Primal Hybrid reports one more guessed weight than it should

[ElenaKirshanova]

Hi,
this piece of code

        if zeta:
            # the number of non-zero entries
            h = params.Xs.hamming_weight
            probability = RR(prob_drop(params.n, h, zeta))
            hw = 1
            while hw < min(h, zeta):
                new_search_space = binomial(zeta, hw) * base**hw
                if svp_cost.repeat(ssf(search_space + new_search_space))["rop"] >= bkz_cost["rop"]:
                    break
                search_space += new_search_space
                probability += prob_drop(params.n, h, zeta, fail=hw)
                hw += 1

            svp_cost = svp_cost.repeat(ssf(search_space))

from https://github.com/malb/lattice-estimator/blob/374f07331e6575d1856b2212f3b8aeac96e0295e/estimator/lwe_primal.py#L406C9-L420C1
computes the weight of the guessed secret on zeta coordinates.

The search space computed within the while-loop corresponds to hw-1 where hw is the weight value at the end of the loop. As the result, hw stored in the returned dictionary ret["wt"] = hw should rather be hw-1.

[bencrts]

Hi @ElenaKirshanova -- thanks! I agree, I will do a quick PR to fix this :).