Brute force method to find the optimal number of harmonics for WaveX, DMWaveX, and CMWaveX

[abhisrkckl]

This method will be slow since this is brute force. I have some ideas for a faster version, but that will be in another PR.

This PR depends on #1802, so it should be merged before merging this.

[abhisrkckl]

Here is an example for achromatic red noise and DM noise

[codecov]

Codecov Report

Attention: Patch coverage is 81.72324% with 70 lines in your changes missing coverage. Please review.

Project coverage is 69.70%. Comparing base (ad88a25) to head (8f43826).
Report is 46 commits behind head on master.

Files	Patch %	Lines
src/pint/utils.py	70.14%	14 Missing and 6 partials ⚠️
src/pint/models/cmwavex.py	88.34%	10 Missing and 9 partials ⚠️
src/pint/models/timing_model.py	76.05%	12 Missing and 5 partials ⚠️
src/pint/noise_analysis.py	82.50%	3 Missing and 11 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1824      +/-   ##
==========================================
+ Coverage   69.53%   69.70%   +0.16%     
==========================================
  Files         108      110       +2     
  Lines       25139    25407     +268     
  Branches     4458     4520      +62     
==========================================
+ Hits        17480    17709     +229     
- Misses       6547     6565      +18     
- Partials     1112     1133      +21     

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

[dlakaplan]

should we discuss having a new requirement more broadly? If this is only needed for a subset of tasks, should it be optional?

[scottransom]

I'm personally fine with adding new requirements if the requirement is pure python and doesn't have a bunch of its own new requirements. nestle seems like it is of that (good) type.

[dlakaplan]

I think this was actually adding joblib?

[dlakaplan]

In particular, I wonder about how using joblib compares to the use of concurrent.futures? I see some discussion online. It seems like we might want to stick with one library for that functionality.

[abhisrkckl]

I will try concurrent.futures.