The reason the urls need to be normalized at all here is because legacy urls from Danbooru 1 don't have their unnormalized counterparts stored anywhere and thus can't be accurately compared with unnormalized urls from Danbooru 2.
The reason the urls need to be normalized at all here is because legacy urls from Danbooru 1 don't have their unnormalized counterparts stored anywhere and thus can't be accurately compared with unnormalized urls from Danbooru 2.