Weighted histograms

[matteobachetti]

Allow our fast histograms to accept weights. Also, uniformize the interface so that we have the same options across all 1d, 2d, 3d, and nd histograms. Use a single function to fail safe to the numpy implementation of each kind of histogram when Numba throws a typing error.
Breaking change: the option for data range is now range like in Numpy. It made no sense to have a different name for the different implementations.

[codecov]

Codecov Report

Merging #764 (5d3c2e1) into main (3fc3c15) will increase coverage by 0.01%.
The diff coverage is 100.00%.

❗ Current head 5d3c2e1 differs from pull request most recent head 4f55b87. Consider uploading reports for the commit 4f55b87 to get more accurate results

@@            Coverage Diff             @@
##             main     #764      +/-   ##
==========================================
+ Coverage   96.14%   96.15%   +0.01%     
==========================================
  Files          43       43              
  Lines        7937     7966      +29     
==========================================
+ Hits         7631     7660      +29     
  Misses        306      306              

Files	Coverage Δ
stingray/utils.py	98.91% <100.00%> (+0.05%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[mgullik]

Hi @matteobachetti,

I'm not an expert on this topic, so please excuse me if my comments are not correct.
I noticed that for the function hist1d_numba_seq_weight you start immediately with an example while for the other histogram definitions (e.g. histogramnd) you wrote a couple of lines of explanation before the example. I think this is very useful, and it would be useful even in hist1d_numba_seq_weight.

I know that a histogram object is pretty self-explanatory, but does it make sense to write a few more lines explaining what is the difference between histogram and histogramdd, for people that are not super familiar with the object of numpy?