Merge 1aec76a000 into 89f535e265

Authored by: bashonly

.github/workflows/core.yml

@@ -53,7 +53,7 @@ jobs:
       with:
         python-version: ${{ matrix.python-version }}
     - name: Install test requirements
-      run: python3 ./devscripts/install_deps.py --include dev --include curl_cffi
+      run: python3 ./devscripts/install_deps.py --include dev --include curl-cffi
     - name: Run tests
       continue-on-error: False
       run: |

yt_dlp/extractor/crunchyroll.py

@@ -24,11 +24,15 @@ class CrunchyrollBaseIE(InfoExtractor):
     _BASE_URL = 'https://www.crunchyroll.com'
     _API_BASE = 'https://api.crunchyroll.com'
     _NETRC_MACHINE = 'crunchyroll'
+    _REFRESH_TOKEN = None
     _AUTH_HEADERS = None
+    _AUTH_EXPIRY = None
     _API_ENDPOINT = None
-    _BASIC_AUTH = None
+    _BASIC_AUTH = 'Basic ' + base64.b64encode(':'.join((
+        't-kdgp2h8c3jub8fn0fq',
+        'yfLDfMfrYvKXh4JXS1LEI2cCqu1v5Wan',
+    )).encode()).decode()
     _IS_PREMIUM = None
-    _CLIENT_ID = ('cr_web', 'noaihdevm_6iyg0a8l0q')
     _LOCALE_LOOKUP = {
         'ar': 'ar-SA',
         'de': 'de-DE',
@@ -43,69 +47,74 @@ class CrunchyrollBaseIE(InfoExtractor):
         'hi': 'hi-IN',
     }
 
-    @property
-    def is_logged_in(self):
-        return bool(self._get_cookies(self._BASE_URL).get('etp_rt'))
+    def _set_auth_info(self, response):
+        CrunchyrollBaseIE._IS_PREMIUM = 'cr_premium' in traverse_obj(response, ('access_token', {jwt_decode_hs256}, 'benefits', ...))
+        CrunchyrollBaseIE._AUTH_HEADERS = {'Authorization': response['token_type'] + ' ' + response['access_token']}
+        CrunchyrollBaseIE._AUTH_EXPIRY = time_seconds(seconds=traverse_obj(response, ('expires_in', {float_or_none}), default=300) - 10)
+
+    def _request_token(self, headers, data, note='Requesting token', errnote='Failed to request token'):
+        try:  # TODO: Add impersonation support here
+            return self._download_json(
+                f'{self._BASE_URL}/auth/v1/token', None, note=note, errnote=errnote,
+                headers=headers, data=urlencode_postdata(data))
+        except ExtractorError as error:
+            if not isinstance(error.cause, HTTPError) or error.cause.status != 403:
+                raise
+            raise ExtractorError(
+                'Request blocked by Cloudflare; navigate to Crunchyroll in your browser, '
+                'then pass the fresh cookies (with --cookies-from-browser or --cookies) '
+                'and your browser\'s User-Agent (with --user-agent)', expected=True)
 
     def _perform_login(self, username, password):
-        if self.is_logged_in:
+        if not CrunchyrollBaseIE._REFRESH_TOKEN:
+            CrunchyrollBaseIE._REFRESH_TOKEN = self.cache.load(self._NETRC_MACHINE, username)
+        if CrunchyrollBaseIE._REFRESH_TOKEN:
             return
 
