Eliminate compute_final_proof

[benjaminsavage]

@danielmasny has pointed out that there is a lot of duplication of logic between compute_proof and compute_final_proof. We could potentially just use one. The caller just needs to be careful to "mask" the "u/v" vectors with randomly distributed points at index 0. By eliminating this method, we do not have any compile time guarantees that it will be used correctly (i.e. the inputs will be masked) but we can follow up with a solution for this later.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.99%. Comparing base (8c48156) to head (ab50440).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1094   +/-   ##
=======================================
  Coverage   90.99%   90.99%           
=======================================
  Files         183      183           
  Lines       26028    25973   -55     
=======================================
- Hits        23683    23633   -50     
+ Misses       2345     2340    -5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[akoshelev]

I am generally against code repetitions even in tests - it is often useful to read them to understand the API. Perhaps a helper function like this can help

       fn zip_chunks<F: PrimeField, U: ArrayLength>(a: &[u128], b: &[u128]) -> Vec<(GenericArray<F, U>, GenericArray<F, U>)> {
        zip(a.chunks(U::USIZE), b.chunks(U::USIZE))
            .map(|(u_chunk, v_chunk)| {
                (
                    GenericArray::generate(|i| F::try_from(u_chunk[i]).unwrap()),
                    GenericArray::generate(|i| F::try_from(v_chunk[i]).unwrap()),
                )
            })
            .collect::<Vec<_>>()
    }

[benjaminsavage]

Great suggestion @akoshelev - this really makes it cleaner and easier to read!