feat: add iso_oscar_singular_[coeff/poly]_ring

[thofma]

Implements the plan we agreed on #976, introducing

    iso_oscar_singular_coeff_ring
    iso_oscar_singular_poly_ring

This is similar to iso_oscar_gap, although here we must have two separate version.

This PR here applies it only as far as we need it to fix #976.

Comments/thoughts:

• There are still many singular_poly_ring and singular_generators uses left, which can (and will with the right input) trigger ambiguity errors. But these might be fixed later.
• @fieker this also fixes FqField if the field is not absolute.
• This opens up the possibility to optimize the coercion for specific fields, so that Singular has a nice field to work with. For example:
  • if $K$ is an arbitrary number field, we could bring it into the form $\mathbf{Q}(\alpha)$,
  • if $R = K[e]/(e^2)$, we could turn $R[x_1,\dotsc,x_n]$ into $K[e,x_1,\dotsc,x_n]/(e^2)$

CC: @fieker @ederc