Numpy eigensolver has failing code with the latest scipy 1.14.0 due to .H being removed. Apparently it was deprecated in 1.11.0. This affects tests using numpy (minimum) eigensolver. This only shows in >= Python 3.10 as lower versions (3.8 and 3.9) are not supported by this new version and so CI still uses a prior version there of scipy there.
File "/home/runner/work/qiskit-algorithms/qiskit-algorithms/qiskit_algorithms/eigensolvers/numpy_eigensolver.py", line 258, in compute_eigenvalues
eigvals, eigvecs = self._solve(operator)
File "/home/runner/work/qiskit-algorithms/qiskit-algorithms/qiskit_algorithms/eigensolvers/numpy_eigensolver.py", line 143, in _solve
eigval, eigvec = self._solve_sparse(op_matrix, self._k)
File "/home/runner/work/qiskit-algorithms/qiskit-algorithms/qiskit_algorithms/eigensolvers/numpy_eigensolver.py", line 155, in _solve_sparse
if (op_matrix != op_matrix.H).nnz == 0:
AttributeError: 'csr_matrix' object has no attribute 'H'
Numpy eigensolver has failing code with the latest scipy 1.14.0 due to
.H
being removed. Apparently it was deprecated in 1.11.0. This affects tests using numpy (minimum) eigensolver. This only shows in >= Python 3.10 as lower versions (3.8 and 3.9) are not supported by this new version and so CI still uses a prior version there of scipy there.