[youtube] Add `thirdParty` to agegate clients (#577)

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

# coding: utf-8\r
from __future__ import unicode_literals\r
\r
import json\r
import time\r
\r
from urllib.error import HTTPError\r
from .common import InfoExtractor\r
from ..compat import compat_str, compat_urllib_parse_unquote, compat_urllib_parse_quote\r
from ..utils import (\r
    ExtractorError,\r
    parse_iso8601,\r
    try_get,\r
    urljoin,\r
)\r
\r
\r
class NebulaIE(InfoExtractor):\r
\r
    _VALID_URL = r'https?://(?:www\.)?(?:watchnebula\.com|nebula\.app)/videos/(?P<id>[-\w]+)'\r
    _TESTS = [\r
        {\r
            'url': 'https://nebula.app/videos/that-time-disney-remade-beauty-and-the-beast',\r
            'md5': 'fe79c4df8b3aa2fea98a93d027465c7e',\r
            'info_dict': {\r
                'id': '5c271b40b13fd613090034fd',\r
                'ext': 'mp4',\r
                'title': 'That Time Disney Remade Beauty and the Beast',\r
                'description': 'Note: this video was originally posted on YouTube with the sponsor read included. We weren’t able to remove it without reducing video quality, so it’s presented here in its original context.',\r
                'upload_date': '20180731',\r
                'timestamp': 1533009600,\r
                'channel': 'Lindsay Ellis',\r
                'uploader': 'Lindsay Ellis',\r
            },\r
            'params': {\r
                'usenetrc': True,\r
            },\r
            'skip': 'All Nebula content requires authentication',\r
        },\r
        {\r
            'url': 'https://nebula.app/videos/the-logistics-of-d-day-landing-craft-how-the-allies-got-ashore',\r
            'md5': '6d4edd14ce65720fa63aba5c583fb328',\r
            'info_dict': {\r
                'id': '5e7e78171aaf320001fbd6be',\r
                'ext': 'mp4',\r
                'title': 'Landing Craft - How The Allies Got Ashore',\r
                'description': r're:^In this episode we explore the unsung heroes of D-Day, the landing craft.',\r
                'upload_date': '20200327',\r
                'timestamp': 1585348140,\r
                'channel': 'The Logistics of D-Day',\r
                'uploader': 'The Logistics of D-Day',\r
            },\r
            'params': {\r
                'usenetrc': True,\r
            },\r
            'skip': 'All Nebula content requires authentication',\r
        },\r
        {\r
            'url': 'https://nebula.app/videos/money-episode-1-the-draw',\r
            'md5': '8c7d272910eea320f6f8e6d3084eecf5',\r
            'info_dict': {\r
                'id': '5e779ebdd157bc0001d1c75a',\r
                'ext': 'mp4',\r
                'title': 'Episode 1: The Draw',\r
                'description': r'contains:There’s free money on offer… if the players can all work together.',\r
                'upload_date': '20200323',\r
                'timestamp': 1584980400,\r
                'channel': 'Tom Scott Presents: Money',\r
                'uploader': 'Tom Scott Presents: Money',\r
            },\r
            'params': {\r
                'usenetrc': True,\r
            },\r
            'skip': 'All Nebula content requires authentication',\r
        },\r
        {\r
            'url': 'https://watchnebula.com/videos/money-episode-1-the-draw',\r
            'only_matching': True,\r
        },\r
    ]\r
    _NETRC_MACHINE = 'watchnebula'\r
\r
    _nebula_token = None\r
\r
    def _retrieve_nebula_auth(self):\r
        """\r
        Log in to Nebula, and returns a Nebula API token\r
        """\r
\r
        username, password = self._get_login_info()\r
        if not (username and password):\r
            self.raise_login_required()\r
\r
        self.report_login()\r
        data = json.dumps({'email': username, 'password': password}).encode('utf8')\r
        response = self._download_json(\r
            'https://api.watchnebula.com/api/v1/auth/login/',\r
            data=data, fatal=False, video_id=None,\r
            headers={\r
                'content-type': 'application/json',\r
                # Submitting the 'sessionid' cookie always causes a 403 on auth endpoint\r
                'cookie': ''\r
            },\r
            note='Authenticating to Nebula with supplied credentials',\r
            errnote='Authentication failed or rejected')\r
        if not response or not response.get('key'):\r
            self.raise_login_required()\r
\r
        # save nebula token as cookie\r
        self._set_cookie(\r
            'nebula.app', 'nebula-auth',\r
            compat_urllib_parse_quote(\r
                json.dumps({\r
                    "apiToken": response["key"],\r
                    "isLoggingIn": False,\r
                    "isLoggingOut": False,\r
                }, separators=(",", ":"))),\r
            expire_time=int(time.time()) + 86400 * 365,\r
        )\r
\r
        return response['key']\r
\r
    def _retrieve_zype_api_key(self, page_url, display_id):\r
        """\r
        Retrieves the Zype API key\r
        """\r
\r
        # Find the js that has the API key from the webpage and download it\r
        webpage = self._download_webpage(page_url, video_id=display_id)\r
        main_script_relpath = self._search_regex(\r
            r'<script[^>]*src="(?P<script_relpath>[^"]*main.[0-9a-f]*.chunk.js)"[^>]*>', webpage,\r
            group='script_relpath', name='script relative path', fatal=True)\r
