Pull request #902: CAS-14195
Merge in CASA/casa6 from CAS-14195 to master
* commit 'd74c9a0ba4ec491506f98bf594b2710a3258278a':
For CAS-14195, tidy up doc info in fringefit.xml
For CAS-14195, improve inline documentation for corrcomb parameter.
changed np.alltrue to np.all for new test case
For CAS-14195, added combine='spw',concatspws=True tests to Fringefit_corrcomb2 test
For CAS-14195, add new Fringefit_corrcomb2 test class to test new corrcomb modes (requires new dataset in the DR)
For CAS-14195, ensure corrcomb='stokes' and 'parallel' duplicate the combined-poln solution is duplicated to both polarizations within the output caltable (as was done for corrcomb='all' previously).
For CAS-14195, in fringefit task and underlying mechanics, add support for correlation combinations: 'stokes' (which does formal Stokes I combination of parallel hands, strictly respecting flags; formerly 'all') and 'parallel' (which does optimally weighted parallel-hand correlation combination for all available unflagged data)
Pull request #901: CAS-14249
Merge in CASA/casa6 from CAS-14249 to master
* commit 'c849f9cdf483096d07fbcdf2d805d0fdae773bd9':
ratesmooth now marked as experimental in the xml
re-toggled on rate smoothing for the test case
Added ratesmooth param to toggle using rates while smoothing rates are now also smoothed by smoothtime
added param passed to ct globals
remove new param from ctglobals
fixed test failure. Still using the data from the CAS-14249 ticket as the test data, Looking to add more test cases using a doctored data set
Changed cycle prediction to use the differences from the current phase value. Changed cycle prediction to iterate over polId instead of ipar. Cleaned up unused variables.
Removed complex conditions since they will never occur in this case. Removed mention of par and parok since only fpar is used. Cleaned up some spacing. Unwrap phase pol uses polid 0 and 1 now before they were set to polId which would lead both to have an id of 1
previous removals were commented out. Now the lines are deleted
removed real/imag conversion to amp/phase and removed channel looping
Changed the FringeJones smoothCT to an override of smooth
cleaned up cout statments in fringefitsmooth and added test case for new behavior
Apply to both pols (ipar==0 and 4)
Cleaned up comments and added phase wrapping for median mode
Added new fringe fit smoothing mode. (Currently just testing with smooth mode 'mean')
