def vanishing_poly(S):
"""
args:
S (m vector)
returns:
p(X) = (X-S1)*(X-S2)*...*(X-Sm)
"""
p = Poly([Fp(1)])
for s in S:
p *= Poly([-s, Fp(1)])
return p
def babysnark_prover(U, n_stmt, CRS, precomp, a):
(m, n) = U.shape
assert n == len(a)
assert len(CRS) == (m + 1) + 2 + (n - n_stmt)
assert len(ROOTS) >= m
# Parse the CRS
taus = CRS[:m + 1]
bUis = CRS[-(n - n_stmt):]
Uis, T = precomp
# Target is the vanishing polynomial
t = vanishing_poly(ROOTS[:m])
# 1. Find the polynomial p(X)
# Convert the basis polynomials Us to coefficient form by interpolating
Us = [Poly.interpolate(ROOTS[:m], U[:, k]) for k in range(n)]
# First compute v
v = Poly([])
for k in range(n):
v += Us[k] * a[k]
# Finally p
p = v * v - 1
# 2. Compute the H term
# Find the polynomial h by dividing p / t
h = p / t
# assert p == h * t
H = evaluate_in_exponent(taus, h)
# 3. Compute the Vw terms, using precomputed Uis
Vw = sum([Uis[k] * a[k] for k in range(n_stmt, n)], G * 0)
# assert G * vw(tau) == Vw
# 4. Compute the Bw terms
Bw = sum([bUis[k - n_stmt] * a[k] for k in range(n_stmt, n)], G * 0)
# assert G * (beta * vw(tau)) == Bw
# V = G * v(tau)
# assert H.pair(T) * GT == V.pair(V)
# print('H:', H)
# print('Bw:', Bw)
# print('Vw:', Vw)
return H, Bw, Vw
def vanishing_poly(omega, n):
# For the special case of evaluating at all n powers of omega,
# the vanishing poly has a special form.
# t(X) = (X-1)(X-omega)....(X-omega^(n-1)) = X^n - 1
return Poly([Fp(-1)] + [Fp(0)] * (n - 1) + [Fp(1)])
def test_two_round(crstype):
numr1players = 6
numr2players = 6
numpolys = 3
crslen = 5
n = numr1players + numr2players + 1
numplayers = n - 1
sends, recvs = get_routing(n)
bcfuncs = gen_blockchain_funcs(sends, recvs)
loop = asyncio.get_event_loop()
public = {}
if crstype == "BabySNARK":
#generate a random alpha, gamma for each party
trapdoors1 = [[random_fp() for j in range(2)]
for i in range(numr1players + 1)]
#generate a random beta for round 2 players
trapdoors2 = [random_fp() for i in range(numr2players)]
trapdoors = trapdoors1 + trapdoors2
polys = [
Poly([random_fp() for j in range(crslen)]) for i in range(numpolys)
]
public['polys'] = polys
crs = init_babysnark(trapdoors[0], crslen)
TestPlayer1 = BabySNARKr1Player
TestPlayer2 = BabySNARKr2Player
bc = Blockchain(sends[0], recvs[0], beacon=beacon_babysnark)
#A proof is unnecessary for the first block as trapdoors are public
pi = None
#set the original CRS
sends[0](0, ["SET_NOWAIT", [pi, crs]])
bctask = loop.create_task(bc.run())
mysleep = loop.create_task(asyncio.sleep(.5))
loop.run_until_complete(mysleep)
options = [{'roles': ['contributor']} for i in range(numplayers)]
players = [None] * numplayers
r1players = [
TestPlayer1(i, trapdoors[i], sends[i], recvs[i], bcfuncs[i],
options[i - 1]) for i in range(1, numr1players + 1)
]
r2players = [
TestPlayer2(i, trapdoors[i], sends[i], recvs[i], bcfuncs[i],
options[i - 1], public)
for i in range(numr1players + 1, numplayers + 1)
]
r1players[numr1players // 2].roles += ["validator"]
r2players[numr2players // 2].roles += ["validator"]
r2players[0].roles += ["roundstarter"]
r1players[-1].roles += ["roundender"]
r2players[-1].roles += ["roundender"]
players = r1players + r2players
for i in range(3, numplayers, 3):
players[i].roles += ["checkpoint"]
#send starting message to first player
sends[0](1, ["INC", None, crs])
playertasks = [loop.create_task(player.run()) for player in players]
for task in [bctask] + playertasks:
task.add_done_callback(print_exception_callback)
loop.run_until_complete(playertasks[-1])
get_block = bcfuncs[1][1]
mytask = loop.create_task(get_block("ALL"))
blocks = loop.run_until_complete(mytask)
assert TestPlayer2.verify_chain(blocks, public)
bctask.cancel()
def _polydemo():
p1 = Poly([1, 2, 3, 4, 5])
def prover_algo(witness, CRS, Qs, p_i_poly, perm_precomp):
print("Starting the Prover Algorithm")
n = int(len(witness) / 3)
assert n & n - 1 == 0, "n must be a power of 2"
