import numpy as np
from scipy.io import loadmat
from neuralNetwork import NeuralNetwork
if __name__ == '__main__':
layer_dims = [400, 25, 10]
data = loadmat('ex4data1.mat')
X, y = (data['X'], data['y'])
y = np.where(y == 10, 0, y).reshape(y.size).astype(float)
data = loadmat('ex4weights.mat')
Theta = [data['Theta1'], data['Theta2']]
theta = np.block([t.reshape(t.size, order='F') for t in Theta])
nn = NeuralNetwork(X, y, Theta, layer_dims, 0)
print('========== Part 1.3: Feedforward and Cost Function ==========')
J = nn.cost(theta)
print('Cost: {0:0.6f} (expected: 0.287629)'.format(J))
print('========== Part 1.4: Regularized Cost Function ==========')
nn.update_lambda(1)
J = nn.cost(theta)
print('Regularized cost: {0:0.6f} (expected: 0.383770)'.format(J))
print('========== Part 2.1: Sigmoid Gradient ==========')
g = nn.sigmoid_gradient(np.array([-1.0, -0.5, 0, 0.5, 1.0]))
e_g = [0.19661193, 0.23500371, 0.250000, 0.23500371, 0.19661193]
print('---------------------')
print(' Actual | Expected')
print('---------------------')
for i, j in zip(g, e_g):
print('{0: .6f} | {1: .6f}'.format(i, j))
