[pp/ffmpeg] Fix uncaught error if bad --ffmpeg-location is given (#15104)

Revert 9f77e04c76

Closes #12829
Authored by: bashonly

yt_dlp/downloader/external.py

@@ -457,6 +457,8 @@ class FFmpegFD(ExternalFD):
 
     @classmethod
     def available(cls, path=None):
+        # TODO: Fix path for ffmpeg
+        # Fixme: This may be wrong when --ffmpeg-location is used
         return FFmpegPostProcessor().available
 
     def on_process_started(self, proc, stdin):

yt_dlp/extractor/medaltv.py

@@ -1,14 +1,9 @@
-import re
-
 from .common import InfoExtractor
 from ..utils import (
-    ExtractorError,
-    float_or_none,
-    format_field,
     int_or_none,
-    str_or_none,
-    traverse_obj,
+    url_or_none,
 )
+from ..utils.traversal import traverse_obj
 
 
 class MedalTVIE(InfoExtractor):
@@ -30,25 +25,8 @@ class MedalTVIE(InfoExtractor):
             'view_count': int,
             'like_count': int,
             'duration': 13,
-        },
-    }, {
-        'url': 'https://medal.tv/games/cod-cold-war/clips/2mA60jWAGQCBH',
-        'md5': 'fc7a3e4552ae8993c1c4006db46be447',
-        'info_dict': {
-            'id': '2mA60jWAGQCBH',
-            'ext': 'mp4',
-            'title': 'Quad Cold',
-            'description': 'Medal,https://medal.tv/desktop/',
-            'uploader': 'MowgliSB',
-            'timestamp': 1603165266,
-            'upload_date': '20201020',
-            'uploader_id': '10619174',
-            'thumbnail': 'https://cdn.medal.tv/10619174/thumbnail-34934644-720p.jpg?t=1080p&c=202042&missing',
-            'uploader_url': 'https://medal.tv/users/10619174',
-            'comment_count': int,
-            'view_count': int,
-            'like_count': int,
-            'duration': 23,
+            'thumbnail': r're:https://cdn\.medal\.tv/ugcp/content-thumbnail/.*\.jpg',
+            'tags': ['headshot', 'valorant', '4k', 'clutch', 'mornu'],
         },
     }, {
         'url': 'https://medal.tv/games/cod-cold-war/clips/2um24TWdty0NA',
@@ -57,12 +35,12 @@ class MedalTVIE(InfoExtractor):
             'id': '2um24TWdty0NA',
             'ext': 'mp4',
             'title': 'u tk me i tk u bigger',
-            'description': 'Medal,https://medal.tv/desktop/',
-            'uploader': 'Mimicc',
+            'description': '',
+            'uploader': 'zahl',
             'timestamp': 1605580939,
             'upload_date': '20201117',
             'uploader_id': '5156321',
-            'thumbnail': 'https://cdn.medal.tv/5156321/thumbnail-36787208-360p.jpg?t=1080p&c=202046&missing',
+            'thumbnail': r're:https://cdn\.medal\.tv/source/.*\.png',
             'uploader_url': 'https://medal.tv/users/5156321',
             'comment_count': int,
             'view_count': int,
@@ -70,91 +48,77 @@ class MedalTVIE(InfoExtractor):
             'duration': 9,
         },
     }, {
-        'url': 'https://medal.tv/games/valorant/clips/37rMeFpryCC-9',
-        'only_matching': True,
-    }, {
+        # API requires auth
         'url': 'https://medal.tv/games/valorant/clips/2WRj40tpY_EU9',
+        'md5': '6c6bb6569777fd8b4ef7b33c09de8dcf',
+        'info_dict': {
+            'id': '2WRj40tpY_EU9',
+            'ext': 'mp4',
+            'title': '1v5 clutch',
+            'description': '',
+            'uploader': 'adny',
+            'uploader_id': '6256941',
+            'uploader_url': 'https://medal.tv/users/6256941',
+            'comment_count': int,
+            'view_count': int,
+            'like_count': int,
+            'duration': 25,
+            'thumbnail': r're:https://cdn\.medal\.tv/source/.*\.jpg',
+            'timestamp': 1612896680,
+            'upload_date': '20210209',
+        },
+        'expected_warnings': ['Video formats are not available through API'],
+    }, {
+        'url': 'https://medal.tv/games/valorant/clips/37rMeFpryCC-9',
         'only_matching': True,
     }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        webpage = self._download_webpage(url, video_id, query={'mobilebypass': 'true'})
-
-        hydration_data = self._search_json(
-            r'<script[^>]*>[^<]*\bhydrationData\s*=', webpage,
-            'next data', video_id, end_pattern='</script>', fatal=False)
-
-        clip = traverse_obj(hydration_data, ('clips', ...), get_all=False)
-        if not clip:
-            raise ExtractorError(
-                'Could not find video information.', video_id=video_id)
-
-        title = clip['contentTitle']
-
-        source_width = int_or_none(clip.get('sourceWidth'))
-        source_height = int_or_none(clip.get('sourceHeight'))
-
-        aspect_ratio = source_width / source_height if source_width and source_height else 16 / 9
-
-        def add_item(container, item_url, height, id_key='format_id', item_id=None):
-            item_id = item_id or '%dp' % height
-            if item_id not in item_url:
-                return
-            container.append({
-                'url': item_url,
-                id_key: item_id,
-                'width': round(aspect_ratio * height),
-                'height': height,
-            })
+        content_data = self._download_json(
+            f'https://medal.tv/api/content/{video_id}', video_id,
+            headers={'Accept': 'application/json'})
 
