Inherited OA status does not match W3ACT due to lack of canonicalisation

[anjackson]

The inherited licenses via python-w3act do not match those determined by W3ACT, because the URL prefix checks are on 'raw' URLs rather than being canonicalised at all. Hence, we have a target 3109 with a https host URL, and a child target 116215 that has a http scheme. Therefore, the host-level check (see also #37) does not match.

(Raised by @crarugal)

The inheritance logic in W3ACT should be checked and reproduced in this code base, ideally using https://github.com/iipc/urlcanon