FileNameNormalizer/URLNormalizer does not support files with long file extensions

[pgrunewald]

The current implementation does not properly normalize filenames with long file extensions (i.e. extensions with more than 4 characters). Example for this: base.geojson

When I upload this file, the generated ID for the filename will be base-geosjon instead of base.geojson. When downloading, the filename of the download however is the correct base.geojson. Shouldn't The FileNameNormalizer also cause the dash here?

The proposed solution would be to update the FILENAME_REGEX from:

FILENAME_REGEX = re.compile(r"^(.+)\.(\w{,4})$")

to:

FILENAME_REGEX = re.compile(r"^(.+)\.(\w+)$")

Additionally we should check if FileNameNormalizer is actually used, since it should've be applied in the filename for downloads.

[davisagli]

Makes sense to me.