yt_dlp/extractor/nytimes.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import hmac
   5 import hashlib
   6 import base64
   7
   8 from .common import InfoExtractor
   9 from ..utils import (
  10     determine_ext,
  11     float_or_none,
  12     int_or_none,
  13     js_to_json,
  14     mimetype2ext,
  15     parse_iso8601,
  16     remove_start,
  17 )
  18
  19
  20 class NYTimesBaseIE(InfoExtractor):
  21     _SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v'
  22
  23     def _extract_video_from_id(self, video_id):
  24         # Authorization generation algorithm is reverse engineered from `signer` in
  25         # http://graphics8.nytimes.com/video/vhs/vhs-2.x.min.js
  26         path = '/svc/video/api/v3/video/' + video_id
  27         hm = hmac.new(self._SECRET, (path + ':vhs').encode(), hashlib.sha512).hexdigest()
  28         video_data = self._download_json('http://www.nytimes.com' + path, video_id, 'Downloading video JSON', headers={
  29             'Authorization': 'NYTV ' + base64.b64encode(hm.encode()).decode(),
  30             'X-NYTV': 'vhs',
  31         }, fatal=False)
  32         if not video_data:
  33             video_data = self._download_json(
  34                 'http://www.nytimes.com/svc/video/api/v2/video/' + video_id,
  35                 video_id, 'Downloading video JSON')
  36
  37         title = video_data['headline']
  38
  39         def get_file_size(file_size):
  40             if isinstance(file_size, int):
  41                 return file_size
  42             elif isinstance(file_size, dict):
  43                 return int(file_size.get('value', 0))
  44             else:
  45                 return None
  46
  47         urls = []
  48         formats = []
  49         subtitles = {}
  50         for video in video_data.get('renditions', []):
  51             video_url = video.get('url')
  52             format_id = video.get('type')
  53             if not video_url or format_id == 'thumbs' or video_url in urls:
  54                 continue
  55             urls.append(video_url)
  56             ext = mimetype2ext(video.get('mimetype')) or determine_ext(video_url)
  57             if ext == 'm3u8':
  58                 m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
  59                     video_url, video_id, 'mp4', 'm3u8_native',
  60                     m3u8_id=format_id or 'hls', fatal=False)
  61                 formats.extend(m3u8_fmts)
  62                 subtitles = self._merge_subtitles(subtitles, m3u8_subs)
  63             elif ext == 'mpd':
  64                 continue
  65             #     formats.extend(self._extract_mpd_formats(
  66             #         video_url, video_id, format_id or 'dash', fatal=False))
  67             else:
  68                 formats.append({
  69                     'url': video_url,
  70                     'format_id': format_id,
  71                     'vcodec': video.get('videoencoding') or video.get('video_codec'),
  72                     'width': int_or_none(video.get('width')),
  73                     'height': int_or_none(video.get('height')),
  74                     'filesize': get_file_size(video.get('file_size') or video.get('fileSize')),
  75                     'tbr': int_or_none(video.get('bitrate'), 1000) or None,
  76                     'ext': ext,
  77                 })
  78         self._sort_formats(formats)
  79
  80         thumbnails = []
  81         for image in video_data.get('images', []):
  82             image_url = image.get('url')
  83             if not image_url:
  84                 continue
  85             thumbnails.append({
  86                 'url': 'http://www.nytimes.com/' + image_url,
  87                 'width': int_or_none(image.get('width')),
  88                 'height': int_or_none(image.get('height')),
  89             })
  90
  91         publication_date = video_data.get('publication_date')
  92         timestamp = parse_iso8601(publication_date[:-8]) if publication_date else None
  93
  94         return {
  95             'id': video_id,
  96             'title': title,
  97             'description': video_data.get('summary'),
  98             'timestamp': timestamp,
  99             'uploader': video_data.get('byline'),
 100             'duration': float_or_none(video_data.get('duration'), 1000),
 101             'formats': formats,
 102             'subtitles': subtitles,
 103             'thumbnails': thumbnails,
 104         }
 105
 106
 107 class NYTimesIE(NYTimesBaseIE):
 108     _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
 109
 110     _TESTS = [{
 111         'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
 112         'md5': 'd665342765db043f7e225cff19df0f2d',
 113         'info_dict': {
 114             'id': '100000002847155',
 115             'ext': 'mov',
 116             'title': 'Verbatim: What Is a Photocopier?',
 117             'description': 'md5:93603dada88ddbda9395632fdc5da260',
 118             'timestamp': 1398631707,
 119             'upload_date': '20140427',
 120             'uploader': 'Brett Weiner',
 121             'duration': 419,
 122         }
 123     }, {
 124         'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
 125         'only_matching': True,
 126     }]
 127
 128     def _real_extract(self, url):
 129         video_id = self._match_id(url)
 130
 131         return self._extract_video_from_id(video_id)
 132
 133
 134 class NYTimesArticleIE(NYTimesBaseIE):
 135     _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))*?/(?:[^/]+/)*(?P<id>[^.]+)(?:\.html)?'
