migrate scipy sparse matrices to sparse arrays

docs/source/petsc-interface.rst

@@ -85,7 +85,7 @@ SciPy in the following way:
     import scipy.sparse as sp
 
     indptr, indices, data = petsc_mat.getValuesCSR()
-    scipy_mat = sp.csr_matrix((data, indices, indptr), shape=petsc_mat.getSize())
+    scipy_mat = sp.csr_array((data, indices, indptr), shape=petsc_mat.getSize())
 
 The sparsity pattern may then be straightforwardly plotted
 using matplotlib:

pyop2/scripts/spydump

@@ -39,7 +39,7 @@ plots if two input file names are given."""
 import matplotlib
 import numpy as np
 import pylab
-from scipy.sparse import csr_matrix
+from scipy.sparse import csr_array
 
 COOKIE = 1211216    # from petscmat.h
 IntType = '>i4'     # big-endian, 4 byte integer
@@ -66,7 +66,7 @@ def readmat(filename):
 
 def dump2csr(filename):
     (M, N), (I, J, V) = readmat(filename)
-    return csr_matrix((V, J, I))
+    return csr_array((V, J, I))
 
 
 def compare_dump(files, outfile=None, marker='.', markersize=.5):

tests/firedrake/regression/test_covariance_operator.py

@@ -1,6 +1,6 @@
 import pytest
 import numpy as np
-from scipy.sparse import csr_matrix
+from scipy.sparse import csr_array
 import petsctools
 from firedrake import *
 from firedrake.adjoint import (
@@ -21,7 +21,7 @@ def petsc2numpy_mat(petsc_mat):
     comm = petsc_mat.getComm()
     local_mat = petsc_mat.getRedundantMatrix(
         comm.size, PETSc.COMM_SELF)
-    return csr_matrix(
+    return csr_array(
         local_mat.getValuesCSR()[::-1],
         shape=local_mat.getSize()
     ).todense()