Trac #32502: make WeierstrassIsomorphism behave (more) like EllipticCurveIsogeny

This ticket is the obvious next step after #32388: We move lots of code
from `EllipticCurveIsogeny` to `EllipticCurveHom`, and implement the
missing `EllipticCurveHom` methods in `WeierstrassIsomorphism`.

This should make `WeierstrassIsomorphism` close to indistinguishable
from `EllipticCurveIsogeny`.

All the really new code is in `weierstrass_morphism.py` and should be
straightforward. The other changes are just moving things around and
some small tweaks that shouldn't cause any observable changes in
behaviour.

(The only "real" dependency is #32388. I've merged the others into this
branch too because it made sense, but in principle this could be rebased
on top of #32388 alone if one of the others gets stalled for some
reason.)

URL: https://trac.sagemath.org/32502
Reported by: lorenz
Ticket author(s): Lorenz Panny
Reviewer(s): John Cremona

Author: Release Manager