Iterating through the numpy array rows

Question

I am trying to take each row of this contrast matrix, perform the operation defined in the denominator variable, add them up, and then divide by the number of rows. When I try to do something like this, I get an "invalid syntax error" (using python 3). This question is from the fMRI univariate analysis context. The X is predefined, you can ignore it. I am just trying to do a dot product with matrices in the denominator variable.

X = simulate_two_predictors(N=350, shift=30, TR=2)

contrast_matrix = np.array([

[0, 1, 0],
[0, 0, 1],
[0, 1, -1]

])

for i in range(contrast_matrix.shape[0]):
     denominator = (contrast_matrix[i,:] @ inv(X.T @ X) @ 
     (contrast_matrix[i,:].T))
     return contrast_matrix.size / denominator

Can you help me with it? I am quite new.

Answer 1

Since I do not know how the output of X simulate_two_predictors look like, I have made a few assumptions.

Correct me if I'm wrong, the return keyword is only valid given that you are returning from a function.

import numpy as np

contrast_matrix = np.array([[0, 1, 0],
                            [0, 0, 1],
                            [0, 1, -1]])

X = simulate_two_predictors(N=350, shift=30, TR=2)

for i in range(contrast_matrix.shape[0]):

    # do something here
    denominator = (contrast_matrix[i,:]) @ (np.linalg.inv(X.T @ X) @ (contrast_matrix[i,:].T))

    # reassign value
    contrast_matrix[i,:] = (contrast_matrix.size / denominator)


print(contrast_matrix)