Merge 9de472f6ea into 5904853ae5

Closes #7442
Authored by: bashonly

README.md

@@ -613,8 +613,7 @@ ## Filesystem Options:
     --no-restrict-filenames         Allow Unicode characters, "&" and spaces in
                                     filenames (default)
     --windows-filenames             Force filenames to be Windows-compatible
-    --no-windows-filenames          Make filenames Windows-compatible only if
-                                    using Windows (default)
+    --no-windows-filenames          Sanitize filenames only minimally
     --trim-filenames LENGTH         Limit the filename length (excluding
                                     extension) to the specified number of
                                     characters

test/test_YoutubeDL.py

@@ -707,6 +707,13 @@ def test(tmpl, expected, *, info=None, **params):
         test('%(width)06d.%%(ext)s', 'NA.%(ext)s')
         test('%%(width)06d.%(ext)s', '%(width)06d.mp4')
 
+        # Sanitization options
+        test('%(title3)s', (None, 'foo⧸bar⧹test'))
+        test('%(title5)s', (None, 'aei_A'), restrictfilenames=True)
+        test('%(title3)s', (None, 'foo_bar_test'), windowsfilenames=False, restrictfilenames=True)
+        if sys.platform != 'win32':
+            test('%(title3)s', (None, 'foo⧸bar\\test'), windowsfilenames=False)
+
         # ID sanitization
         test('%(id)s', '_abcd', info={'id': '_abcd'})
         test('%(some_id)s', '_abcd', info={'some_id': '_abcd'})

yt_dlp/YoutubeDL.py

@@ -254,7 +254,9 @@ class YoutubeDL:
     outtmpl_na_placeholder: Placeholder for unavailable meta fields.
     restrictfilenames: Do not allow "&" and spaces in file names
     trim_file_name:    Limit length of filename (extension excluded)
-    windowsfilenames:  Force the filenames to be windows compatible
+    windowsfilenames:  True: Force filenames to be Windows compatible
+                       False: Sanitize filenames only minimally
+                       This option has no effect when running on Windows
     ignoreerrors:      Do not stop on download/postprocessing errors.
                        Can be 'only_download' to ignore only download errors.
                        Default is 'only_download' for CLI, but False for API
@@ -1175,8 +1177,7 @@ def _copy_infodict(info_dict):
 
     def prepare_outtmpl(self, outtmpl, info_dict, sanitize=False):
         """ Make the outtmpl and info_dict suitable for substitution: ydl.escape_outtmpl(outtmpl) % info_dict
-        @param sanitize    Whether to sanitize the output as a filename.
-                           For backward compatibility, a function can also be passed
+        @param sanitize    Whether to sanitize the output as a filename
         """
 
         info_dict.setdefault('epoch', int(time.time()))  # keep epoch consistent once set
@@ -1292,14 +1293,23 @@ def get_value(mdict):
 
         na = self.params.get('outtmpl_na_placeholder', 'NA')
 
-        def filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames')):
+        def filename_sanitizer(key, value, restricted):
             return sanitize_filename(str(value), restricted=restricted, is_id=(
                 bool(re.search(r'(^|[_.])id(\.|$)', key))
                 if 'filename-sanitization' in self.params['compat_opts']
                 else NO_DEFAULT))
 
-        sanitizer = sanitize if callable(sanitize) else filename_sanitizer
-        sanitize = bool(sanitize)
+        if callable(sanitize):
+            self.deprecation_warning('Passing a callable "sanitize" to YoutubeDL.prepare_outtmpl is deprecated')
+        elif not sanitize:
+            pass
+        elif (sys.platform != 'win32' and not self.params.get('restrictfilenames')
+                and self.params.get('windowsfilenames') is False):
+            def sanitize(key, value):
+                return value.replace('/', '\u29F8').replace('\0', '')
+        else:
+            def sanitize(key, value):
+                return filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames'))
 
         def _dumpjson_default(obj):
             if isinstance(obj, (set, LazyList)):
@@ -1382,13 +1392,13 @@ def create_key(outer_mobj):
 
             if sanitize:
                 # If value is an object, sanitize might convert it to a string
-                # So we convert it to repr first
+                # So we manually convert it before sanitizing
                 if fmt[-1] == 'r':
                     value, fmt = repr(value), str_fmt
                 elif fmt[-1] == 'a':
                     value, fmt = ascii(value), str_fmt
                 if fmt[-1] in 'csra':
-                    value = sanitizer(last_field, value)
+                    value = sanitize(last_field, value)
 
             key = '%s\0%s' % (key.replace('%', '%\0'), outer_mobj.group('format'))
             TMPL_DICT[key] = value

yt_dlp/extractor/cbc.py

@@ -151,7 +151,7 @@ def _real_extract(self, url):
 
