我试图用下面的代码找到Pypark中2个numpy数组的所有行之间的欧几里德距离的平均值：

def distance_matrix(array_t, array_s):
    from scipy.spatial import distance
    r_count = array_t.shape[0]
    c_count = array_s.shape[0]
    score_mat = np.empty([r_count, 1])
    dist_mat = np.empty([1, c_count])
    for r in range(r_count):
        for c in range(c_count):
            dist_mat[0, c] = distance.euclidean(array_t[r, :], array_s[c, :])
        score_mat[r, 0] = np.mean(dist_mat)
    return score_mat

dist_mat = distance_matrix(df_T_train_test_nparray, df_S_train_test_nparray)

我如何在Pypark中分发这个？