[yt-dlp.git] / yt_dlp / extractor / nytimes.py

import hmac
import hashlib
import base64

from .common import InfoExtractor
from ..utils import (
    determine_ext,
    float_or_none,
    int_or_none,
    js_to_json,
    mimetype2ext,
    parse_iso8601,
    remove_start,
)


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 = []
        subtitles = {}
        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':
                m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
                    video_url, video_id, 'mp4', 'm3u8_native',
                    m3u8_id=format_id or 'hls', fatal=False)
                formats.extend(m3u8_fmts)
                subtitles = self._merge_subtitles(subtitles, m3u8_subs)
            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) or None,
                    '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,
            'subtitles': subtitles,
            'thumbnails': thumbnails,
        }


class NYTimesIE(NYTimesBaseIE):
    _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
    _EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//graphics8\.nytimes\.com/bcvideo/[^/]+/iframe/embed\.html.+?)\1>']

    _TESTS = [{
        'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
        'md5': 'd665342765db043f7e225cff19df0f2d',
        'info_dict': {
            'id': '100000002847155',
            'ext': 'mov',
            'title': 'Verbatim: What Is a Photocopier?',
            'description': 'md5:93603dada88ddbda9395632fdc5da260',
            'timestamp': 1398631707,
            'upload_date': '20140427',
            'uploader': 'Brett Weiner',
            'duration': 419,
        }
    }, {
        'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
        'only_matching': True,
    }]

    def _real_extract(self, url):
        video_id = self._match_id(url)

        return self._extract_video_from_id(video_id)


class NYTimesArticleIE(NYTimesBaseIE):
    _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))*?/(?:[^/]+/)*(?P<id>[^.]+)(?:\.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/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
        'md5': 'e0d52040cafb07662acf3c9132db3575',
        'info_dict': {
            'id': '100000004709062',
            'title': 'The Run-Up: ‘He Was Like an Octopus’',
            'ext': 'mp3',
            'description': 'md5:fb5c6b93b12efc51649b4847fe066ee4',
            'series': 'The Run-Up',
            'episode': '‘He Was Like an Octopus’',
            'episode_number': 20,
            'duration': 2130,
        }
    }, {
        'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
        'info_dict': {
            'id': '100000004709479',
            'title': 'The Rise of Hitler',
            'ext': 'mp3',
            'description': 'md5:bce877fd9e3444990cb141875fab0028',
            'creator': 'Pamela Paul',
            'duration': 3475,
        },
        'params': {
            'skip_download': True,
        },
    }, {
        '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 _extract_podcast_from_json(self, json, page_id, webpage):
        podcast_audio = self._parse_json(
            json, page_id, transform_source=js_to_json)

        audio_data = podcast_audio['data']
        track = audio_data['track']

        episode_title = track['title']
        video_url = track['source']

        description = track.get('description') or self._html_search_meta(
            ['og:description', 'twitter:description'], webpage)

        podcast_title = audio_data.get('podcast', {}).get('title')
        title = ('%s: %s' % (podcast_title, episode_title)
                 if podcast_title else episode_title)

        episode = audio_data.get('podcast', {}).get('episode') or ''
        episode_number = int_or_none(self._search_regex(
            r'[Ee]pisode\s+(\d+)', episode, 'episode number', default=None))

        return {
            'id': remove_start(podcast_audio.get('target'), 'FT') or page_id,
            'url': video_url,
            'title': title,
            'description': description,
            'creator': track.get('credit'),
            'series': podcast_title,
            'episode': episode_title,
            'episode_number': episode_number,
            'duration': int_or_none(track.get('duration')),
        }

    def _real_extract(self, url):
        page_id = self._match_id(url)

        webpage = self._download_webpage(url, page_id)

        video_id = self._search_regex(
            r'data-videoid=["\'](\d+)', webpage, 'video id',
            default=None, fatal=False)
        if video_id is not None:
            return self._extract_video_from_id(video_id)