 136     _TESTS = [{
 137         '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',
 138         'md5': 'e2076d58b4da18e6a001d53fd56db3c9',
 139         'info_dict': {
 140             'id': '100000003628438',
 141             'ext': 'mov',
 142             'title': 'New Minimum Wage: $70,000 a Year',
 143             '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.',
 144             'timestamp': 1429033037,
 145             'upload_date': '20150414',
 146             'uploader': 'Matthew Williams',
 147         }
 148     }, {
 149         'url': 'http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
 150         'md5': 'e0d52040cafb07662acf3c9132db3575',
 151         'info_dict': {
 152             'id': '100000004709062',
 153             'title': 'The Run-Up: ‘He Was Like an Octopus’',
 154             'ext': 'mp3',
 155             'description': 'md5:fb5c6b93b12efc51649b4847fe066ee4',
 156             'series': 'The Run-Up',
 157             'episode': '‘He Was Like an Octopus’',
 158             'episode_number': 20,
 159             'duration': 2130,
 160         }
 161     }, {
 162         'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
 163         'info_dict': {
 164             'id': '100000004709479',
 165             'title': 'The Rise of Hitler',
 166             'ext': 'mp3',
 167             'description': 'md5:bce877fd9e3444990cb141875fab0028',
 168             'creator': 'Pamela Paul',
 169             'duration': 3475,
 170         },
 171         'params': {
 172             'skip_download': True,
 173         },
 174     }, {
 175         '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',
 176         'only_matching': True,
 177     }]
 178
 179     def _extract_podcast_from_json(self, json, page_id, webpage):
 180         podcast_audio = self._parse_json(
 181             json, page_id, transform_source=js_to_json)
 182
 183         audio_data = podcast_audio['data']
 184         track = audio_data['track']
 185
 186         episode_title = track['title']
 187         video_url = track['source']
 188
 189         description = track.get('description') or self._html_search_meta(
 190             ['og:description', 'twitter:description'], webpage)
 191
 192         podcast_title = audio_data.get('podcast', {}).get('title')
 193         title = ('%s: %s' % (podcast_title, episode_title)
 194                  if podcast_title else episode_title)
 195
 196         episode = audio_data.get('podcast', {}).get('episode') or ''
 197         episode_number = int_or_none(self._search_regex(
 198             r'[Ee]pisode\s+(\d+)', episode, 'episode number', default=None))
 199
 200         return {
 201             'id': remove_start(podcast_audio.get('target'), 'FT') or page_id,
 202             'url': video_url,
 203             'title': title,
 204             'description': description,
 205             'creator': track.get('credit'),
 206             'series': podcast_title,
 207             'episode': episode_title,
 208             'episode_number': episode_number,
 209             'duration': int_or_none(track.get('duration')),
 210         }
 211
 212     def _real_extract(self, url):
 213         page_id = self._match_id(url)
 214
 215         webpage = self._download_webpage(url, page_id)
 216
 217         video_id = self._search_regex(
 218             r'data-videoid=["\'](\d+)', webpage, 'video id',
 219             default=None, fatal=False)
 220         if video_id is not None:
 221             return self._extract_video_from_id(video_id)
 222
 223         podcast_data = self._search_regex(
 224             (r'NYTD\.FlexTypes\.push\s*\(\s*({.+?})\s*\)\s*;\s*</script',
 225              r'NYTD\.FlexTypes\.push\s*\(\s*({.+})\s*\)\s*;'),
 226             webpage, 'podcast data')
 227         return self._extract_podcast_from_json(podcast_data, page_id, webpage)
 228
 229
 230 class NYTimesCookingIE(NYTimesBaseIE):
 231     _VALID_URL = r'https?://cooking\.nytimes\.com/(?:guid|recip)es/(?P<id>\d+)'
 232     _TESTS = [{
 233         'url': 'https://cooking.nytimes.com/recipes/1017817-cranberry-curd-tart',
 234         'md5': 'dab81fa2eaeb3f9ed47498bdcfcdc1d3',
 235         'info_dict': {
 236             'id': '100000004756089',
 237             'ext': 'mov',
 238             'timestamp': 1479383008,
 239             'uploader': 'By SHAW LASH, ADAM SAEWITZ and JAMES HERRON',
 240             'title': 'Cranberry Tart',
 241             'upload_date': '20161117',
 242             'description': 'If you are a fan of lemon curd or the classic French tarte au citron, you will love this cranberry version.',
 243         },
 244     }, {
 245         'url': 'https://cooking.nytimes.com/guides/13-how-to-cook-a-turkey',
 246         'md5': '4b2e8c70530a89b8d905a2b572316eb8',
 247         'info_dict': {
 248             'id': '100000003951728',
 249             'ext': 'mov',
 250             'timestamp': 1445509539,
 251             'description': 'Turkey guide',
 252             'upload_date': '20151022',
 253             'title': 'Turkey',
 254         }
 255     }]
 256
 257     def _real_extract(self, url):
 258         page_id = self._match_id(url)
 259
 260         webpage = self._download_webpage(url, page_id)
 261
 262         video_id = self._search_regex(
 263             r'data-video-id=["\'](\d+)', webpage, 'video id')
 264
 265         return self._extract_video_from_id(video_id)