[extractor] Support multiple archive ids for one video (#4307)

[yt-dlp.git] / yt_dlp / utils.py

diff --git a/yt_dlp/utils.py b/yt_dlp/utils.py
index f0e9ee8c4270f9a4d79d4be5e2d489509d09d5ae..545c027635da2213809c01b746155ca0e3e436d4 100644 (file)
--- a/yt_dlp/utils.py
+++ b/yt_dlp/utils.py
@@ -150,21 +150,6 @@ def random_user_agent():
         'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
 }
 
-KNOWN_EXTENSIONS = (
-    'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
-    'flv', 'f4v', 'f4a', 'f4b',
-    'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
-    'mkv', 'mka', 'mk3d',
-    'avi', 'divx',
-    'mov',
-    'asf', 'wmv', 'wma',
-    '3gp', '3g2',
-    'mp3',
-    'flac',
-    'ape',
-    'wav',
-    'f4f', 'f4m', 'm3u8', 'smil')
-
 # needed for sanitizing filenames in restricted mode
 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
                         itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
@@ -223,6 +208,7 @@ def random_user_agent():
     '%d/%m/%Y',
     '%d/%m/%y',
     '%d/%m/%Y %H:%M:%S',
+    '%d-%m-%Y %H:%M',
 ])
 
 DATE_FORMATS_MONTH_FIRST = list(DATE_FORMATS)
@@ -613,7 +599,9 @@ def sanitize_open(filename, open_mode):
     if filename == '-':
         if sys.platform == 'win32':
             import msvcrt
-            msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
+            # stdout may be any IO stream. Eg, when using contextlib.redirect_stdout
+            with contextlib.suppress(io.UnsupportedOperation):
+                msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
         return (sys.stdout.buffer if hasattr(sys.stdout, 'buffer') else sys.stdout, filename)
 
     for attempt in range(2):
@@ -717,13 +705,13 @@ def sanitize_path(s, force=False):
     return os.path.join(*sanitized_path)
 
 
-def sanitize_url(url):
+def sanitize_url(url, *, scheme='http'):
     # Prepend protocol-less URLs with `http:` scheme in order to mitigate
     # the number of unwanted failures due to missing protocol
     if url is None:
         return
     elif url.startswith('//'):
-        return 'http:%s' % url
+        return f'{scheme}:{url}'
     # Fix some common typos seen so far
     COMMON_TYPOS = (
         # https://github.com/ytdl-org/youtube-dl/issues/15649
@@ -1072,6 +1060,14 @@ def __init__(self, msg, countries=None, **kwargs):
         self.countries = countries
 
 
+class UserNotLive(ExtractorError):
+    """Error when a channel/user is not live"""
+
+    def __init__(self, msg=None, **kwargs):
+        kwargs['expected'] = True
+        super().__init__(msg or 'The channel is not currently live', **kwargs)
+
+
 class DownloadError(YoutubeDLError):
     """Download Error exception.
 
@@ -3666,7 +3662,7 @@ def _match_func(info_dict, incomplete=False):
         if not filters or any(match_str(f, info_dict, incomplete) for f in filters):
             return NO_DEFAULT if interactive and not incomplete else None
         else:
-            video_title = info_dict.get('title') or info_dict.get('id') or 'video'
+            video_title = info_dict.get('title') or info_dict.get('id') or 'entry'
             filter_str = ') | ('.join(map(str.strip, filters))
             return f'{video_title} does not pass filter ({filter_str}), skipping ..'
     return _match_func
@@ -5638,6 +5634,22 @@ def items_(self):
         return self.__dict__.items()
 
 
+MEDIA_EXTENSIONS = Namespace(
+    common_video=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
+    video=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
+    common_audio=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
+    audio=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma'),
+    thumbnails=('jpg', 'png', 'webp'),
+    storyboards=('mhtml', ),
+    subtitles=('srt', 'vtt', 'ass', 'lrc'),
+    manifests=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
+)
+MEDIA_EXTENSIONS.video += MEDIA_EXTENSIONS.common_video
+MEDIA_EXTENSIONS.audio += MEDIA_EXTENSIONS.common_audio
+
+KNOWN_EXTENSIONS = (*MEDIA_EXTENSIONS.video, *MEDIA_EXTENSIONS.audio, *MEDIA_EXTENSIONS.manifests)
+
+
 # Deprecated
 has_certifi = bool(certifi)
 has_websockets = bool(websockets)