Take multiplicity into account in local_spike_counter

[stinebuu]

With this PR, the local_spike_counter will also add the multiplicity so that we get the correct spike count.

This fixes #1100.

[suku248]

Thanks @stinebuu !
I think the test could be simplified as described below.

Description:
Creates a parrot neuron, which receives one spike with multiplicity two from a spike generator. The parrot neuron should then emit a spike with multiplicity two, which should be counted as two spikes by the local spike counter.

@jougs this PR is similar to #1105 . Could you please take a look?

[jougs]

Other than that it looks really terrible with all these cvgidcollections sprinkled throughout the code, I think this is a good fix and the best test for it until the full glory of the GidCollection is unleashed. Many thanks!

[stinebuu]

@suku248 The test in this PR creates 4000 source parrot neurons and 10 target parrot neurons and the source neurons receive spikes from a poisson_generator, not a spike_generator. With the poisson_generator some of the spikes will have multiplicity greater than one. Should I update the test to just have two parrot neurons and use a spike_generator with multiplicity two? It would make the test more explicit. Or would you be ok with keeping the test as is, and having this description:

Description:
Creates 4000 source parrot neurons and 10 target parrot neurons. The source parrots receive spikes from a poisson generator, some of which will have multiplicity greater than one. The source neurons should then emit a spike with multiplicity greater than one to the target neurons, which should be counted as multiple spikes by the local spike counter.

[suku248]

@stinebuu As this is a regression test, I'd prefer it to be as minimalist as possible. A minimalist test that exposes the bug (spikes of parrot neurons with multiplicity > 1 are not counted multiple times) is the one that I described above: One spike generator is connected to one parrot neuron. The spike generator feeds one spike with multiplicity two to the parrot neuron, which then emits a spike with multiplicity two. In the current master, the local spike counter only counts one spike, which is incorrect 🐛. With this PR, the local spike counter should count two spikes 🦋.

[stinebuu]

@suku248 the emojis are making me so happy! I have now simplified the test, which had the added bonus of loosing the scattered cvgidcollections, @jougs 😁