Generic Factor does not use Public Variables

[F-Lehmann]

Describe the bug Generic Factor treat Public Variables as Uniformly distributed

To Reproduce

import numpy as np
import scalib.attacks 

nc = 2
fg = scalib.attacks.FactorGraph(f'''
NC {nc}
PUB SINGLE A
VAR SINGLE NOT_A
GENERIC SINGLE NOT
PROPERTY NOT(A,NOT_A)
''')

NOT = []

for a in range(nc):
    NOT.append([a,not a])

gen_factors = {
    'NOT':scalib.attacks.GenFactor.sparse_functional(np.array(NOT,dtype=np.uint32))
}

A = 0
bp = scalib.attacks.BPState(fg,1,public_values={'A':A},gen_factors=gen_factors)
bp.bp_acyclic('NOT_A')
result = bp.get_distribution('NOT_A')
print(result)

A = 1
bp = scalib.attacks.BPState(fg,1,public_values={'A':A},gen_factors=gen_factors)
bp.bp_acyclic('NOT_A')
result = bp.get_distribution('NOT_A')
print(result)

Observed behavior

Results are

[0.5 0.5]
[0.5 0.5]

Expected behavior

If the Public whould have been respected the expected result is:

[0.0 1.0]
[1.0 0.0]

Environment (please complete the following information):

• OS: GNU/Linux 11 (bullseye)
  • Python version: 3.8
  • numpy version: 1.23.5
  • SCALib version: 0.5.7.dev18+g5305549
  • How did you install SCALib? build from Xource
  • What CPU do you use? 2x Intel(R) Xeon(R) Gold 5317