Fix NumpyVectorArray.__mul__ when other is a NumPy array

[pmli]

No description provided.

[pmli]

There is still the issue with __rmul__. It seems that in array * vectorarray, np.ndarray.__mul__ wants to iterate over vectorarray, which raises AssertionError when the index is out of bounds.

>>> import numpy as np
>>> from pymor.vectorarrays.numpy import NumpyVectorSpace
>>> a = np.array([1, 2])
>>> v = NumpyVectorSpace(3).ones(2)
>>> v * a
NumpyVectorArray(
    [[1. 1. 1.]
     [2. 2. 2.]],
    NumpyVectorSpace(3))
>>> a * v
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/.../src/pymor/vectorarrays/numpy.py", line 69, in __getitem__
    return NumpyVectorArrayView(self, ind)
  File "/.../src/pymor/vectorarrays/numpy.py", line 446, in __init__
    assert array.check_ind(ind)
AssertionError

[sdrave]

@pmli, fixing this will probably require some thought: we cannot prevent ndarray.__mul__ from being called first. Probably we would have to raise IndexError at the very least (no idea what will happen then). I am a bit hesitant to do so, however, since I am not sure how expensive check_ind is and I would prefer not having to call it when running with -O. Maybe it's ok to accept this issue for the release?

[pmli]

I'm fine with postponing this particular issue (I'm also not sure what is the best approach). I removed the failing test.

[codecov]

Codecov Report

Merging #824 into master will decrease coverage by 0.01%.
The diff coverage is 50%.

Impacted Files	Coverage Δ
src/pymor/vectorarrays/numpy.py	83.37% <50%> (-0.68%)	⬇️