naive replacement of np histograms with boost_histogram histograms
boost_histograms
are somehow much faster than numpy.histograms
. This PR is a very naive replacement of all calls to numpy.histogram
.
There are some places where it would make more sense to keep at least the same bh.axis
object but I went for a quick drop in replacement.
I think boost_histogram can also keep track of the variance internally so these calls to bin with weight**2 should not be necessary. However, I don't want to invest the time to do more than a simple drop in replacement right now.