Since only the Nationalbibliografin dataset is explicitly CC0-licensed, this is something that has to be checked for every processed entry. If we get a dump of only the NB entries, this won't be necessary, but we don't yet know the final format of the data.