        main_script_abspath = urljoin(page_url, main_script_relpath)\r
        main_script = self._download_webpage(main_script_abspath, video_id=display_id,\r
                                             note='Retrieving Zype API key')\r
\r
        api_key = self._search_regex(\r
            r'REACT_APP_ZYPE_API_KEY\s*:\s*"(?P<api_key>[\w-]*)"', main_script,\r
            group='api_key', name='API key', fatal=True)\r
\r
        return api_key\r
\r
    def _call_zype_api(self, path, params, video_id, api_key, note):\r
        """\r
        A helper for making calls to the Zype API.\r
        """\r
        query = {'api_key': api_key, 'per_page': 1}\r
        query.update(params)\r
        return self._download_json('https://api.zype.com' + path, video_id, query=query, note=note)\r
\r
    def _call_nebula_api(self, path, video_id, access_token, note):\r
        """\r
        A helper for making calls to the Nebula API.\r
        """\r
        return self._download_json('https://api.watchnebula.com/api/v1' + path, video_id, headers={\r
            'Authorization': 'Token {access_token}'.format(access_token=access_token)\r
        }, note=note)\r
\r
    def _fetch_zype_access_token(self, video_id):\r
        try:\r
            user_object = self._call_nebula_api('/auth/user/', video_id, self._nebula_token, note='Retrieving Zype access token')\r
        except ExtractorError as exc:\r
            # if 401, attempt credential auth and retry\r
            if exc.cause and isinstance(exc.cause, HTTPError) and exc.cause.code == 401:\r
                self._nebula_token = self._retrieve_nebula_auth()\r
                user_object = self._call_nebula_api('/auth/user/', video_id, self._nebula_token, note='Retrieving Zype access token')\r
            else:\r
                raise\r
\r
        access_token = try_get(user_object, lambda x: x['zype_auth_info']['access_token'], compat_str)\r
        if not access_token:\r
            if try_get(user_object, lambda x: x['is_subscribed'], bool):\r
                # TODO: Reimplement the same Zype token polling the Nebula frontend implements\r
                # see https://github.com/ytdl-org/youtube-dl/pull/24805#issuecomment-749231532\r
                raise ExtractorError(\r
                    'Unable to extract Zype access token from Nebula API authentication endpoint. '\r
                    'Open an arbitrary video in a browser with this account to generate a token',\r
                    expected=True)\r
            raise ExtractorError('Unable to extract Zype access token from Nebula API authentication endpoint')\r
        return access_token\r
\r
    def _extract_channel_title(self, video_meta):\r
        # TODO: Implement the API calls giving us the channel list,\r
        # so that we can do the title lookup and then figure out the channel URL\r
        categories = video_meta.get('categories', []) if video_meta else []\r
        # the channel name is the value of the first category\r
        for category in categories:\r
            if category.get('value'):\r
                return category['value'][0]\r
\r
    def _real_initialize(self):\r
        # check cookie jar for valid token\r
        nebula_cookies = self._get_cookies('https://nebula.app')\r
        nebula_cookie = nebula_cookies.get('nebula-auth')\r
        if nebula_cookie:\r
            self.to_screen('Authenticating to Nebula with token from cookie jar')\r
            nebula_cookie_value = compat_urllib_parse_unquote(nebula_cookie.value)\r
            self._nebula_token = self._parse_json(nebula_cookie_value, None).get('apiToken')\r
\r
        # try to authenticate using credentials if no valid token has been found\r
        if not self._nebula_token:\r
            self._nebula_token = self._retrieve_nebula_auth()\r
\r
    def _real_extract(self, url):\r
        display_id = self._match_id(url)\r
        api_key = self._retrieve_zype_api_key(url, display_id)\r
\r
        response = self._call_zype_api('/videos', {'friendly_title': display_id},\r
                                       display_id, api_key, note='Retrieving metadata from Zype')\r
        if len(response.get('response') or []) != 1:\r
            raise ExtractorError('Unable to find video on Zype API')\r
        video_meta = response['response'][0]\r
\r
        video_id = video_meta['_id']\r
        zype_access_token = self._fetch_zype_access_token(display_id)\r
\r
        channel_title = self._extract_channel_title(video_meta)\r
\r
        return {\r
            'id': video_id,\r
            'display_id': display_id,\r
            '_type': 'url_transparent',\r
            'ie_key': 'Zype',\r
            'url': 'https://player.zype.com/embed/%s.html?access_token=%s' % (video_id, zype_access_token),\r
            'title': video_meta.get('title'),\r
            'description': video_meta.get('description'),\r
            'timestamp': parse_iso8601(video_meta.get('published_at')),\r
            'thumbnails': [{\r
                'id': tn.get('name'),  # this appears to be null\r
                'url': tn['url'],\r
                'width': tn.get('width'),\r
                'height': tn.get('height'),\r
            } for tn in video_meta.get('thumbnails', [])],\r
            'duration': video_meta.get('duration'),\r
            'channel': channel_title,\r
            'uploader': channel_title,  # we chose uploader = channel name\r
            # TODO: uploader_url, channel_id, channel_url\r
        }\r