import torch
from safetensors.torch import load_file

def load_model(path='model.safetensors'):
    return load_file(path)

def s134(x3, x2, x1, x0, weights):
    """Symmetric function S_{1,3}^4: output 1 iff exactly 1 or 3 of 4 inputs are 1."""
    inp = torch.tensor([float(x3), float(x2), float(x1), float(x0)])

    al1 = int((inp @ weights['at_least_1.weight'].T + weights['at_least_1.bias'] >= 0).item())
    am1 = int((inp @ weights['at_most_1.weight'].T + weights['at_most_1.bias'] >= 0).item())
    al3 = int((inp @ weights['at_least_3.weight'].T + weights['at_least_3.bias'] >= 0).item())
    am3 = int((inp @ weights['at_most_3.weight'].T + weights['at_most_3.bias'] >= 0).item())

    l1 = torch.tensor([float(al1), float(am1), float(al3), float(am3)])
    e1 = int((l1 @ weights['exactly_1.weight'].T + weights['exactly_1.bias'] >= 0).item())
    e3 = int((l1 @ weights['exactly_3.weight'].T + weights['exactly_3.bias'] >= 0).item())

    l2 = torch.tensor([float(e1), float(e3)])
    y = int((l2 @ weights['y.weight'].T + weights['y.bias'] >= 0).item())

    return y

if __name__ == '__main__':
    w = load_model()
    print('Symmetric S_{1,3}^4 (exactly 1 or 3 of 4):')
    for i in range(16):
        x3, x2, x1, x0 = (i >> 3) & 1, (i >> 2) & 1, (i >> 1) & 1, i & 1
        y = s134(x3, x2, x1, x0, w)
        s = x3 + x2 + x1 + x0
        print(f'  {x3}{x2}{x1}{x0} (sum={s}) -> {y}')