_ Git - youtube-dl/blob - youtube_dl/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     float_or_none,
  11     int_or_none,
  12     js_to_json,
  13     parse_iso8601,
  14     mimetype2ext,
  15     determine_ext,
  16 )
  17
  18
  19 class NYTimesBaseIE(InfoExtractor):
  20     _SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v'
  21
  22     def _extract_video_from_id(self, video_id):
  23         # Authorization generation algorithm is reverse engineered from `signer` in
  24         # http://graphics8.nytimes.com/video/vhs/vhs-2.x.min.js
  25         path = '/svc/video/api/v3/video/' + video_id
  26         hm = hmac.new(self._SECRET, (path + ':vhs').encode(), hashlib.sha512).hexdigest()
  27         video_data = self._download_json('http://www.nytimes.com' + path, video_id, 'Downloading video JSON', headers={
  28             'Authorization': 'NYTV ' + base64.b64encode(hm.encode()).decode(),
  29             'X-NYTV': 'vhs',
  30         }, fatal=False)
  31         if not video_data:
  32             video_data = self._download_json(
  33                 'http://www.nytimes.com/svc/video/api/v2/video/' + video_id,
  34                 video_id, 'Downloading video JSON')
  35
  36         title = video_data['headline']
  37
  38         def get_file_size(file_size):
  39             if isinstance(file_size, int):
  40                 return file_size
  41             elif isinstance(file_size, dict):
  42                 return int(file_size.get('value', 0))
  43             else:
  44                 return None
  45
  46         urls = []
  47         formats = []
  48         for video in video_data.get('renditions', []):
  49             video_url = video.get('url')
  50             format_id = video.get('type')
  51             if not video_url or format_id == 'thumbs' or video_url in urls:
  52                 continue
  53             urls.append(video_url)
  54             ext = mimetype2ext(video.get('mimetype')) or determine_ext(video_url)
  55             if ext == 'm3u8':
  56                 formats.extend(self._extract_m3u8_formats(
  57                     video_url, video_id, 'mp4', 'm3u8_native',
  58                     m3u8_id=format_id or 'hls', fatal=False))
  59             elif ext == 'mpd':
  60                 continue
  61             #     formats.extend(self._extract_mpd_formats(
  62             #         video_url, video_id, format_id or 'dash', fatal=False))
  63             else:
  64                 formats.append({
  65                     'url': video_url,
  66                     'format_id': format_id,
  67                     'vcodec': video.get('videoencoding') or video.get('video_codec'),
  68                     'width': int_or_none(video.get('width')),
  69                     'height': int_or_none(video.get('height')),
  70                     'filesize': get_file_size(video.get('file_size') or video.get('fileSize')),
  71                     'tbr': int_or_none(video.get('bitrate'), 1000),
  72                     'ext': ext,
  73                 })
  74         self._sort_formats(formats)
  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             'thumbnails': thumbnails,
  99         }
 100
 101     def _extract_podcast_from_json(self, json, page_id, webpage):
 102         audio_data = self._parse_json(json, page_id, transform_source=js_to_json)['data']
 103
 104         description = audio_data['track'].get('description')
 105         if not description:
 106             description = self._html_search_meta(['og:description', 'twitter:description'], webpage)
 107
 108         episode_title = audio_data['track']['title']
 109         episode_number = None
 110         episode = audio_data['podcast']['episode'].split()
 111         if episode:
 112             episode_number = int_or_none(episode[-1])
 113             video_id = episode[-1]
 114         else:
 115             video_id = page_id
 116
 117         podcast_title = audio_data['podcast']['title']
 118         title = None
 119         if podcast_title:
 120             title = "%s: %s" % (podcast_title, episode_title)
 121         else:
 122             title = episode_title
 123
 124         info_dict = {
 125             'id': video_id,
 126             'title': title,
 127             'creator': audio_data['track'].get('credit'),
 128             'series': podcast_title,
 129             'episode': episode_title,
 130             'episode_number': episode_number,
 131             'url': audio_data['track']['source'],
 132             'duration': audio_data['track'].get('duration'),
 133             'description': description,
 134         }
 135
 136         return info_dict
 137
 138
 139 class NYTimesIE(NYTimesBaseIE):
 140     _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
 141
 142     _TESTS = [{
 143         'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
 144         'md5': 'd665342765db043f7e225cff19df0f2d',
 145         'info_dict': {
 146             'id': '100000002847155',
 147             'ext': 'mov',
 148             'title': 'Verbatim: What Is a Photocopier?',
 149             'description': 'md5:93603dada88ddbda9395632fdc5da260',
 150             'timestamp': 1398631707,
 151             'upload_date': '20140427',
 152             'uploader': 'Brett Weiner',
 153             'duration': 419,
 154         }
 155     }, {
 156         'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
 157         'only_matching': True,
 158     }]
 159
 160     def _real_extract(self, url):
 161         video_id = self._match_id(url)
 162
 163         return self._extract_video_from_id(video_id)
 164
 165
 166 class NYTimesArticleIE(NYTimesBaseIE):
 167     _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))*?/(?:[^/]+/)*(?P<id>[^.]+)(?:\.html)?'
 168     _TESTS = [{
 169         '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',
 170         'md5': 'e2076d58b4da18e6a001d53fd56db3c9',
 171         'info_dict': {
 172             'id': '100000003628438',
 173             'ext': 'mov',
 174             'title': 'New Minimum Wage: $70,000 a Year',
 175             '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.',
 176             'timestamp': 1429033037,
 177             'upload_date': '20150414',
 178             'uploader': 'Matthew Williams',
 179         }
 180     }, {
 181         'url': 'http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
 182         'md5': 'e0d52040cafb07662acf3c9132db3575',
 183         'info_dict': {
 184             'id': '20',
 185             'title': "The Run-Up: \u2018He Was Like an Octopus\u2019",
 186             'ext': 'mp3',
 187             'description': 'We go behind the story of the two women who told us that Donald Trump touched them inappropriately (which he denies) and check in on Hillary Clinton’s campaign.',
 188         }
 189     }, {
 190         'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
 191         'md5': '66fb5471d7ef15da98af176dc1af4cb9',
 192         'info_dict': {
 193             'id': 'inside-the-new-york-times-book-review-the-rise-of-hitler',
 194             'title': "The Rise of Hitler",
 195             'ext': 'mp3',
 196             'description': 'Adam Kirsch discusses Volker Ullrich\'s new biography of Hitler; Billy Collins talks about his latest collection of poems; and iO Tillett Wright on his new memoir, "Darling Days."',
 197             }
 198     }, {
 199         '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',
 200         'only_matching': True,
 201     }]
 202
 203     def _real_extract(self, url):
 204         page_id = self._match_id(url)
 205
 206         webpage = self._download_webpage(url, page_id)
 207
 208         video_id = self._html_search_regex(r'data-videoid="(\d+)"', webpage, 'video id', None, False)
 209         if video_id is not None:
 210             return self._extract_video_from_id(video_id)
 211
 212         data_json = self._html_search_regex(r'NYTD\.FlexTypes\.push\(({.*})\);', webpage, 'json data')
 213         return self._extract_podcast_from_json(data_json, page_id, webpage)