holopy-developers team mailing list archive

[Tom Dimiduk <tdimiduk@xxxxxxxxxxxxxxxxxxx>]

I have put in place a good first step of of output for fits.

holopy.fit now returns a FitResult object that contains the result of 
the fit as well as all of the fitter internal state (covariance 
matrices, etc).

These objects (and indeed all scatterer and theory objects as well) can 
now be saved into yaml files with:
> holopy.io.save('filename', object)
and loaded with
> l = holopy.io.yaml_io.load('filename')

(holopy.io.load is currently taken as the way to load holograms, hence 
the extra namespace in the name.  I may change something there 
eventually).

I would suggest using this capability to save each frame of fits, 
something kind of like

for frame in frames:
   fitresult = fit(...)
   holopy.io.save('name_including_frame_info', fitresult)

and then also make sure to save your fit script with the results since 
it will have all the info on how you set up the fits.

We can think about doing more automated logging in the future, but this 
should get us back a lot closer to parity with the logging of the old 
stable.

Tom