Exception with np.ndarray transpose because it's not row major

[jamesETsmith]

General

np.transpose() and the .T operator often (always in the case of .T) return views of the transposed array. Since it's not actually transposed, it's not row major in memory and will lead to a runtime error on the C++ side.

Details

For now we suggest, copying after the transpose to actually produce row major ordering.

op = fp.PauliOp(...)
psi = np.random.rand(7, 64) # (n_states, dim)
op.apply(psi.T) # fails
op.apply(psi.T.copy()) # succeeds