[ie/digitalconcerthall] Support login with access/refresh tokens (#11571)

Removes broken support for login with email and password
Removes obsolete `prefer_combined_hls` extractor-arg

Closes #11404, Closes #11436
Authored by: bashonly

README.md

@@ -1867,9 +1867,6 @@ #### orfon (orf:on)
 #### bilibili
 * `prefer_multi_flv`: Prefer extracting flv formats over mp4 for older videos that still provide legacy formats
 
-#### digitalconcerthall
-* `prefer_combined_hls`: Prefer extracting combined/pre-merged video and audio HLS formats. This will exclude 4K/HEVC video and lossless/FLAC audio formats, which are only available as split video/audio HLS formats
-
 #### sonylivseries
 * `sort_order`: Episode sort order for series extraction - one of `asc` (ascending, oldest first) or `desc` (descending, newest first). Default is `asc`
 

yt_dlp/extractor/digitalconcerthall.py

@@ -1,7 +1,10 @@
+import time
+
 from .common import InfoExtractor
 from ..networking.exceptions import HTTPError
 from ..utils import (
     ExtractorError,
+    jwt_decode_hs256,
     parse_codecs,
     try_get,
     url_or_none,
@@ -13,9 +16,6 @@
 class DigitalConcertHallIE(InfoExtractor):
     IE_DESC = 'DigitalConcertHall extractor'
     _VALID_URL = r'https?://(?:www\.)?digitalconcerthall\.com/(?P<language>[a-z]+)/(?P<type>film|concert|work)/(?P<id>[0-9]+)-?(?P<part>[0-9]+)?'
-    _OAUTH_URL = 'https://api.digitalconcerthall.com/v2/oauth2/token'
-    _USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.5 Safari/605.1.15'
-    _ACCESS_TOKEN = None
     _NETRC_MACHINE = 'digitalconcerthall'
     _TESTS = [{
         'note': 'Playlist with only one video',
@@ -69,59 +69,157 @@ class DigitalConcertHallIE(InfoExtractor):
         'params': {'skip_download': 'm3u8'},
         'playlist_count': 1,
     }]
+    _LOGIN_HINT = ('Use  --username token --password ACCESS_TOKEN  where ACCESS_TOKEN '
+                   'is the "access_token_production" from your browser local storage')
+    _REFRESH_HINT = 'or else use a "refresh_token" with  --username refresh --password REFRESH_TOKEN'
+    _OAUTH_URL = 'https://api.digitalconcerthall.com/v2/oauth2/token'
+    _CLIENT_ID = 'dch.webapp'
+    _CLIENT_SECRET = '2ySLN+2Fwb'
+    _USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.5 Safari/605.1.15'
+    _OAUTH_HEADERS = {
+        'Accept': 'application/json',
+        'Content-Type': 'application/x-www-form-urlencoded;charset=UTF-8',
+        'Origin': 'https://www.digitalconcerthall.com',
+        'Referer': 'https://www.digitalconcerthall.com/',
+        'User-Agent': _USER_AGENT,
+    }
+    _access_token = None
+    _access_token_expiry = 0
+    _refresh_token = None
 
