X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fnytimes.py;h=142c34256f20fea317dfa91736f38fdeb6ad5126;hb=95843da5297965bb535262002c92a4d0afcb7e12;hp=03f0a4de6dccd8485091b802422bbfaf5f7402df;hpb=d08225edf454dd28d0bc15d94b23d62c4deab7f4;p=youtube-dl diff --git a/youtube_dl/extractor/nytimes.py b/youtube_dl/extractor/nytimes.py index 03f0a4de6..142c34256 100644 --- a/youtube_dl/extractor/nytimes.py +++ b/youtube_dl/extractor/nytimes.py @@ -1,19 +1,108 @@ from __future__ import unicode_literals +import hmac +import hashlib +import base64 + from .common import InfoExtractor from ..utils import ( float_or_none, int_or_none, parse_iso8601, + mimetype2ext, + determine_ext, ) -class NYTimesIE(InfoExtractor): +class NYTimesBaseIE(InfoExtractor): + _SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v' + + def _extract_video_from_id(self, video_id): + # Authorization generation algorithm is reverse engineered from `signer` in + # http://graphics8.nytimes.com/video/vhs/vhs-2.x.min.js + path = '/svc/video/api/v3/video/' + video_id + hm = hmac.new(self._SECRET, (path + ':vhs').encode(), hashlib.sha512).hexdigest() + video_data = self._download_json('http://www.nytimes.com' + path, video_id, 'Downloading video JSON', headers={ + 'Authorization': 'NYTV ' + base64.b64encode(hm.encode()).decode(), + 'X-NYTV': 'vhs', + }, fatal=False) + if not video_data: + video_data = self._download_json( + 'http://www.nytimes.com/svc/video/api/v2/video/' + video_id, + video_id, 'Downloading video JSON') + + title = video_data['headline'] + + def get_file_size(file_size): + if isinstance(file_size, int): + return file_size + elif isinstance(file_size, dict): + return int(file_size.get('value', 0)) + else: + return None + + urls = [] + formats = [] + for video in video_data.get('renditions', []): + video_url = video.get('url') + format_id = video.get('type') + if not video_url or format_id == 'thumbs' or video_url in urls: + continue + urls.append(video_url) + ext = mimetype2ext(video.get('mimetype')) or determine_ext(video_url) + if ext == 'm3u8': + formats.extend(self._extract_m3u8_formats( + video_url, video_id, 'mp4', 'm3u8_native', + m3u8_id=format_id or 'hls', fatal=False)) + elif ext == 'mpd': + continue + # formats.extend(self._extract_mpd_formats( + # video_url, video_id, format_id or 'dash', fatal=False)) + else: + formats.append({ + 'url': video_url, + 'format_id': format_id, + 'vcodec': video.get('videoencoding') or video.get('video_codec'), + 'width': int_or_none(video.get('width')), + 'height': int_or_none(video.get('height')), + 'filesize': get_file_size(video.get('file_size') or video.get('fileSize')), + 'tbr': int_or_none(video.get('bitrate'), 1000), + 'ext': ext, + }) + self._sort_formats(formats) + + thumbnails = [] + for image in video_data.get('images', []): + image_url = image.get('url') + if not image_url: + continue + thumbnails.append({ + 'url': 'http://www.nytimes.com/' + image_url, + 'width': int_or_none(image.get('width')), + 'height': int_or_none(image.get('height')), + }) + + publication_date = video_data.get('publication_date') + timestamp = parse_iso8601(publication_date[:-8]) if publication_date else None + + return { + 'id': video_id, + 'title': title, + 'description': video_data.get('summary'), + 'timestamp': timestamp, + 'uploader': video_data.get('byline'), + 'duration': float_or_none(video_data.get('duration'), 1000), + 'formats': formats, + 'thumbnails': thumbnails, + } + + +class NYTimesIE(NYTimesBaseIE): _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P\d+)' _TESTS = [{ 'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263', - 'md5': '18a525a510f942ada2720db5f31644c0', + 'md5': 'd665342765db043f7e225cff19df0f2d', 'info_dict': { 'id': '100000002847155', 'ext': 'mov', @@ -32,52 +121,33 @@ class NYTimesIE(InfoExtractor): def _real_extract(self, url): video_id = self._match_id(url) - video_data = self._download_json( - 'http://www.nytimes.com/svc/video/api/v2/video/%s' % video_id, - video_id, 'Downloading video JSON') + return self._extract_video_from_id(video_id) - title = video_data['headline'] - description = video_data.get('summary') - duration = float_or_none(video_data.get('duration'), 1000) - uploader = video_data['byline'] - timestamp = parse_iso8601(video_data['publication_date'][:-8]) +class NYTimesArticleIE(NYTimesBaseIE): + _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?[^.]+)(?:\.html)?' + _TESTS = [{ + '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', + 'md5': 'e2076d58b4da18e6a001d53fd56db3c9', + 'info_dict': { + 'id': '100000003628438', + 'ext': 'mov', + 'title': 'New Minimum Wage: $70,000 a Year', + '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.', + 'timestamp': 1429033037, + 'upload_date': '20150414', + 'uploader': 'Matthew Williams', + } + }, { + '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', + 'only_matching': True, + }] - def get_file_size(file_size): - if isinstance(file_size, int): - return file_size - elif isinstance(file_size, dict): - return int(file_size.get('value', 0)) - else: - return 0 - - formats = [ - { - 'url': video['url'], - 'format_id': video.get('type'), - 'vcodec': video.get('video_codec'), - 'width': int_or_none(video.get('width')), - 'height': int_or_none(video.get('height')), - 'filesize': get_file_size(video.get('fileSize')), - } for video in video_data['renditions'] - ] - self._sort_formats(formats) + def _real_extract(self, url): + video_id = self._match_id(url) - thumbnails = [ - { - 'url': 'http://www.nytimes.com/%s' % image['url'], - 'width': int_or_none(image.get('width')), - 'height': int_or_none(image.get('height')), - } for image in video_data['images'] - ] + webpage = self._download_webpage(url, video_id) - return { - 'id': video_id, - 'title': title, - 'description': description, - 'timestamp': timestamp, - 'uploader': uploader, - 'duration': duration, - 'formats': formats, - 'thumbnails': thumbnails, - } + video_id = self._html_search_regex(r'data-videoid="(\d+)"', webpage, 'video id') + + return self._extract_video_from_id(video_id)