import json
import uuid
from .common import InfoExtractor
from ..utils import (
ExtractorError,
clean_html,
determine_ext,
extract_attributes,
float_or_none,
get_elements_html_by_class,
int_or_none,
merge_dicts,
mimetype2ext,
parse_iso8601,
remove_end,
remove_start,
str_or_none,
traverse_obj,
url_or_none,
)
class NYTimesBaseIE(InfoExtractor):
_DNS_NAMESPACE = uuid.UUID('36dd619a-56dc-595b-9e09-37f4152c7b5d')
_TOKEN = 'MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAuNIzKBOFB77aT/jN/FQ+/QVKWq5V1ka1AYmCR9hstz1pGNPH5ajOU9gAqta0T89iPnhjwla+3oec/Z3kGjxbpv6miQXufHFq3u2RC6HyU458cLat5kVPSOQCe3VVB5NRpOlRuwKHqn0txfxnwSSj8mqzstR997d3gKB//RO9zE16y3PoWlDQXkASngNJEWvL19iob/xwAkfEWCjyRILWFY0JYX3AvLMSbq7wsqOCE5srJpo7rRU32zsByhsp1D5W9OYqqwDmflsgCEQy2vqTsJjrJohuNg+urMXNNZ7Y3naMoqttsGDrWVxtPBafKMI8pM2ReNZBbGQsQXRzQNo7+QIDAQAB'
_GRAPHQL_API = 'https://samizdat-graphql.nytimes.com/graphql/v2'
_GRAPHQL_QUERY = '''query VideoQuery($id: String!) {
video(id: $id) {
... on Video {
bylines {
renderedRepresentation
}
duration
promotionalHeadline
promotionalMedia {
... on Image {
crops {
name
renditions {
name
width
height
url
}
}
}
}
renditions {
type
width
height
url
bitrate
}
summary
}
}
}'''
def _call_api(self, media_id):
# reference: `id-to-uri.js`
video_uuid = uuid.uuid5(self._DNS_NAMESPACE, 'video')
media_uuid = uuid.uuid5(video_uuid, media_id)
return traverse_obj(self._download_json(
self._GRAPHQL_API, media_id, 'Downloading JSON from GraphQL API', data=json.dumps({
'query': self._GRAPHQL_QUERY,
'variables': {'id': f'nyt://video/{media_uuid}'},
}, separators=(',', ':')).encode(), headers={
'Content-Type': 'application/json',
'Nyt-App-Type': 'vhs',
'Nyt-App-Version': 'v3.52.21',
'Nyt-Token': self._TOKEN,
'Origin': 'https://nytimes.com',
}, fatal=False), ('data', 'video', {dict})) or {}
def _extract_thumbnails(self, thumbs):
return traverse_obj(thumbs, (lambda _, v: url_or_none(v['url']), {
'url': 'url',
'width': ('width', {int_or_none}),
'height': ('height', {int_or_none}),
}), default=None)
def _extract_formats_and_subtitles(self, video_id, content_media_json):
urls = []
formats = []
subtitles = {}
for video in traverse_obj(content_media_json, ('renditions', ..., {dict})):
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)
self._merge_subtitles(m3u8_subs, target=subtitles)
elif ext == 'mpd':
continue # all mpd urls give 404 errors
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': traverse_obj(video, (
('file_size', 'fileSize'), (None, ('value')), {int_or_none}), get_all=False),
'tbr': int_or_none(video.get('bitrate'), 1000) or None,
'ext': ext,
})
return formats, subtitles
def _extract_video(self, media_id):
data = self._call_api(media_id)
formats, subtitles = self._extract_formats_and_subtitles(media_id, data)
return {
'id': media_id,
'title': data.get('promotionalHeadline'),
'description': data.get('summary'),
'duration': float_or_none(data.get('duration'), scale=1000),
'creator': ', '.join(traverse_obj(data, ( # TODO: change to 'creators'
'bylines', ..., 'renderedRepresentation', {lambda x: remove_start(x, 'By ')}))),
'formats': formats,
'subtitles': subtitles,
'thumbnails': self._extract_thumbnails(
traverse_obj(data, ('promotionalMedia', 'crops', ..., 'renditions', ...))),
}
class NYTimesIE(NYTimesBaseIE):
_VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P