pi-clans refactor

[Jlh18]

In the branch clans1, PR #164.
Here is a roadmap for (natural) model semantics in the setting of a pi-clan, if we settle on doing things this way:

Polynomials:

• use a different version of mathlib (currently in a PR) to generalize some pullback results so that we can define pullback functors without all pullbacks existing on the category
• refactor the UvPoly results in this pi-clan setting. I will try to write out enough definitions and sorrys for us to continue the rest of the development => Pi-clan polynomials #148

General semantics:

• define semantics in the setting of pi-clans, starting with a universe in the pi-clan. (done for pis and sigmas) pi-clan refactor general semantics #149
• refactor the interpretation used in SynthLean to target Ctx with a pi-clan instead of Psh(Ctx)? (should be very easy, and is totally optional)

Groupoids:

• Define the class of isofibrations in Grpd and show that it is a pi-clan (that it is closed under pushforwards is a fairly involved but self-contained theorem) => Groupoid isofibrations are closed under pushforwards. #147
• Make the previous constructions instantiate the new semantics => FunctorOperation.Sigma --> GroupoidModel.Universe.Sigma #150