weighted KDE returns nan

[Clownshift]

The example0-notebook is actually broken:
I don't know since when, but if weights are given, the kde returns an array full of nan as kde.pdf

To be clear: I believe that this is a problem in the code not in the example.

Not sure exactly what the actual problem is. I will investigate in the next few days (no time today)

[Clownshift]

Still investigating, where exactly the issue comes from...

@jokr91 Do you have a hunch why normalized weights might lead to an issue and non-normalized weights seem to not cause the issue?
(There is no normalization in the python frontend. Afaik, the cpp backend takes care of that. Not sure, if this is generally done, or whether it is being checked, if the weights are already normalized eventually)

[Clownshift]

I think the weights are just always normalized, right?

X-Entropy/package/src/kde.cpp

Line 105 in 732e19d

double histogramNormalizer{calcHistogramNormalizer(weights)};

anyways, I am not sure, whether the histogram itself is still normalized if weights are given...?

[Clownshift]

Side note, when debugging this, mind that currently, we do not check for negative weights (#58 )

[Clownshift]

I suggest to pull #60 first as soon as it is accepted.

[Clownshift]

I started investigating a little bit and I noticed that the example notebook apparently still worked before #51.
@jokr91 Can you maybe check that out?