         formats = []
-        thumbnails = []
-        for k, v in clip.items():
-            if not (v and isinstance(v, str)):
-                continue
-            mobj = re.match(r'(contentUrl|thumbnail)(?:(\d+)p)?$', k)
-            if not mobj:
-                continue
-            prefix = mobj.group(1)
-            height = int_or_none(mobj.group(2))
-            if prefix == 'contentUrl':
-                add_item(
-                    formats, v, height or source_height,
-                    item_id=None if height else 'source')
-            elif prefix == 'thumbnail':
-                add_item(thumbnails, v, height, 'id')
-
-        error = clip.get('error')
-        if not formats and error:
-            if error == 404:
-                self.raise_no_formats(
-                    'That clip does not exist.',
-                    expected=True, video_id=video_id)
-            else:
-                self.raise_no_formats(
-                    f'An unknown error occurred ({error}).',
-                    video_id=video_id)
-
-        # Necessary because the id of the author is not known in advance.
-        # Won't raise an issue if no profile can be found as this is optional.
-        author = traverse_obj(hydration_data, ('profiles', ...), get_all=False) or {}
-        author_id = str_or_none(author.get('userId'))
-        author_url = format_field(author_id, None, 'https://medal.tv/users/%s')
+        if m3u8_url := url_or_none(content_data.get('contentUrlHls')):
+            formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', m3u8_id='hls'))
+        if http_url := url_or_none(content_data.get('contentUrl')):
+            formats.append({
+                'url': http_url,
+                'format_id': 'http-source',
+                'ext': 'mp4',
+                'quality': 1,
+            })
+        formats = [fmt for fmt in formats if 'video/privacy-protected-guest' not in fmt['url']]
+        if not formats:
+            # Fallback, does not require auth
+            self.report_warning('Video formats are not available through API, falling back to social video URL')
+            urlh = self._request_webpage(
+                f'https://medal.tv/api/content/{video_id}/socialVideoUrl', video_id,
+                note='Checking social video URL')
+            formats.append({
+                'url': urlh.url,
+                'format_id': 'social-video',
+                'ext': 'mp4',
+                'quality': -1,
+            })
 
         return {
             'id': video_id,
-            'title': title,
             'formats': formats,
-            'thumbnails': thumbnails,
-            'description': clip.get('contentDescription'),
-            'uploader': author.get('displayName'),
-            'timestamp': float_or_none(clip.get('created'), 1000),
-            'uploader_id': author_id,
-            'uploader_url': author_url,
-            'duration': int_or_none(clip.get('videoLengthSeconds')),
-            'view_count': int_or_none(clip.get('views')),
-            'like_count': int_or_none(clip.get('likes')),
-            'comment_count': int_or_none(clip.get('comments')),
+            **traverse_obj(content_data, {
+                'title': ('contentTitle', {str}),
+                'description': ('contentDescription', {str}),
+                'timestamp': ('created', {int_or_none(scale=1000)}),
+                'duration': ('videoLengthSeconds', {int_or_none}),
+                'view_count': ('views', {int_or_none}),
+                'like_count': ('likes', {int_or_none}),
+                'comment_count': ('comments', {int_or_none}),
+                'uploader': ('poster', 'displayName', {str}),
+                'uploader_id': ('poster', 'userId', {str}),
+                'uploader_url': ('poster', 'userId', {str}, filter, {lambda x: x and f'https://medal.tv/users/{x}'}),
+                'tags': ('tags', ..., {str}),
+                'thumbnail': ('thumbnailUrl', {url_or_none}),
+            }),
         }

yt_dlp/extractor/thisoldhouse.py

@@ -1,18 +1,17 @@
-import json
+import urllib.parse
 
 from .brightcove import BrightcoveNewIE
 from .common import InfoExtractor
 from .zype import ZypeIE
 from ..networking import HEADRequest
-from ..networking.exceptions import HTTPError
 from ..utils import (
     ExtractorError,
     filter_dict,
     parse_qs,
     smuggle_url,
-    try_call,
     urlencode_postdata,
 )
+from ..utils.traversal import traverse_obj
 
 
 class ThisOldHouseIE(InfoExtractor):
@@ -77,46 +76,43 @@ class ThisOldHouseIE(InfoExtractor):
         'only_matching': True,
     }]
 
