DCDFT FWHM as a measure of period error

[dbenn]

FWHM, the period range near the peak over which the periodogram power is at least half its peak value.

See #255 for background.

I'm keeping these issues separate from an implementation viewpoint because they are very different with respect to how they will be handled within VStar.

[dbenn]

Make FWHM available for a selected peak period via a FWHM button and also in the Models dialog for comparison with the standard error values.

[dbenn]

Show as lower and upper bound of FWHM as per Grant's R code, e.g.

         freq    power      amp      per      freq.se     amp.se    fwhm.lo    fwhm.up
46 0.00299914 156.4582 2.825638 333.4292 2.532946e-06 0.05835603 0.00289931 0.00308787
     per.lo   per.up
46 323.8478 344.9097

[dbenn]

Although a model creation step is not required for FWHM, it could just be simpler from a UI and consistency viewpoint to only show FWHM alongside other error measures in the Models dialog. The only downside is that a model must be created and a period can't be inspected interactively to see FWHM. That could be added later.

[dbenn]

Uncertainty output in models dialog now includes std errors and FWHM:

FWHM lower bound: 0.002653393
FWHM upper bound: 0.003459425
Standard Error of the Frequency: 0.000023577
Standard Error of the Semi-Amplitude: 0.124175477

When harmonics are specified in the model, only FWHM is generated, as per comments in #255.

Compare the above with Grant's R code output for the same data and DCDFT:

         freq    power      amp      per      freq.se    amp.se   fwhm.lo    fwhm.up  per.lo
17 0.00300604 36.41817 2.960859 332.6638 2.357736e-05 0.1241754 0.0026534 0.00345943 289.065
    per.up
17 376.875