        podcast_data = self._search_regex(
            (r'NYTD\.FlexTypes\.push\s*\(\s*({.+?})\s*\)\s*;\s*</script',
             r'NYTD\.FlexTypes\.push\s*\(\s*({.+})\s*\)\s*;'),
            webpage, 'podcast data')
        return self._extract_podcast_from_json(podcast_data, page_id, webpage)


class NYTimesCookingIE(NYTimesBaseIE):
    _VALID_URL = r'https?://cooking\.nytimes\.com/(?:guid|recip)es/(?P<id>\d+)'
    _TESTS = [{
        'url': 'https://cooking.nytimes.com/recipes/1017817-cranberry-curd-tart',
        'md5': 'dab81fa2eaeb3f9ed47498bdcfcdc1d3',
        'info_dict': {
            'id': '100000004756089',
            'ext': 'mov',
            'timestamp': 1479383008,
            'uploader': 'By SHAW LASH, ADAM SAEWITZ and JAMES HERRON',
            'title': 'Cranberry Tart',
            'upload_date': '20161117',
            'description': 'If you are a fan of lemon curd or the classic French tarte au citron, you will love this cranberry version.',
        },
    }, {
        'url': 'https://cooking.nytimes.com/guides/13-how-to-cook-a-turkey',
        'md5': '4b2e8c70530a89b8d905a2b572316eb8',
        'info_dict': {
            'id': '100000003951728',
            'ext': 'mov',
            'timestamp': 1445509539,
            'description': 'Turkey guide',
            'upload_date': '20151022',
            'title': 'Turkey',
        }
    }]

    def _real_extract(self, url):
        page_id = self._match_id(url)

        webpage = self._download_webpage(url, page_id)

        video_id = self._search_regex(
            r'data-video-id=["\'](\d+)', webpage, 'video id')

