Implemented normalisation for weights with sum larger than one. Also …

[BStoelzner]

This fixes the normalisation issue in #146

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (a38c210) 100.00% compared to head (0e2a811) 100.00%.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #185   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           33        33           
  Lines         2279      2282    +3     
=========================================
+ Hits          2279      2282    +3     

Files	Coverage Δ
src/qp/mixmod_pdf.py	100.00% <100.00%> (ø)

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[aimalz]

This is good! It just needs one little unit test to be complete.

[BStoelzner]

@aimalz I've added a unit test, but forgot to let you know. Feel free to take a look :)

[BStoelzner]

The conflict should be resolved now.

[aimalz]

Note the to-do item for an additional test to catch the original way the bug cropped up. I can write it later this week and add it to this PR, or we can make an issue to come back to it some other time.

[aimalz]

This is great! There is an even simpler test that I think is more essential. Recalling that the behavior that first alerted us to the bug was that weights [1,1,1] led to the .pdf() and .cdf() yielding results identical to that for weights of [1,0,0], the most basic unit test for this fix would show that it indeed fixes that: when weights that sum to >1 are provided, we want to check that the .cdf() or .pdf() of the ensemble is not equal to what it would be for just the first element of the mixture model.

[aimalz]

Ack, sorry, I meant this comment as a suggestion that could be made now or in the future but failed to change the status of the review to approved to actually indicate that, my bad!