Reduce filename sanitization when --no-windows-filenames is given

This commit is contained in:
pukkandan 2024-04-01 14:45:27 +05:30
parent a25a424323
commit 0f58c5628a
No known key found for this signature in database
GPG Key ID: 7EEE9E1E817D0A39
3 changed files with 21 additions and 12 deletions

View File

@ -609,8 +609,7 @@ ## 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

View File

@ -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

View File

@ -1348,12 +1348,12 @@ 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,