Proper ZK treatment in plonky2

[Nashtare]

@LindaGuiga I'm opening a draft PR to be able to comment on the code

[Nashtare]

Maybe add a one-liner for the -2

[LindaGuiga]

This PR aims at addressing #1625, based on this note https://eprint.iacr.org/2024/1037.pdf .

• For the batch FRI polynomial, we take a random polynomial with twice the degree of the subgroup, so that we can add a FRI step with arity 2 instead of computing 2 different FRI proofs (for the lower half and higher half of the polynomial, as mentioned in the note).
• For the quotient polynomial chunks, we reduce the degree n of each chunk to n - h, so that we can add to them random polynomials with degree n.
• For the third point in Proper ZK treatment in plonky2 #1625, the current implementation to randomize the wire polynomials seems to follow the guidelines in the paper. Indeed, currently, the degree h is computed as: h_1 = D + num_fri_openings for wire polynomials and h_2 = 2 * D + num_fri_openings for the permutation polynomial Z where D is the extension degree. The differnece between the two values is because the wire polynomials are openings are only opened at zeta, while the Z polynomial is also opened at g*Z. h_1 is added to all wire polynomials, while h_2 is only added to the routed wires. This is in accordance with the prescription of having h >= 2 * (D * n_DEEP + n_FRI) (Eq. 13) in the paper, for the case where the quotient chunks are computed the canonical way and randomized. (Note that the factor 2 in Eq 13 comes from the evaluation at zeta and g*zeta, but we only evaluate at zeta for the wire polynomials, as explained before.)