        return self._extract_video_from_id(video_id)
Commit	Line	Data
4191779d RA	1	import hmac
	2	import hashlib
	3	import base64
	4
d664de44	5	from .common import InfoExtractor
f3c0c667	6	from ..utils import (
47da7823	7	determine_ext,
f3c0c667 S	8	float_or_none,
f3c0c667 S	9	int_or_none,
74324a7a	10	js_to_json,
4191779d	11	mimetype2ext,
47da7823 S	12	parse_iso8601,
47da7823 S	13	remove_start,
f3c0c667	14	)
d664de44 S	15
d664de44 S	16
50aa43b3	17	class NYTimesBaseIE(InfoExtractor):
4191779d RA	18	_SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v'
4191779d RA	19
50aa43b3	20	def _extract_video_from_id(self, video_id):
4191779d RA	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')
d664de44 S	33
d664de44 S	34	title = video_data['headline']
d664de44	35
65157783 S	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:
4191779d RA	42	return None
	43
	44	urls = []
	45	formats = []
47f4203d	46	subtitles = {}
4191779d RA	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':
47f4203d	55	m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
4191779d	56	video_url, video_id, 'mp4', 'm3u8_native',
47f4203d F	57	m3u8_id=format_id or 'hls', fatal=False)
	58	formats.extend(m3u8_fmts)
	59	subtitles = self._merge_subtitles(subtitles, m3u8_subs)
4191779d RA	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')),
f377edec	72	'tbr': int_or_none(video.get('bitrate'), 1000) or None,
4191779d RA	73	'ext': ext,
4191779d RA	74	})
54f37eea	75	self._sort_formats(formats)
d664de44	76
4191779d RA	77	thumbnails = []
	78	for image in video_data.get('images', []):
	79	image_url = image.get('url')
	80	if not image_url:
	81	continue
	82	thumbnails.append({
	83	'url': 'http://www.nytimes.com/' + image_url,
f3c0c667 S	84	'width': int_or_none(image.get('width')),
f3c0c667 S	85	'height': int_or_none(image.get('height')),
4191779d RA	86	})
	87
	88	publication_date = video_data.get('publication_date')
	89	timestamp = parse_iso8601(publication_date[:-8]) if publication_date else None
d664de44 S	90
	91	return {
	92	'id': video_id,
	93	'title': title,
4191779d	94	'description': video_data.get('summary'),
d664de44	95	'timestamp': timestamp,
4191779d RA	96	'uploader': video_data.get('byline'),
4191779d RA	97	'duration': float_or_none(video_data.get('duration'), 1000),
d664de44	98	'formats': formats,
47f4203d	99	'subtitles': subtitles,
d664de44	100	'thumbnails': thumbnails,
5f6a1245	101	}
50aa43b3 YCH	102
	103
	104	class NYTimesIE(NYTimesBaseIE):
	105	_VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?\|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
bfd973ec	106	_EMBED_REGEX = [r'<iframe[^>]+src=(["\'])(?P<url>(?:https?:)?//graphics8\.nytimes\.com/bcvideo/[^/]+/iframe/embed\.html.+?)\1>']
50aa43b3 YCH	107
	108	_TESTS = [{
	109	'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
4191779d	110	'md5': 'd665342765db043f7e225cff19df0f2d',
50aa43b3 YCH	111	'info_dict': {
	112	'id': '100000002847155',
	113	'ext': 'mov',
	114	'title': 'Verbatim: What Is a Photocopier?',
	115	'description': 'md5:93603dada88ddbda9395632fdc5da260',
	116	'timestamp': 1398631707,
	117	'upload_date': '20140427',
	118	'uploader': 'Brett Weiner',
	119	'duration': 419,
	120	}
	121	}, {
	122	'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
	123	'only_matching': True,
	124	}]
	125
	126	def _real_extract(self, url):
	127	video_id = self._match_id(url)
	128
	129	return self._extract_video_from_id(video_id)
	130
	131
	132	class NYTimesArticleIE(NYTimesBaseIE):
5332fd91	133	_VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))?/(?:[^/]+/)(?P<id>[^.]+)(?:\.html)?'
df8418ff	134	_TESTS = [{
50aa43b3 YCH	135	'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',
	136	'md5': 'e2076d58b4da18e6a001d53fd56db3c9',
	137	'info_dict': {
	138	'id': '100000003628438',
	139	'ext': 'mov',
	140	'title': 'New Minimum Wage: $70,000 a Year',
	141	'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.',
	142	'timestamp': 1429033037,
	143	'upload_date': '20150414',
	144	'uploader': 'Matthew Williams',
	145	}
74324a7a JH	146	}, {
	147	'url': 'http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
