Using A = scipy.sparse.*_array fails and is hard to debug

[wolfgang-noichl]

Minimum example:

import numpy as np
import scipy

from pyamg.gallery import poisson
pyamg.solve(scipy.sparse.csr_array(poisson((100,100))), b=np.random.randn(100*100))

This can easily happen when building A "by hand".

I didn't look into the source so far, but believe this is because some dot product becomes a element-wise multiplication somewhere.
I think checking the type or replacing * by @ or something alike should do the trick? Scipy documentation recommends *_array instead of the matrix classes, making this kind of error more likely.