-    def _perform_login(self, username, password):
+    @property
-        login_token = self._download_json(
+    def _access_token_is_expired(self):
-            self._OAUTH_URL,
+        return self._access_token_expiry - 30 <= int(time.time())
-            None, 'Obtaining token', errnote='Unable to obtain token', data=urlencode_postdata({
+
+    def _set_access_token(self, value):
+        self._access_token = value
+        self._access_token_expiry = traverse_obj(value, ({jwt_decode_hs256}, 'exp', {int})) or 0
+
+    def _cache_tokens(self, /):
+        self.cache.store(self._NETRC_MACHINE, 'tokens', {
+            'access_token': self._access_token,
+            'refresh_token': self._refresh_token,
+        })
+
+    def _fetch_new_tokens(self, invalidate=False):
+        if invalidate:
+            self.report_warning('Access token has been invalidated')
+            self._set_access_token(None)
+
+        if not self._access_token_is_expired:
+            return
+
+        if not self._refresh_token:
+            self._set_access_token(None)
+            self._cache_tokens()
+            raise ExtractorError(
+                'Access token has expired or been invalidated. '
+                'Get a new "access_token_production" value from your browser '
+                f'and try again, {self._REFRESH_HINT}', expected=True)
+
+        # If we only have a refresh token, we need a temporary "initial token" for the refresh flow
+        bearer_token = self._access_token or self._download_json(
+            self._OAUTH_URL, None, 'Obtaining initial token', 'Unable to obtain initial token',
+            data=urlencode_postdata({
                 'affiliate': 'none',
                 'grant_type': 'device',
                 'device_vendor': 'unknown',
-                # device_model 'Safari' gets split streams of 4K/HEVC video and lossless/FLAC audio
+                # device_model 'Safari' gets split streams of 4K/HEVC video and lossless/FLAC audio,
-                'device_model': 'unknown' if self._configuration_arg('prefer_combined_hls') else 'Safari',
+                # but this is no longer effective since actual login is not possible anymore
-                'app_id': 'dch.webapp',
+                'device_model': 'unknown',
+                'app_id': self._CLIENT_ID,
                 'app_distributor': 'berlinphil',
-                'app_version': '1.84.0',
+                'app_version': '1.95.0',
-                'client_secret': '2ySLN+2Fwb',
+                'client_secret': self._CLIENT_SECRET,
-            }), headers={
+            }), headers=self._OAUTH_HEADERS)['access_token']
-                'Accept': 'application/json',
+
-                'Content-Type': 'application/x-www-form-urlencoded;charset=UTF-8',
-                'User-Agent': self._USER_AGENT,
-            })['access_token']
         try:
-            login_response = self._download_json(
+            response = self._download_json(
-                self._OAUTH_URL,
+                self._OAUTH_URL, None, 'Refreshing token', 'Unable to refresh token',
-                None, note='Logging in', errnote='Unable to login', data=urlencode_postdata({
+                data=urlencode_postdata({
-                    'grant_type': 'password',
+                    'grant_type': 'refresh_token',
-                    'username': username,
+                    'refresh_token': self._refresh_token,
-                    'password': password,
+                    'client_id': self._CLIENT_ID,
+                    'client_secret': self._CLIENT_SECRET,
                 }), headers={
-                    'Accept': 'application/json',
+                    **self._OAUTH_HEADERS,
-                    'Content-Type': 'application/x-www-form-urlencoded;charset=UTF-8',
+                    'Authorization': f'Bearer {bearer_token}',
-                    'Referer': 'https://www.digitalconcerthall.com',
-                    'Authorization': f'Bearer {login_token}',
-                    'User-Agent': self._USER_AGENT,
                 })
-        except ExtractorError as error:
+        except ExtractorError as e:
-            if isinstance(error.cause, HTTPError) and error.cause.status == 401:
+            if isinstance(e.cause, HTTPError) and e.cause.status == 401:
-                raise ExtractorError('Invalid username or password', expected=True)
+                self._set_access_token(None)
+                self._refresh_token = None
+                self._cache_tokens()
+                raise ExtractorError('Your tokens have been invalidated', expected=True)
             raise
-        self._ACCESS_TOKEN = login_response['access_token']
+
+        self._set_access_token(response['access_token'])
+        if refresh_token := traverse_obj(response, ('refresh_token', {str})):
+            self.write_debug('New refresh token granted')
+            self._refresh_token = refresh_token
+        self._cache_tokens()
+
+    def _perform_login(self, username, password):
+        self.report_login()
+
+        if username == 'refresh':
+            self._refresh_token = password
+            self._fetch_new_tokens()
+
+        if username == 'token':
+            if not traverse_obj(password, {jwt_decode_hs256}):
+                raise ExtractorError(
+                    f'The access token passed to yt-dlp is not valid. {self._LOGIN_HINT}', expected=True)
+            self._set_access_token(password)
+            self._cache_tokens()
+
+        if username in ('refresh', 'token'):
+            if self.get_param('cachedir') is not False:
+                token_type = 'access' if username == 'token' else 'refresh'
+                self.to_screen(f'Your {token_type} token has been cached to disk. To use the cached '
+                               'token next time, pass  --username cache  along with any password')
+            return
+
+        if username != 'cache':
+            raise ExtractorError(
+                'Login with username and password is no longer supported '
+                f'for this site. {self._LOGIN_HINT}, {self._REFRESH_HINT}', expected=True)
+
+        # Try cached access_token
+        cached_tokens = self.cache.load(self._NETRC_MACHINE, 'tokens', default={})
+        self._set_access_token(cached_tokens.get('access_token'))
+        self._refresh_token = cached_tokens.get('refresh_token')
+        if not self._access_token_is_expired:
+            return
+
+        # Try cached refresh_token
+        self._fetch_new_tokens(invalidate=True)
 
     def _real_initialize(self):
-        if not self._ACCESS_TOKEN:
+        if not self._access_token:
-            self.raise_login_required(method='password')
+            self.raise_login_required(
+                'All content on this site is only available for registered users. '
+                f'{self._LOGIN_HINT}, {self._REFRESH_HINT}', method=None)
 
     def _entries(self, items, language, type_, **kwargs):
         for item in items:
             video_id = item['id']
-            stream_info = self._download_json(
+
-                self._proto_relative_url(item['_links']['streams']['href']), video_id, headers={
+            for should_retry in (True, False):
-                    'Accept': 'application/json',
+                self._fetch_new_tokens(invalidate=not should_retry)
-                    'Authorization': f'Bearer {self._ACCESS_TOKEN}',
+                try:
-                    'Accept-Language': language,
+                    stream_info = self._download_json(
-                    'User-Agent': self._USER_AGENT,
+                        self._proto_relative_url(item['_links']['streams']['href']), video_id, headers={
-                })
+                            'Accept': 'application/json',
+                            'Authorization': f'Bearer {self._access_token}',
+                            'Accept-Language': language,
+                            'User-Agent': self._USER_AGENT,
+                        })
+                    break
+                except ExtractorError as error:
+                    if should_retry and isinstance(error.cause, HTTPError) and error.cause.status == 401:
+                        continue
+                    raise
 
             formats = []
             for m3u8_url in traverse_obj(stream_info, ('channel', ..., 'stream', ..., 'url', {url_or_none})):
@@ -157,7 +255,6 @@ def _real_extract(self, url):
                 'Accept': 'application/json',
                 'Accept-Language': language,
                 'User-Agent': self._USER_AGENT,
-                'Authorization': f'Bearer {self._ACCESS_TOKEN}',
             })
         videos = [vid_info] if type_ == 'film' else traverse_obj(vid_info, ('_embedded', ..., ...))