 class CBCPlayerIE(InfoExtractor):
     IE_NAME = 'cbc.ca:player'
-    _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/|i/caffeine/syndicate/\?mediaId=))(?P<id>(?:\d\.)?\d+)'
+    _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/(?:video/)?|i/caffeine/syndicate/\?mediaId=))(?P<id>(?:\d\.)?\d+)'
     _TESTS = [{
         'url': 'http://www.cbc.ca/player/play/2683190193',
         'md5': '64d25f841ddf4ddb28a235338af32e2c',
@@ -277,6 +277,28 @@ class CBCPlayerIE(InfoExtractor):
             'location': 'Canada',
             'media_type': 'Full Program',
         },
+    }, {
+        'url': 'https://www.cbc.ca/player/play/video/1.7194274',
+        'md5': '188b96cf6bdcb2540e178a6caa957128',
+        'info_dict': {
+            'id': '2334524995812',
+            'ext': 'mp4',
+            'title': '#TheMoment a rare white spirit moose was spotted in Alberta',
+            'description': 'md5:18ae269a2d0265c5b0bbe4b2e1ac61a3',
+            'timestamp': 1714788791,
+            'duration': 77.678,
+            'subtitles': {'eng': [{'ext': 'vtt', 'protocol': 'm3u8_native'}]},
+            'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/201/543/THE_MOMENT.jpg',
+            'uploader': 'CBCC-NEW',
+            'chapters': 'count:0',
+            'upload_date': '20240504',
+            'categories': 'count:3',
+            'series': 'The National',
+            'tags': 'count:15',
+            'creators': ['encoder'],
+            'location': 'Canada',
+            'media_type': 'Excerpt',
+        },
     }, {
         'url': 'cbcplayer:1.7159484',
         'only_matching': True,

yt_dlp/extractor/crunchyroll.py

@@ -53,15 +53,19 @@ def _set_auth_info(self, response):
         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
+        try:
             return self._download_json(
                 f'{self._BASE_URL}/auth/v1/token', None, note=note, errnote=errnote,
-                headers=headers, data=urlencode_postdata(data))
+                headers=headers, data=urlencode_postdata(data), impersonate=True)
         except ExtractorError as error:
             if not isinstance(error.cause, HTTPError) or error.cause.status != 403:
                 raise
+            if target := error.cause.response.extensions.get('impersonate'):
+                raise ExtractorError(f'Got HTTP Error 403 when using impersonate target "{target}"')
             raise ExtractorError(
-                'Request blocked by Cloudflare; navigate to Crunchyroll in your browser, '
+                'Request blocked by Cloudflare. '
+                'Install the required impersonation dependency if possible, '
+                'or else 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)
 

yt_dlp/extractor/youporn.py

@@ -72,15 +72,15 @@ class YouPornIE(InfoExtractor):
             'id': '16290308',
             'age_limit': 18,
             'categories': [],
-            'description': 'md5:00ea70f642f431c379763c17c2f396bc',
+            'description': str,  # TODO: detect/remove SEO spam description in ytdl backport
             'display_id': 'tinderspecial-trailer1',
             'duration': 298.0,
             'ext': 'mp4',
             'upload_date': '20201123',
             'uploader': 'Ersties',
             'tags': [],
-            'thumbnail': 'https://fi1.ypncdn.com/202011/23/16290308/original/8/tinderspecial-trailer1-8(m=eaAaaEPbaaaa).jpg',
-            'timestamp': 1606089600,
+            'thumbnail': r're:https://.+\.jpg',
+            'timestamp': 1606147564,
             'title': 'Tinder In Real Life',
             'view_count': int,
         }
@@ -88,11 +88,17 @@ class YouPornIE(InfoExtractor):
 
     def _real_extract(self, url):
         video_id, display_id = self._match_valid_url(url).group('id', 'display_id')
-        definitions = self._download_json(
-            f'https://www.youporn.com/api/video/media_definitions/{video_id}/', display_id or video_id)
+        self._set_cookie('.youporn.com', 'age_verified', '1')
+        webpage = self._download_webpage(f'https://www.youporn.com/watch/{video_id}', video_id)
+        definitions = self._search_json(r'\bplayervars\s*:', webpage, 'player vars', video_id)['mediaDefinitions']
 
-        def get_format_data(data, f):
-            return traverse_obj(data, lambda _, v: v['format'] == f and url_or_none(v['videoUrl']))
+        def get_format_data(data, stream_type):
+            info_url = traverse_obj(data, (lambda _, v: v['format'] == stream_type, 'videoUrl', {url_or_none}, any))
+            if not info_url:
+                return []
+            return traverse_obj(
+                self._download_json(info_url, video_id, f'Downloading {stream_type} info JSON', fatal=False),
+                lambda _, v: v['format'] == stream_type and url_or_none(v['videoUrl']))
 
         formats = []
         # Try to extract only the actual master m3u8 first, avoiding the duplicate single resolution "master" m3u8s
@@ -123,10 +129,6 @@ def get_format_data(data, f):
             f['height'] = height
             formats.append(f)
 
-        webpage = self._download_webpage(
-            'http://www.youporn.com/watch/%s' % video_id, display_id,
-            headers={'Cookie': 'age_verified=1'})
-
         title = self._html_search_regex(
             r'(?s)<div[^>]+class=["\']watchVideoTitle[^>]+>(.+?)</div>',
             webpage, 'title', default=None) or self._og_search_title(

yt_dlp/options.py

@@ -1356,12 +1356,12 @@ def _alias_callback(option, opt_str, value, parser, opts, nargs):
         help='Allow Unicode characters, "&" and spaces in filenames (default)')
     filesystem.add_option(
         '--windows-filenames',
-        action='store_true', dest='windowsfilenames', default=False,
+        action='store_true', dest='windowsfilenames', default=None,
         help='Force filenames to be Windows-compatible')
     filesystem.add_option(
         '--no-windows-filenames',
         action='store_false', dest='windowsfilenames',
-        help='Make filenames Windows-compatible only if using Windows (default)')
+        help='Sanitize filenames only minimally')
     filesystem.add_option(
         '--trim-filenames', '--trim-file-names', metavar='LENGTH',
         dest='trim_file_name', default=0, type=int,