Change aggregate stats type and added unique method.

[nwlandry]

This PR does the following:

• Makes the type of the statistic aggregate methods (max, min, etc.) the classic float type instead of the new numpy type.
• added the unique method to the statistics interface.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.24%. Comparing base (1fa9839) to head (3e59dc4).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #603      +/-   ##
==========================================
+ Coverage   93.21%   93.24%   +0.02%     
==========================================
  Files          62       62              
  Lines        4733     4735       +2     
==========================================
+ Hits         4412     4415       +3     
+ Misses        321      320       -1     

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

[leotrs]

Could you add a little context? What is this new numpy float type? What versions is it available in? Do we want to pin the numpy version?

[nwlandry]

Could you add a little context? What is this new numpy float type? What versions is it available in? Do we want to pin the numpy version?

@leotrs --- thanks for the review! Numpy 2.0 created breaking changes, which resulted in Issues #544 and #566. Numpy made the decision to use their own float and int datatypes (np.int32, np.int64, etc.) (not sure why), and we made the decision that the traditional Python types are preferable.

[nwlandry]

The way we are handling the Numpy interface now is largely backwards-compatible, so we will also be able to use Numpy versions <2.0.