-        upsell_response = self._download_json(
-            f'{self._API_BASE}/get_upsell_data.0.json', None, 'Getting session id',
-            query={
-                'sess_id': 1,
-                'device_id': 'whatvalueshouldbeforweb',
-                'device_type': 'com.crunchyroll.static',
-                'access_token': 'giKq5eY27ny3cqz',
-                'referer': f'{self._BASE_URL}/welcome/login'
-            })
-        if upsell_response['code'] != 'ok':
-            raise ExtractorError('Could not get session id')
-        session_id = upsell_response['data']['session_id']
-
-        login_response = self._download_json(
-            f'{self._API_BASE}/login.1.json', None, 'Logging in',
-            data=urlencode_postdata({
-                'account': username,
-                'password': password,
-                'session_id': session_id
-            }))
-        if login_response['code'] != 'ok':
-            raise ExtractorError('Login failed. Server message: %s' % login_response['message'], expected=True)
-        if not self.is_logged_in:
-            raise ExtractorError('Login succeeded but did not set etp_rt cookie')
-
-    def _update_auth(self):
-        if CrunchyrollBaseIE._AUTH_HEADERS and CrunchyrollBaseIE._AUTH_REFRESH > time_seconds():
-            return
-
-        if not CrunchyrollBaseIE._BASIC_AUTH:
-            cx_api_param = self._CLIENT_ID[self.is_logged_in]
-            self.write_debug(f'Using cxApiParam={cx_api_param}')
-            CrunchyrollBaseIE._BASIC_AUTH = 'Basic ' + base64.b64encode(f'{cx_api_param}:'.encode()).decode()
-
-        auth_headers = {'Authorization': CrunchyrollBaseIE._BASIC_AUTH}
-        if self.is_logged_in:
-            grant_type = 'etp_rt_cookie'
-        else:
-            grant_type = 'client_id'
-            auth_headers['ETP-Anonymous-ID'] = uuid.uuid4()
         try:
-            auth_response = self._download_json(
-                f'{self._BASE_URL}/auth/v1/token', None, note=f'Authenticating with grant_type={grant_type}',
-                headers=auth_headers, data=f'grant_type={grant_type}'.encode())
+            login_response = self._request_token(
+                headers={'Authorization': self._BASIC_AUTH}, data={
+                    'username': username,
+                    'password': password,
+                    'grant_type': 'password',
+                    'scope': 'offline_access',
+                }, note='Logging in', errnote='Failed to log in')
         except ExtractorError as error:
-            if isinstance(error.cause, HTTPError) and error.cause.status == 403:
-                raise ExtractorError(
-                    'Request blocked by Cloudflare; navigate to Crunchyroll in your browser, '
-                    'then pass the fresh cookies (with --cookies-from-browser or --cookies) '
-                    'and your browser\'s User-Agent (with --user-agent)', expected=True)
+            if isinstance(error.cause, HTTPError) and error.cause.status == 401:
+                raise ExtractorError('Invalid username and/or password', expected=True)
             raise
 
-        CrunchyrollBaseIE._IS_PREMIUM = 'cr_premium' in traverse_obj(auth_response, ('access_token', {jwt_decode_hs256}, 'benefits', ...))
-        CrunchyrollBaseIE._AUTH_HEADERS = {'Authorization': auth_response['token_type'] + ' ' + auth_response['access_token']}
-        CrunchyrollBaseIE._AUTH_REFRESH = time_seconds(seconds=traverse_obj(auth_response, ('expires_in', {float_or_none}), default=300) - 10)
+        CrunchyrollBaseIE._REFRESH_TOKEN = login_response['refresh_token']
+        self.cache.store(self._NETRC_MACHINE, username, CrunchyrollBaseIE._REFRESH_TOKEN)
+        self._set_auth_info(login_response)
+
+    def _update_auth(self):
+        if CrunchyrollBaseIE._AUTH_HEADERS and CrunchyrollBaseIE._AUTH_EXPIRY > time_seconds():
+            return
+
+        auth_headers = {'Authorization': self._BASIC_AUTH}
+        if CrunchyrollBaseIE._REFRESH_TOKEN:
+            data = {
+                'refresh_token': CrunchyrollBaseIE._REFRESH_TOKEN,
+                'grant_type': 'refresh_token',
+                'scope': 'offline_access',
+            }
+        else:
+            data = {'grant_type': 'client_id'}
+            auth_headers['ETP-Anonymous-ID'] = uuid.uuid4()
+        try:
+            auth_response = self._request_token(auth_headers, data)
+        except ExtractorError as error:
+            username, password = self._get_login_info()
+            if not username or not isinstance(error.cause, HTTPError) or error.cause.status != 400:
+                raise
+            self.to_screen('Refresh token has expired. Re-logging in')
+            CrunchyrollBaseIE._REFRESH_TOKEN = None
+            self.cache.store(self._NETRC_MACHINE, username, None)
+            self._perform_login(username, password)
+            return
+
+        self._set_auth_info(auth_response)
 
     def _locale_from_language(self, language):
         config_locale = self._configuration_arg('metadata', ie_key=CrunchyrollBetaIE, casesense=True)
@@ -168,7 +177,8 @@ class CrunchyrollBaseIE(InfoExtractor):
         self._update_auth()
         stream_response = self._download_json(
             f'https://cr-play-service.prd.crunchyrollsvc.com/v1/{identifier}/console/switch/play',
-            display_id, note='Downloading stream info', headers=CrunchyrollBaseIE._AUTH_HEADERS)
+            display_id, note='Downloading stream info', errnote='Failed to download stream info',
+            headers=CrunchyrollBaseIE._AUTH_HEADERS)
 
         available_formats = {'': ('', '', stream_response['url'])}
         for hardsub_lang, stream in traverse_obj(stream_response, ('hardSubs', {dict.items}, lambda _, v: v[1]['url'])):
@@ -383,9 +393,9 @@ class CrunchyrollBetaIE(CrunchyrollCmsBaseIE):
 
         if not self._IS_PREMIUM and traverse_obj(response, (f'{object_type}_metadata', 'is_premium_only')):
             message = f'This {object_type} is for premium members only'
-            if self.is_logged_in:
+            if CrunchyrollBaseIE._REFRESH_TOKEN:
                 raise ExtractorError(message, expected=True)
-            self.raise_login_required(message)
+            self.raise_login_required(message, method='password')
 
         result['formats'], result['subtitles'] = self._extract_stream(internal_id)
 