	148	'md5': 'e0d52040cafb07662acf3c9132db3575',
	149	'info_dict': {
47da7823 S	150	'id': '100000004709062',
47da7823 S	151	'title': 'The Run-Up: ‘He Was Like an Octopus’',
74324a7a	152	'ext': 'mp3',
47da7823 S	153	'description': 'md5:fb5c6b93b12efc51649b4847fe066ee4',
	154	'series': 'The Run-Up',
	155	'episode': '‘He Was Like an Octopus’',
	156	'episode_number': 20,
	157	'duration': 2130,
74324a7a JH	158	}
	159	}, {
	160	'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
74324a7a	161	'info_dict': {
47da7823 S	162	'id': '100000004709479',
47da7823 S	163	'title': 'The Rise of Hitler',
74324a7a	164	'ext': 'mp3',
47da7823 S	165	'description': 'md5:bce877fd9e3444990cb141875fab0028',
	166	'creator': 'Pamela Paul',
	167	'duration': 3475,
	168	},
	169	'params': {
	170	'skip_download': True,
	171	},
df8418ff YCH	172	}, {
	173	'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',
	174	'only_matching': True,
	175	}]
50aa43b3	176
47da7823 S	177	def _extract_podcast_from_json(self, json, page_id, webpage):
	178	podcast_audio = self._parse_json(
	179	json, page_id, transform_source=js_to_json)
	180
	181	audio_data = podcast_audio['data']
	182	track = audio_data['track']
	183
	184	episode_title = track['title']
	185	video_url = track['source']
	186
	187	description = track.get('description') or self._html_search_meta(
	188	['og:description', 'twitter:description'], webpage)
	189
	190	podcast_title = audio_data.get('podcast', {}).get('title')
	191	title = ('%s: %s' % (podcast_title, episode_title)
	192	if podcast_title else episode_title)
	193
	194	episode = audio_data.get('podcast', {}).get('episode') or ''
	195	episode_number = int_or_none(self._search_regex(
df4939b1	196	r'[Ee]pisode\s+(\d+)', episode, 'episode number', default=None))
47da7823 S	197
	198	return {
	199	'id': remove_start(podcast_audio.get('target'), 'FT') or page_id,
	200	'url': video_url,
	201	'title': title,
	202	'description': description,
	203	'creator': track.get('credit'),
	204	'series': podcast_title,
	205	'episode': episode_title,
	206	'episode_number': episode_number,
	207	'duration': int_or_none(track.get('duration')),
	208	}
	209
50aa43b3	210	def _real_extract(self, url):
74324a7a	211	page_id = self._match_id(url)
50aa43b3	212
74324a7a	213	webpage = self._download_webpage(url, page_id)
50aa43b3	214
47da7823 S	215	video_id = self._search_regex(
	216	r'data-videoid=["\'](\d+)', webpage, 'video id',
	217	default=None, fatal=False)
74324a7a JH	218	if video_id is not None:
74324a7a JH	219	return self._extract_video_from_id(video_id)
47da7823 S	220
	221	podcast_data = self._search_regex(
	222	(r'NYTD\.FlexTypes\.push\s\(\s({.+?})\s\)\s;\s*</script',
	223	r'NYTD\.FlexTypes\.push\s\(\s({.+})\s\)\s;'),
	224	webpage, 'podcast data')
	225	return self._extract_podcast_from_json(podcast_data, page_id, webpage)
70c5802b	226
	227
	228	class NYTimesCookingIE(NYTimesBaseIE):
	229	_VALID_URL = r'https?://cooking\.nytimes\.com/(?:guid\|recip)es/(?P<id>\d+)'
	230	_TESTS = [{
	231	'url': 'https://cooking.nytimes.com/recipes/1017817-cranberry-curd-tart',
	232	'md5': 'dab81fa2eaeb3f9ed47498bdcfcdc1d3',
	233	'info_dict': {
	234	'id': '100000004756089',
	235	'ext': 'mov',
	236	'timestamp': 1479383008,
	237	'uploader': 'By SHAW LASH, ADAM SAEWITZ and JAMES HERRON',
	238	'title': 'Cranberry Tart',
	239	'upload_date': '20161117',
	240	'description': 'If you are a fan of lemon curd or the classic French tarte au citron, you will love this cranberry version.',
	241	},
	242	}, {
	243	'url': 'https://cooking.nytimes.com/guides/13-how-to-cook-a-turkey',
	244	'md5': '4b2e8c70530a89b8d905a2b572316eb8',
	245	'info_dict': {
	246	'id': '100000003951728',
	247	'ext': 'mov',
	248	'timestamp': 1445509539,
	249	'description': 'Turkey guide',
	250	'upload_date': '20151022',
	251	'title': 'Turkey',
	252	}
	253	}]
	254
	255	def _real_extract(self, url):
	256	page_id = self._match_id(url)
	257
	258	webpage = self._download_webpage(url, page_id)
	259
	260	video_id = self._search_regex(
	261	r'data-video-id=["\'](\d+)', webpage, 'video id')
	262
	263	return self._extract_video_from_id(video_id)