[ bug/refactor ] fix spurious dependencies in `Algebra.Morphism.*Monomorphism`

[jamesmckinna]

Recent work in Algebra wrt 'sub'algebras (#2852 , #2854 , #2855 , #2859 , #2863 ...) exposes that, at least for Algebra.Morphism.MagmaMonomorphism, certain properties make spurious appeal to the underlying isMagma property (which exposes ∙-cong), when no appeal is made to ∙-cong... so the dependency in the anonymous module is bogus.

Fixing this in any sensible way, however, will be a breaking change, unless we regard this as a bug fix?

UPDATED: ugh. We need setoid, but not cong for these properties, so arguably, the design is OK as is, but at the same time... ugh.

TODO:

• Algebra.Morphism.MagmaMonomorphism: comm, sel, idem, cancel
• Algebra.Morphism.GroupMonomorphism: ⁻¹-cong
• ... more?
• knock-ons for the whole of Algebra.Module.Morphism.*Monomorphism as a consequence.