@@ -575,9 +585,9 @@ class CrunchyrollMusicIE(CrunchyrollBaseIE):
 
         if not self._IS_PREMIUM and response.get('isPremiumOnly'):
             message = f'This {response.get("type") or "media"} is for premium members only'
-            if self.is_logged_in:
+            if CrunchyrollBaseIE._REFRESH_TOKEN:
                 raise ExtractorError(message, expected=True)
-            self.raise_login_required(message)
+            self.raise_login_required(message, method='password')
 
         result = self._transform_music_response(response)
         result['formats'], _ = self._extract_stream(f'music/{internal_id}', internal_id)

yt_dlp/extractor/manyvids.py

@@ -1,14 +1,6 @@
-import re
+import json
 
 from .common import InfoExtractor
-from ..utils import (
-    determine_ext,
-    extract_attributes,
-    int_or_none,
-    str_to_int,
-    url_or_none,
-    urlencode_postdata,
-)
 
 
 class ManyVidsIE(InfoExtractor):
@@ -16,27 +8,24 @@ class ManyVidsIE(InfoExtractor):
     _VALID_URL = r'(?i)https?://(?:www\.)?manyvids\.com/video/(?P<id>\d+)'
     _TESTS = [{
         # preview video
-        'url': 'https://www.manyvids.com/Video/133957/everthing-about-me/',
-        'md5': '03f11bb21c52dd12a05be21a5c7dcc97',
+        'url': 'https://www.manyvids.com/Video/1856601/Horny-Asian-Bunny-Needs-Cock/',
         'info_dict': {
-            'id': '133957',
+            'id': '1856601',
             'ext': 'mp4',
-            'title': 'everthing about me (Preview)',
-            'uploader': 'ellyxxix',
-            'view_count': int,
+            'title': 'Horny Asian Bunny Needs Cock',
+            'description': "This naughty little Asian bunny is bored and horny. Feed her something hard and big.",
+            'uploader': 'nicoledoshi',
             'like_count': int,
         },
     }, {
         # full video
-        'url': 'https://www.manyvids.com/Video/935718/MY-FACE-REVEAL/',
-        'md5': 'bb47bab0e0802c2a60c24ef079dfe60f',
+        'url': 'https://www.manyvids.com/Video/4957054/Latex-Dress-JOI/',
         'info_dict': {
-            'id': '935718',
+            'id': '4957054',
             'ext': 'mp4',
-            'title': 'MY FACE REVEAL',
-            'description': 'md5:ec5901d41808b3746fed90face161612',
-            'uploader': 'Sarah Calanthe',
-            'view_count': int,
+            'title': 'Latex Dress JOI',
+            'description': "My new latex dress feels so good on my skin! Just wearing it for you makes me horny. Jerk your cock while I show off my latex-clad body from all angles, rubbing my hands over my body and describing how good it feels. I can't help but touch my pussy, and I give you a countdown when I'm close so we can cum together for my new latex dress!",
+            'uploader': "Kylee Nash",
             'like_count': int,
         },
     }]
@@ -44,119 +33,33 @@ class ManyVidsIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        real_url = 'https://www.manyvids.com/video/%s/gtm.js' % (video_id, )
-        try:
-            webpage = self._download_webpage(real_url, video_id)
-        except Exception:
-            # probably useless fallback
-            webpage = self._download_webpage(url, video_id)
+        infos_url = f'https://www.manyvids.com/bff/store/video/{video_id}'
+        info_webpage = self._download_webpage(infos_url, video_id, expected_status=200)
+        info: dict = json.loads(info_webpage).get('data')
 
-        info = self._search_regex(
-            r'''(<div\b[^>]*\bid\s*=\s*(['"])pageMetaDetails\2[^>]*>)''',
-            webpage, 'meta details', default='')
-        info = extract_attributes(info)
+        download_url = f'https://www.manyvids.com/bff/store/video/{video_id}/private'
+        download_webpage = self._download_webpage(download_url, video_id, expected_status=200)
+        download: dict = json.loads(download_webpage).get('data')
 
-        player = self._search_regex(
-            r'''(<div\b[^>]*\bid\s*=\s*(['"])rmpPlayerStream\2[^>]*>)''',
-            webpage, 'player details', default='')
-        player = extract_attributes(player)
+        is_free: bool = info["isFree"]
 
