coverting matlab circshift function on sparse matrix to python

0
votes

I am trying to convert this matlab code to python which uses circshift on sparse matrix :

    W = [W_A;circshift(W_B, m,2);circshift(W_N, 2*m,2)];

where W_B is a sparse matrix of dim 103 by 309 and looks like this

  (1,1)       10
  (2,1)      -10
  (1,2)      -10
  (2,2)       20
  .............

W_A and W_N are sparse matrices of same dim 103 by 309 but with different values. The output W is a sparse matrix of 309 by 309 which looks something like this :

 (1,1)       10
 (2,1)      -10
 (1,2)      -10
 ................

and m =103

This my python code for the matlab syntax :

 W = np.array([W_A_check, np.roll(W_B_check, m-1, 2), np.roll(W_N_check, 2*m-1, 2)])

I am not sure if this is the correct converted syntax as I am looking for the same kind of output like matlab . Also, the numpy roll function is giving me the axis error below :

   axis 2 is out of bounds for array of dimension 0

How can I use numpy roll correctly to get the same output as matlab. Any sugesstions

pythonnumpymatlabmatrix