-    _LOGIN_URL = 'https://login.thisoldhouse.com/usernamepassword/login'
-
     def _perform_login(self, username, password):
-        self._request_webpage(
-            HEADRequest('https://www.thisoldhouse.com/insider'), None, 'Requesting session cookies')
-        urlh = self._request_webpage(
-            'https://www.thisoldhouse.com/wp-login.php', None, 'Requesting login info',
-            errnote='Unable to login', query={'redirect_to': 'https://www.thisoldhouse.com/insider'})
+        login_page = self._download_webpage(
+            'https://www.thisoldhouse.com/insider-login', None, 'Downloading login page')
+        hidden_inputs = self._hidden_inputs(login_page)
+        response = self._download_json(
+            'https://www.thisoldhouse.com/wp-admin/admin-ajax.php', None, 'Logging in',
+            headers={
+                'Accept': 'application/json',
+                'X-Requested-With': 'XMLHttpRequest',
+            }, data=urlencode_postdata(filter_dict({
+                'action': 'onebill_subscriber_login',
+                'email': username,
+                'password': password,
+                'pricingPlanTerm': hidden_inputs['pricing_plan_term'],
+                'utm_parameters': hidden_inputs.get('utm_parameters'),
+                'nonce': hidden_inputs['mdcr_onebill_login_nonce'],
+            })))
 
-        try:
-            auth_form = self._download_webpage(
-                self._LOGIN_URL, None, 'Submitting credentials', headers={
-                    'Content-Type': 'application/json',
-                    'Referer': urlh.url,
-                }, data=json.dumps(filter_dict({
-                    **{('client_id' if k == 'client' else k): v[0] for k, v in parse_qs(urlh.url).items()},
-                    'tenant': 'thisoldhouse',
-                    'username': username,
-                    'password': password,
-                    'popup_options': {},
-                    'sso': True,
-                    '_csrf': try_call(lambda: self._get_cookies(self._LOGIN_URL)['_csrf'].value),
-                    '_intstate': 'deprecated',
-                }), separators=(',', ':')).encode())
-        except ExtractorError as e:
-            if isinstance(e.cause, HTTPError) and e.cause.status == 401:
+        message = traverse_obj(response, ('data', 'message', {str}))
+        if not response['success']:
+            if message and 'Something went wrong' in message:
                 raise ExtractorError('Invalid username or password', expected=True)
-            raise
-
-        self._request_webpage(
-            'https://login.thisoldhouse.com/login/callback', None, 'Completing login',
-            data=urlencode_postdata(self._hidden_inputs(auth_form)))
+            raise ExtractorError(message or 'Login was unsuccessful')
+        if message and 'Your subscription is not active' in message:
+            self.report_warning(
+                f'{self.IE_NAME} said your subscription is not active. '
+                f'If your subscription is active, this could be caused by too many sign-ins, '
+                f'and you should instead try using {self._login_hint(method="cookies")[4:]}')
+        else:
+            self.write_debug(f'{self.IE_NAME} said: {message}')
 
     def _real_extract(self, url):
         display_id = self._match_id(url)
-        webpage = self._download_webpage(url, display_id)
-        if 'To Unlock This content' in webpage:
-            self.raise_login_required(
-                'This video is only available for subscribers. '
-                'Note that --cookies-from-browser may not work due to this site using session cookies')
+        webpage, urlh = self._download_webpage_handle(url, display_id)
+        # If login response says inactive subscription, site redirects to frontpage for Insider content
+        if 'To Unlock This content' in webpage or urllib.parse.urlparse(urlh.url).path in ('', '/'):
+            self.raise_login_required('This video is only available for subscribers')
 
         video_url, video_id = self._search_regex(
             r'<iframe[^>]+src=[\'"]((?:https?:)?//(?:www\.)?thisoldhouse\.(?:chorus\.build|com)/videos/zype/([0-9a-f]{24})[^\'"]*)[\'"]',

yt_dlp/postprocessor/ffmpeg.py

@@ -192,7 +192,10 @@ class FFmpegPostProcessor(PostProcessor):
 
     @property
     def available(self):
-        return bool(self._ffmpeg_location.get()) or self.basename is not None
+        # If we return that ffmpeg is available, then the basename property *must* be run
+        # (as doing so has side effects), and its value can never be None
+        # See: https://github.com/yt-dlp/yt-dlp/issues/12829
+        return self.basename is not None
 
     @property
     def executable(self):