id_domain, perm_domain, k, Ss = perm_precomp
# We need to convert between representations to multiply and divide more
# efficiently. In round 3 we have to divide a polynomial of degree 4*n+6
# Not sure if there is a big benefit in keeping the lower order
# representations or if it makes sense to do everything in the highest
# order 8*n right away...
# polys represented with n points
omega = omega_base ** (2 ** 32 // n)
ROOTS = [omega ** i for i in range(n)]
PolyEvalRep = polynomialsEvalRep(Fp, omega, n)
witness = [Fp(i) for i in witness]
witness_poly_a = PolyEvalRep(ROOTS, witness[:n])
witness_poly_b = PolyEvalRep(ROOTS, witness[n:n*2])
witness_poly_c = PolyEvalRep(ROOTS, witness[2*n:])
vanishing_pol_coeff = vanishing_poly(omega, n)
# polys represented with 2*n points
omega2 = omega_base ** (2 ** 32 // (2 * n))
PolyEvalRep2 = polynomialsEvalRep(Fp, omega2, 2 * n)
vanishing_poly_ext = PolyEvalRep2.from_coeffs(vanishing_pol_coeff)
witness_poly_a_ext = PolyEvalRep2.from_coeffs(witness_poly_a.to_coeffs())
witness_poly_b_ext = PolyEvalRep2.from_coeffs(witness_poly_b.to_coeffs())
witness_poly_c_ext = PolyEvalRep2.from_coeffs(witness_poly_c.to_coeffs())
# polys represented with 8*n points
omega3 = omega_base ** (2 ** 32 // (8 * n))
PolyEvalRep3 = polynomialsEvalRep(Fp, omega3, 8 * n)
roots3 = [omega3 ** i for i in range(8 * n)]
S1, S2, S3 = Ss
S1_ext3 = PolyEvalRep3.from_coeffs(S1.to_coeffs())
S2_ext3 = PolyEvalRep3.from_coeffs(S2.to_coeffs())
S3_ext3 = PolyEvalRep3.from_coeffs(S3.to_coeffs())
p_i_poly_ext3 = PolyEvalRep3.from_coeffs(p_i_poly.to_coeffs())
qs_ext3 = [PolyEvalRep3.from_coeffs(q.to_coeffs()) for q in Qs]
q_L_ext3, q_R_ext3, q_M_ext3, q_O_ext3, q_C_ext3 = qs_ext3
# Following the paper, we are using the Fiat Shamir Heuristic. We are
# Simulating 5 rounds of communication with the verifier using a
# random oracle for verifier answers
print("Starting Round 1...")
# Generate "random" blinding scalars
rand_scalars = [random_fp_seeded("1234") for i in range(9)]
# Generate polys with the random scalars as coefficients and convert to
# evaluation representation. These are needed for zero knowledge to
# obfuscate the witness.
a_blind_poly_ext = Poly([rand_scalars[1], rand_scalars[0]])
b_blind_poly_ext = Poly([rand_scalars[3], rand_scalars[2]])
c_blind_poly_ext = Poly([rand_scalars[5], rand_scalars[4]])
a_blind_poly_ext = PolyEvalRep2.from_coeffs(a_blind_poly_ext)
b_blind_poly_ext = PolyEvalRep2.from_coeffs(b_blind_poly_ext)
c_blind_poly_ext = PolyEvalRep2.from_coeffs(c_blind_poly_ext)
# These polynomals have random evaluations at all points except ROOTS where
# they evaluate to the witness
a_poly_ext = a_blind_poly_ext * vanishing_poly_ext + witness_poly_a_ext
b_poly_ext = b_blind_poly_ext * vanishing_poly_ext + witness_poly_b_ext
c_poly_ext = c_blind_poly_ext * vanishing_poly_ext + witness_poly_c_ext
# Evaluate the witness polynomials in the exponent using the CRS
a_eval_exp = evaluate_in_exponent(CRS, a_poly_ext.to_coeffs())
b_eval_exp = evaluate_in_exponent(CRS, b_poly_ext.to_coeffs())
c_eval_exp = evaluate_in_exponent(CRS, c_poly_ext.to_coeffs())
first_output = [a_eval_exp, b_eval_exp, c_eval_exp]
print("Round 1 Finished with output: ", first_output)
print("Starting Round 2...")
