Make GroupAlgebra(G, R) an alias for G.algebra(R), and remove its code

[nthiery]

Goal: remove the code for GroupAlgebra, and make GroupAlgebra(G,R) an
alias for G.algebra(R).

Rationale: G.algebra(R) uses categories to share code with semigroup
and monoid algebras. It also uses CombinatorialFreeModule which is
more featurefull than FormalSums. There might be a few features in
GroupAlgebra that is not yet in the categories and that need to be
ported first.

This depends on #10963 for the support of additive group algebras.

Component: categories

Issue created by migration from https://trac.sagemath.org/ticket/11318

[jhpalmieri]

comment:1

See #6670 for a related ticket: it implements group algebras using CombinatorialFreeModule, so it accomplishes some of what's being discussed here. By the way, with G.algebra(R), I don't like the way elements are printed (using "B[...]") — it's not a very standard way to write elements in a group algebra.