AnsMath sparse matrices and SciPy sparse matrices

This example shows how to get AnsMath sparse matrices into SciPy sparse matrices.

Perform required imports and start PyAnsys

Perform required imports.

from ansys.mapdl.core.examples import vmfiles
import matplotlib.pylab as plt

import ansys.math.core.math as pymath

# Start PyAnsys Math as a service.
mm = pymath.AnsMath()

Get matrices

Run the input file from Verification Manual 153 and then get the stiff (k) matrix from the FULL file.

out = mm._mapdl.input(vmfiles["vm153"])
fullfile = mm._mapdl.jobname + ".full"
k = mm.stiff(fname=fullfile)

Copy AnsMath sparse matrix to SciPy CSR matrix and plot

Copy the AnsMath sparse matrix to a SciPy CSR matrix. Then, plot the graph of the sparse matrix.

pk = k.asarray()
plt.spy(pk, color="orange", markersize=3)
plt.title("AnsMath sparse matrix")
plt.show()

Access vectors

You can access the three vectors that describe this sparse matrix with:

• pk.data

• pk.indices

• pk.indptr

For more information, see SciPy’s class description for the CSR (compressed sparse row) matrix.

print(pk.data[:10])
print(pk.indices[:10])
print(pk.indptr[:10])

[ 0.20292539  0.00378143 -0.05003569 -0.01392161 -0.05003569 -0.01392161
 -0.05142701  0.02406179 -0.05142701  0.20090661]
[ 0  1  4  7 18 21 30 47 50  1]
[ 0  9 17 29 40 47 57 66 72 79]

Create AnsMath sparse matrix from SciPy sparse CSR matrix

Create an AnsMath sparse matrix from a SciPy sparse CSR matrix. Then, transfer the SciPy CSR matrix back to PyAnsys Math.

While this code uses a matrix that was originally within MAPDL, you can load any CSR matrix into PyAnsys Math.

my_mat = mm.matrix(pk, "my_mat", triu=True)
my_mat

AnsMath sparse matrix (63, 63)

Check that the matrices k and my_mat are exactly the sames. The norm of the difference should be zero.

msub = k - my_mat
mm.norm(msub)

0.0

Print CSR representation in PyAnsys Math

Printing the list of objects for the CSR representation in the PyAnsys Math space finds these objects:

• Two SMAT objects, corresponding to the k, MSub matrices, with encrypted names.

• The my_mat SMAT object. Its size is zero because the three vectors are stored separately.

• The three vectors of the CSR my_mat structure: MY_MAT_PTR, MY_MAT_IND, and MY_MAT_DATA.

mm.status()

APDLMATH PARAMETER STATUS-  (      6 PARAMETERS DEFINED)

  Name                   Type            Mem. (MB)       Dims            Workspace

   MY_MAT                SMAT            0.000           [63:63]         1
   VBHKYR                SMAT            0.006           [63:63]         1
   VXDWOF                SMAT            0.006           [63:63]         1
   MY_MAT_DATA           VEC             0.003           344             1
   MY_MAT_IND            VEC             0.000           64              1
   MY_MAT_PTR            VEC             0.001           344             1

Access ID of Python object

To determine which PyAnsys Math object corresponds to which Python object, access the id property of the Python object.

print("name(k)=" + k.id)
print("name(my_mat)=" + my_mat.id)
print("name(msub)=" + msub.id)

name(k)=VXDWOF
name(my_mat)=my_mat
name(msub)=VBHKYR

Stop PyAnsys Math

Stop PyAnsys Math.

mm._mapdl.exit()

/opt/hostedtoolcache/Python/3.10.14/x64/lib/python3.10/site-packages/ansys/mapdl/core/launcher.py:811: UserWarning: The environment variable 'PYMAPDL_START_INSTANCE' is set, hence the argument 'start_instance' is overwritten.
  warnings.warn(