# Compute permutation challenges from imaginary verifier
beta = random_fp_seeded(str(first_output) + "0")
gamma = random_fp_seeded(str(first_output) + "1")
# Compute permutation polynomial. z_1 is the blinding summand needed for ZK
z_1 = Poly([rand_scalars[8], rand_scalars[7], rand_scalars[6]])
z_1 = PolyEvalRep2.from_coeffs(z_1)
z_1 = z_1 * vanishing_poly_ext
accumulator_poly_eval = [Fp(1)]
accumulator_poly_eval += [accumulator_factor(n,
i,
witness, beta,
id_domain,
perm_domain,
gamma)
for i in range(n-1)]
accumulator_poly = PolyEvalRep(ROOTS, accumulator_poly_eval)
accumulator_poly = z_1 + PolyEvalRep2.from_coeffs(accumulator_poly.to_coeffs())
second_output = evaluate_in_exponent(CRS, accumulator_poly.to_coeffs())
print("Round 2 Finished with output: ", second_output)
print("Starting Round 3...")
alpha = random_fp_seeded(str(first_output) + str(second_output))
accumulator_poly_ext3 = PolyEvalRep3.from_coeffs(accumulator_poly.to_coeffs())
# The third summand of t has the accumulator poly evaluated at a shift
accumulator_poly_shift_evaluations = eval_poly(accumulator_poly,
roots3,
ROOTS[1])
accumulator_poly_ext3_shift = PolyEvalRep3(roots3,
accumulator_poly_shift_evaluations)
a_poly_ext3 = PolyEvalRep3.from_coeffs(a_poly_ext.to_coeffs())
b_poly_ext3 = PolyEvalRep3.from_coeffs(b_poly_ext.to_coeffs())
c_poly_ext3 = PolyEvalRep3.from_coeffs(c_poly_ext.to_coeffs())
id_poly_1_ext3 = PolyEvalRep3.from_coeffs(Poly([gamma, beta]))
id_poly_2_ext3 = PolyEvalRep3.from_coeffs(Poly([gamma, beta * k]))
id_poly_3_ext3 = PolyEvalRep3.from_coeffs(Poly([gamma, beta * k**2]))
gamma_poly = PolyEvalRep3.from_coeffs(Poly([gamma]))
L_1 = PolyEvalRep(ROOTS, [Fp(1)] + [Fp(0) for i in range(len(ROOTS)-1)])
# Compute quotient polynomial: we are dividing by the vanishing poly which
# has zeros at n roots so we need to do division by swapping to a coset
# first summand should have degree 3n+1, second and third should have
# degree 4n + 5
t = ((a_poly_ext3 * b_poly_ext3 * q_M_ext3) +
(a_poly_ext3 * q_L_ext3) +
(b_poly_ext3 * q_R_ext3) +
(c_poly_ext3 * q_O_ext3) +
q_C_ext3 + p_i_poly_ext3)
t += ((a_poly_ext3 + id_poly_1_ext3) *
(b_poly_ext3 + id_poly_2_ext3) *
(c_poly_ext3 + id_poly_3_ext3) * accumulator_poly_ext3 * alpha)
t -= ((a_poly_ext3 + S1_ext3 * beta + gamma_poly) *
(b_poly_ext3 + S2_ext3 * beta + gamma_poly) *
(c_poly_ext3 + S3_ext3 * beta + gamma_poly) *
accumulator_poly_ext3_shift * alpha)
t += ((accumulator_poly_ext3 - PolyEvalRep3.from_coeffs(Poly([Fp(1)]))) *
PolyEvalRep3.from_coeffs(L_1.to_coeffs()) * alpha ** 2)
t = PolyEvalRep3.divideWithCoset(t.to_coeffs(), vanishing_pol_coeff)
t_coeff = t.coefficients
# We split up the polynomial t in three polynomials so that:
# t= t_lo + x^n*t_mid + t^2n*t_hi
# I found that n has actually to be (n+2) to accomodate the CRS because
# t can be of degree 4n+5
t_lo = Poly(t_coeff[:n+2])
t_mid = Poly(t_coeff[n+2:2*(n+2)])
t_hi = Poly(t_coeff[2*(n+2):])
t_lo_eval_exp = evaluate_in_exponent(CRS, t_lo)
t_mid_eval_exp = evaluate_in_exponent(CRS, t_mid)
t_hi_eval_exp = evaluate_in_exponent(CRS, t_hi)
third_output = [t_lo_eval_exp, t_mid_eval_exp, t_hi_eval_exp]
print("Round 3 Finished with output: ", third_output)
print("Starting Round 4...")
