Single Tensor Summation using Einsum with Sparse Partials

This is an example of how to properly use the einsum function to compute the summation of a single vector while using sparse partial derivative.

from openmdao.api import Problem
import numpy as np
from omtools.api import Group
import omtools.api as ot


class ExampleTensorSummationSparse(Group):
    def setup(self):
        # Shape of Tensor
        shape3 = (2, 4, 3)
        c = np.arange(24).reshape(shape3)

        # Declaring tensor
        tens = self.declare_input('c', val=c)

        self.register_output(
            'einsum_summ2_sparse_derivs',
            ot.einsum_new_api(tens,
                              operation=[(33, 66, 99)],
                              partial_format='sparse'))


prob = Problem()
prob.model = ExampleTensorSummationSparse()
prob.setup(force_alloc_complex=True)
prob.run_model()

print('c', prob['c'].shape)
print(prob['c'])
print('einsum_summ2_sparse_derivs', prob['einsum_summ2_sparse_derivs'].shape)
print(prob['einsum_summ2_sparse_derivs'])

c (2, 4, 3)
[[[ 0.  1.  2.]
  [ 3.  4.  5.]
  [ 6.  7.  8.]
  [ 9. 10. 11.]]

 [[12. 13. 14.]
  [15. 16. 17.]
  [18. 19. 20.]
  [21. 22. 23.]]]
einsum_summ2_sparse_derivs (1,)
[276.]