-        video_urls_and_ids = (
-            (info.get('data-meta-video'), 'video'),
-            (player.get('data-video-transcoded'), 'transcoded'),
-            (player.get('data-video-filepath'), 'filepath'),
-            (self._og_search_video_url(webpage, secure=False, default=None), 'og_video'),
-        )
-
-        def txt_or_none(s, default=None):
-            return (s.strip() or default) if isinstance(s, str) else default
-
-        uploader = txt_or_none(info.get('data-meta-author'))
-
-        def mung_title(s):
-            if uploader:
-                s = re.sub(r'^\s*%s\s+[|-]' % (re.escape(uploader), ), '', s)
-            return txt_or_none(s)
-
-        title = (
-            mung_title(info.get('data-meta-title'))
-            or self._html_search_regex(
-                (r'<span[^>]+class=["\']item-title[^>]+>([^<]+)',
-                 r'<h2[^>]+class=["\']h2 m-0["\'][^>]*>([^<]+)'),
-                webpage, 'title', default=None)
-            or self._html_search_meta(
-                'twitter:title', webpage, 'title', fatal=True))
-
-        title = re.sub(r'\s*[|-]\s+ManyVids\s*$', '', title) or title
-
-        if any(p in webpage for p in ('preview_videos', '_preview.mp4')):
-            title += ' (Preview)'
-
-        mv_token = self._search_regex(
-            r'data-mvtoken=(["\'])(?P<value>(?:(?!\1).)+)\1', webpage,
-            'mv token', default=None, group='value')
-
-        if mv_token:
-            # Sets some cookies
-            self._download_webpage(
-                'https://www.manyvids.com/includes/ajax_repository/you_had_me_at_hello.php',
-                video_id, note='Setting format cookies', fatal=False,
-                data=urlencode_postdata({
-                    'mvtoken': mv_token,
-                    'vid': video_id,
-                }), headers={
-                    'Referer': url,
-                    'X-Requested-With': 'XMLHttpRequest'
-                })
-
-        formats = []
-        for v_url, fmt in video_urls_and_ids:
-            v_url = url_or_none(v_url)
-            if not v_url:
-                continue
-            if determine_ext(v_url) == 'm3u8':
-                formats.extend(self._extract_m3u8_formats(
-                    v_url, video_id, 'mp4', entry_protocol='m3u8_native',
-                    m3u8_id='hls'))
-            else:
-                formats.append({
-                    'url': v_url,
-                    'format_id': fmt,
-                })
-
-        self._remove_duplicate_formats(formats)
-
-        for f in formats:
-            if f.get('height') is None:
-                f['height'] = int_or_none(
-                    self._search_regex(r'_(\d{2,3}[02468])_', f['url'], 'video height', default=None))
-            if '/preview/' in f['url']:
-                f['format_id'] = '_'.join(filter(None, (f.get('format_id'), 'preview')))
-                f['preference'] = -10
-            if 'transcoded' in f['format_id']:
-                f['preference'] = f.get('preference', -1) - 1
-
-        def get_likes():
-            likes = self._search_regex(
-                r'''(<a\b[^>]*\bdata-id\s*=\s*(['"])%s\2[^>]*>)''' % (video_id, ),
-                webpage, 'likes', default='')
-            likes = extract_attributes(likes)
-            return int_or_none(likes.get('data-likes'))
-
-        def get_views():
-            return str_to_int(self._html_search_regex(
-                r'''(?s)<span\b[^>]*\bclass\s*=["']views-wrapper\b[^>]+>.+?<span\b[^>]+>\s*(\d[\d,.]*)\s*</span>''',
-                webpage, 'view count', default=None))
+        title = info["title"]
+        if title and is_free is False:
+            title += " (Preview)"
 
+        if is_free:
+            download_url = download["filepath"]
+        else:
+            download_url = download.get("teaser", {}).get("filepath")
+            if download_url is None:
+                raise ValueError('this video has no preview')
+                # They are rare, but they exist. You can find such videos using the following link:
+                # https://www.manyvids.com/Vids/?content_type=1,2,3&other=blocked&search_type=0&sort=13&page=1
         return {
             'id': video_id,
+            'url': download_url,
             'title': title,
-            'formats': formats,
-            'description': txt_or_none(info.get('data-meta-description')),
-            'uploader': txt_or_none(info.get('data-meta-author')),
-            'thumbnail': (
-                url_or_none(info.get('data-meta-image'))
-                or url_or_none(player.get('data-video-screenshot'))),
-            'view_count': get_views(),
-            'like_count': get_likes(),
+            'description': info.get("description") or None,
+            'uploader': info["model"]["displayName"],
+            'like_count': int(info["likes"]),
         }