3/video/' + video_id
  24         hm = hmac.new(self._SECRET, (path + ':vhs').encode(), hashlib.sha512).hexdigest()
  25         video_data = self._download_json('http://www.nytimes.com' + path, video_id, 'Downloading video JSON', headers={
  26             'Authorization': 'NYTV ' + base64.b64encode(hm.encode()).decode(),
  27             'X-NYTV': 'vhs',
  28         }, fatal=False)
  29         if not video_data:
  30             video_data = self._download_json(
  31                 'http://www.nytimes.com/svc/video/api/v2/video/' + video_id,
  32                 video_id, 'Downloading video JSON')
  33
  34         title = video_data['headline']
  35
  36         def get_file_size(file_size):
  37             if isinstance(file_size, int):
  38                 return file_size
  39             elif isinstance(file_size, dict):
  40                 return int(file_size.get('value', 0))
  41             else:
  42                 return None
  43
  44         urls = []
  45         formats = []
  46         subtitles = {}
  47         for video in video_data.get('renditions', []):
  48             video_url = video.get('url')
  49             format_id = video.get('type')
  50             if not video_url or format_id == 'thumbs' or video_url in urls:
  51                 continue
  52             urls.append(video_url)
  53             ext = mimetype2ext(video.get('mimetype')) or determine_ext(video_url)
  54             if ext == 'm3u8':
  55                 m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
  56                     video_url, video_id, 'mp4', 'm3u8_native',
  57                     m3u8_id=format_id or 'hls', fatal=False)
  58                 formats.extend(m3u8_fmts)
  59                 subtitles = self._merge_subtitles(subtitles, m3u8_subs)
  60             elif ext == 'mpd':
  61                 continue
  62             #     formats.extend(self._extract_mpd_formats(
  63             #         video_url, video_id, format_id or 'dash', fatal=False))
  64             else:
  65                 formats.append({
  66                     'url': video_url,
  67                     'format_id': format_id,
  68                     'vcodec': video.get('videoencoding') or video.get('video_codec'),
  69                     'width': int_or_none(video.get('width')),
  70                     'height': int_or_none(video.get('height')),
  71                     'filesize': get_file_size(video.get('file_size') or video.get('fileSize')),
  72                     'tbr': int_or_none(video.get('bitrate'), 1000) or None,
  73                     'ext': ext,
  74                 })
  75
  76         thumbnails = []
  77         for image in video_data.get('images', []):
  78             image_url = image.get('url')
  79             if not image_url:
  80                 continue
  81             thumbnails.append({
  82                 'url': 'http://www.nytimes.com/' + image_url,
  83                 'width': int_or_none(image.get('width')),
  84                 'height': int_or_none(image.get('height')),
  85             })
  86
  87         publication_date = video_data.get('publication_date')
  88         timestamp = parse_iso8601(publication_date[:-8]) if publication_date else None
  89
  90         return {
  91             'id': video_id,
  92             'title': title,
  93             'description': video_data.get('summary'),
  94             'timestamp': timestamp,
  95             'uploader': video_data.get('byline'),
  96             'duration': float_or_none(video_data.get('duration'), 1000),
  97             'formats': formats,
  98             'subtitles': subtitles,
  99             'thumbnails': thumbnails,
 100         }
 101
 102
 103 class NYTimesIE(NYTimesBaseIE):
 104     _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
 105     _EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//graphics8\.nytimes\.com/bcvideo/[^/]+/iframe/embed\.html.+?)\1>']
 106
 107     _TESTS = [{
 108         'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
 109         'md5': 'd665342765db043f7e225cff19df0f2d',
 110         'info_dict': {
 111             'id': '100000002847155',
 112             'ext': 'mov',
 113             'title': 'Verbatim: What Is a Photocopier?',
 114             'description': 'md5:93603dada88ddbda9395632fdc5da260',
 115             'timestamp': 1398631707,
 116             'upload_date': '20140427',
 117             'uploader': 'Brett Weiner',
 118             'duration': 419,
 119         }
 120     }, {
 121         'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
 122         'only_matching': True,
 123     }]
 124
 125     def _real_extract(self, url):
 126         video_id = self._match_id(url)
 127
 128         return self._extract_video_from_id(video_id)
 129
 130
 131 class NYTimesArticleIE(NYTimesBaseIE):
 132     _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))*?/(?:[^/]+/)*(?P<id>[^.]+)(?:\.html)?'
