It's nice to see a modern alternative for URL parsing, specially as the java.net.URL equals does not work as expected and the java.net.URI is outdated.
By the way, I took a look at your test cases and you seem to test for some crazy edge cases, like Chinese characters in the host part of the URL and even smileys. Are these valid URLs in general?
As I understand it, they are ok, but it depends on the point at which you process the IDN (ToASCII and ToUnicode), this is specified in RFC 3490 (https://tools.ietf.org/html/rfc3490).
By the way, I took a look at your test cases and you seem to test for some crazy edge cases, like Chinese characters in the host part of the URL and even smileys. Are these valid URLs in general?