# Compute the evaluation challenge
zeta = random_fp_seeded(str(first_output) +
str(second_output) +
str(third_output))
# Compute the opening evaluations
a_zeta = eval_poly(a_poly_ext, [zeta])[0]
b_zeta = eval_poly(b_poly_ext, [zeta])[0]
c_zeta = eval_poly(c_poly_ext, [zeta])[0]
S_1_zeta = eval_poly(S1, [zeta])[0]
S_2_zeta = eval_poly(S2, [zeta])[0]
t_zeta = eval_poly(PolyEvalRep3.from_coeffs(t), [zeta])[0]
accumulator_shift_zeta = eval_poly(accumulator_poly_ext3,
[zeta * ROOTS[1]])[0]
# Compute linerisation polynomial
r = (q_M_ext3 * a_zeta * b_zeta +
q_L_ext3 * a_zeta +
q_R_ext3 * b_zeta +
q_O_ext3 * c_zeta +
q_C_ext3)
r += (accumulator_poly_ext3 *
(a_zeta + beta * zeta + gamma) *
(b_zeta + beta * k * zeta + gamma) *
(c_zeta + beta * (k ** 2) * zeta + gamma) * alpha)
r -= (S3_ext3 *
(a_zeta + beta * S_1_zeta + gamma) *
(b_zeta + beta * S_2_zeta + gamma) *
alpha * beta * accumulator_shift_zeta)
r += accumulator_poly_ext3 * eval_poly(L_1, [zeta])[0] * alpha ** 2
# Evaluate r at zeta
r_zeta = eval_poly(r, [zeta])[0]
fourth_output = [a_zeta, b_zeta, c_zeta, S_1_zeta, S_2_zeta,
accumulator_shift_zeta, t_zeta, r_zeta]
print("Round 4 Finished with output: ", fourth_output)
print("Starting Round 5...")
# Compute opening challeng
nu = random_fp_seeded(str(first_output) +
str(second_output) +
str(third_output) +
str(fourth_output))
# Compute opening proof polynomial
W_zeta = (PolyEvalRep3.from_coeffs(t_lo) +
PolyEvalRep3.from_coeffs(t_mid) * zeta ** (n+2) +
PolyEvalRep3.from_coeffs(t_hi) * zeta ** (2*(n+2)) -
PolyEvalRep3.from_coeffs(Poly([t_zeta])) +
(r - PolyEvalRep3.from_coeffs(Poly([r_zeta]))) * nu +
(a_poly_ext3 - PolyEvalRep3.from_coeffs(Poly([a_zeta]))) * nu ** 2 +
(b_poly_ext3 - PolyEvalRep3.from_coeffs(Poly([b_zeta]))) * nu ** 3 +
(c_poly_ext3 - PolyEvalRep3.from_coeffs(Poly([c_zeta]))) * nu ** 4 +
(S1_ext3 - PolyEvalRep3.from_coeffs(Poly([S_1_zeta]))) * nu ** 5 +
(S2_ext3 - PolyEvalRep3.from_coeffs(Poly([S_2_zeta]))) * nu ** 6)
W_zeta = W_zeta / PolyEvalRep3.from_coeffs(Poly([-zeta, Fp(1)]))
# Compute the opening proof polynomial
W_zeta_omega = accumulator_poly_ext3 - PolyEvalRep3.from_coeffs(Poly([accumulator_shift_zeta]))
W_zeta_omega = W_zeta_omega / PolyEvalRep3.from_coeffs(Poly([-zeta*ROOTS[1], Fp(1)]))
W_zeta_eval_exp = evaluate_in_exponent(CRS, W_zeta.to_coeffs())
W_zeta_omega_eval_exp = evaluate_in_exponent(CRS, W_zeta_omega.to_coeffs())
fifth_output = [W_zeta_eval_exp, W_zeta_omega_eval_exp]
proof_SNARK = [first_output, second_output, third_output, fifth_output, fourth_output]
print("Round 5 Finished with output: ", fifth_output)
u = random_fp_seeded(str(proof_SNARK))
return proof_SNARK, u