i(#-1,pos_x,pos_y)=i(#-2,0,x);
These indices look off. Shouldn’t this be
i(#-2,pos_x,pos_y)=i(#-3,0,x);
In other words, the first image on the list has the unrolled pattern of the image (#-3 is leftmost - unroll. #-2 is the hamiltonian. #-1 drives the counters. No?
-meigen only returns eigenvalues. Looks like @David_Tschumperle was thinking eigenvalues/eigenvectors when he wrote the docstrings, but just pulled the eigenvalues through store. Doesn’t look hard to fix…