Add additional fast sky-based calibration to abscal module

[tyler-a-cox]

This PR adds a new sky-based calibration function to the abscal module. There are currently functions for performing traditional sky-based calibration (amp_logcal and phs_logcal), but both are quite slow as they involve matrix inversions. This new function, sky_calibration, implements the StEfCal algorithm (Salvini et al 2014) which scales as nants^2. Below is a time comparison between the existing abscal functions and the new method

And here is a comparison of the residuals of the two methods, showing nearly identical results for the two approaches

I also tested on H6C data using redundantly-calibrated, redundantly-averaged data as the sky model and found that the redundantly-calibrated data and sky-calibrated data agree quite well. Below are the delay-transformed redundantly calibration and sky-calibrated visibilities for a single baseline type.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (b045b6e) 97.16% compared to head (5241467) 97.18%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #892      +/-   ##
==========================================
+ Coverage   97.16%   97.18%   +0.02%     
==========================================
  Files          23       23              
  Lines       10638    10722      +84     
==========================================
+ Hits        10336    10420      +84     
  Misses        302      302              

Flag	Coverage Δ
unittests	97.18% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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