 133     _TESTS = [{
 134         'url': 'http://www.nytimes.com/2015/04/14/business/owner-of-gravity-payments-a-credit-card-processor-is-setting-a-new-minimum-wage-70000-a-year.html?_r=0',
 135         'md5': 'e2076d58b4da18e6a001d53fd56db3c9',
 136         'info_dict': {
 137             'id': '100000003628438',
 138             'ext': 'mov',
 139             'title': 'New Minimum Wage: $70,000 a Year',
 140             'description': 'Dan Price, C.E.O. of Gravity Payments, surprised his 120-person staff by announcing that he planned over the next three years to raise the salary of every employee to $70,000 a year.',
 141             'timestamp': 1429033037,
 142             'upload_date': '20150414',
 143             'uploader': 'Matthew Williams',
 144         }
 145     }, {
 146         'url': 'http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
 147         'md5': 'e0d52040cafb07662acf3c9132db3575',
 148         'info_dict': {
 149             'id': '100000004709062',
 150             'title': 'The Run-Up: ‘He Was Like an Octopus’',
 151             'ext': 'mp3',
 152             'description': 'md5:fb5c6b93b12efc51649b4847fe066ee4',
 153             'series': 'The Run-Up',
 154             'episode': '‘He Was Like an Octopus’',
 155             'episode_number': 20,
 156             'duration': 2130,
 157         }
 158     }, {
 159         'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
 160         'info_dict': {
 161             'id': '100000004709479',
 162             'title': 'The Rise of Hitler',
 163             'ext': 'mp3',
 164             'description': 'md5:bce877fd9e3444990cb141875fab0028',
 165             'creator': 'Pamela Paul',
 166             'duration': 3475,
 167         },
 168         'params': {
 169             'skip_download': True,
 170         },
 171     }, {
 172         'url': 'http://www.nytimes.com/news/minute/2014/03/17/times-minute-whats-next-in-crimea/?_php=true&_type=blogs&_php=true&_type=blogs&_r=1',
 173         'only_matching': True,
 174     }]
 175
 176     def _extract_podcast_from_json(self, json, page_id, webpage):
 177         podcast_audio = self._parse_json(
 178             json, page_id, transform_source=js_to_json)
 179
 180         audio_data = podcast_audio['data']
 181         track = audio_data['track']
 182
 183         episode_title = track['title']
 184         video_url = track['source']
 185
 186         description = track.get('description') or self._html_search_meta(
 187             ['og:description', 'twitter:description'], webpage)
 188
 189         podcast_title = audio_data.get('podcast', {}).get('title')
 190         title = ('%s: %s' % (podcast_title, episode_title)
 191                  if podcast_title else episode_title)
 192
 193         episode = audio_data.get('podcast', {}).get('episode') or ''
 194         episode_number = int_or_none(self._search_regex(
 195             r'[Ee]pisode\s+(\d+)', episode, 'episode number', default=None))
 196
 197         return {
 198             'id': remove_start(podcast_audio.get('target'), 'FT') or page_id,
 199             'url': video_url,
 200             'title': title,
 201             'description': description,
 202             'creator': track.get('credit'),
 203             'series': podcast_title,
 204             'episode': episode_title,
 205             'episode_number': episode_number,
 206             'duration': int_or_none(track.get('duration')),
 207         }
 208
 209     def _real_extract(self, url):
 210         page_id = self._match_id(url)
 211
 212         webpage = self._download_webpage(url, page_id)
 213
 214         video_id = self._search_regex(
 215             r'data-videoid=["\'](\d+)', webpage, 'video id',
 216             default=None, fatal=False)
 217         if video_id is not None:
 218             return self._extract_video_from_id(video_id)
 219
 220         podcast_data = self._search_regex(
 221             (r'NYTD\.FlexTypes\.push\s*\(\s*({.+?})\s*\)\s*;\s*</script',
 222              r'NYTD\.FlexTypes\.push\s*\(\s*({.+})\s*\)\s*;'),
 223             webpage, 'podcast data')
 224         return self._extract_podcast_from_json(podcast_data, page_id, webpage)
 225
 226
 227 class NYTimesCookingIE(NYTimesBaseIE):
 228     _VALID_URL = r'https?://cooking\.nytimes\.com/(?:guid|recip)es/(?P<id>\d+)'
 229     _TESTS = [{
 230         'url': 'https://cooking.nytimes.com/recipes/1017817-cranberry-curd-tart',
 231         'md5': 'dab81fa2eaeb3f9ed47498bdcfcdc1d3',
 232         'info_dict': {
 233             'id': '100000004756089',
 234             'ext': 'mov',
 235             'timestamp': 1479383008,
 236             'uploader': 'By SHAW LASH, ADAM SAEWITZ and JAMES HERRON',
 237             'title': 'Cranberry Tart',
 238             'upload_date': '20161117',
 239             'description': 'If you are a fan of lemon curd or the classic French tarte au citron, you will love this cranberry version.',
 240         },
 241     }, {
 242         'url': 'https://cooking.nytimes.com/guides/13-how-to-cook-a-turkey',
 243         'md5': '4b2e8c70530a89b8d905a2b572316eb8',
 244         'info_dict': {
 245             'id': '100000003951728',
 246             'ext': 'mov',
 247             'timestamp': 1445509539,
 248             'description': 'Turkey guide',
 249             'upload_date': '20151022',
 250             'title': 'Turkey',
 251         }
 252     }]
 253
 254     def _real_extract(self, url):
 255         page_id = self._match_id(url)
 256
 257         webpage = self._download_webpage(url, page_id)
 258
 259         video_id = self._search_regex(
 260             r'data-video-id=["\'](\d+)', webpage, 'video id')
 261
 262         return self._extract_video_from_id(video_id)