feat: multiplicity depend on payload size (#670)

* compute actual multiplicity from max_multiplicity

This commit defines a function `min_multiplicity` which can compute the actual multiplicity that will be used from `max_multiplicity` and `payload_len`. The original argument `multiplicity` has been renamed to `max_multiplicity` to indicate that this is an upper bound.

* compute multiplicity

* rename AdvzParams::multiplicity -> max_multiplicity

* new test max_multiplicity

* fix min_multiplicity

* WIP polynomial_internal construct its own eval domain

* WIP

* WIP

* fix: correctness test now passes

* remove println

* fix test

* fix recover_payload

* delete field Advz::eval_domain (yay)

* fix test max_multiplicity

* remove unneeded arg from min_multiplicity

* remove unneeded arg from bytes_to_polys

* tidy bytes_to_polys

* refactor disperse, disperse_precompute into disperse_with_polys_and_commits

* move code from evaluate_polys, assemble_shares into disperse_with_polys_and_commits

* delete method Advz::polynomial

* rename polynomial_internal -> interpolate_polynomial, return VidResult, tidy

* delete field Advz::multi_open_domain, derive it on-the-fly from common.multiplicity (yay)

* min_multiplicity return VidResult, replace panic with error

* refactor eval_domain

* remove stupid comment

* remove more stupid things

* modify test to allow nontrivial multiplicity

* test use nontrivial multiplicity, fails

* fix payload prover with nontrivial multiplicity

* tests use Bn254 instead of Bls12_381 as per #670 (comment)

* uncomment some test code as per #670 (comment)

* address #670 (comment)

* address #670 (comment)

* typo as per #670 (comment)

* delete debugging comment as per #670 (comment)

* remove superfluous log message as per #670 (comment)

* remove superfluous timer as per #670 (comment)

* paranoia as per #670 (comment)

* clarify comment as per #670 (comment)

* stop being such a dumbass

---------

Co-authored-by: Anders Konring <anders.konring@gmail.com>