Improvements to WebbPSF-WFIRST tutorial notebook

[josePhoenix]

Based on feedback from @rpvdmarel (and things I noticed when looking with fresh eyes) I have a few things I / we should look at for notebook improvements:

• [x] Acknowledge warnings and pink text that users may see, tell them they aren't reasons to mistrust the calculation
• [x] In particular, maybe comment out the warning about pysynphot / "local throughput files" for filters, since that's every filter for JWST and WFIRST at this point (I think?)
• There are certain FAQ items that don't have a good home, but should be discoverable from the notebook (what to do if the kernel dies / what that means, how to restart and clear output)
• [x] Explain the panels in the plot shown when display=True in the PSF calculation call. Potentially add colorbars. (Does the last panel update with the broadband PSF in the last step? Or is that the last monochromatic step?)
• [x] Change the EE example to instead use display_profiles and mention the FWHM is marked
• [x] remove GRS_throughput.fits from the data package and filters.tsv
• [x] Change explanation of final download link to say 'hover over this link'... potentially expand the explanation by a sentence or two
• [x] ensure image.interpolation nearest is set
• [x] Display the final PSF when the calculation completes in the notebook interface (with colorbar, labels, etc)
• [x] override the figure size in the notebook interface for the "Display Optical System" button

[mperrin]

agreed re "maybe comment out the warning"

I think we should avoid writing our own version of a 'Jupyter for new users' basics tutorial, and instead have pointers to existing resources such as http://jupyter-notebook-beginner-guide.readthedocs.org/en/latest/

[mperrin]

How much of this do you want to get done in the next 24 hrs for 0.4, versus punting to a hypothetical 0.4.1 in mid-December along with the jupyterhub rollout?

[josePhoenix]

Most, if not all, is pretty straightforward. The only thing I told Roeland was not necessarily straightforward was customizing the appearance of some of the default plots (e.g. adding colorbars to the intermediate optical planes in the calcPSF(display=True) call).

[josePhoenix]

That said, the only thing I think is essential for the release is removing the GRS thing. I think I put that in for testing and never removed it... and it's only going to confuse people.

[josePhoenix]

FYI, I removed the GRS top hat curve in the webbpsf data sources directory and in filters.tsv. I also doubled the nlambda for the wide filter. (We should come up with less-hand-wavey nlambda at some point, but I can't look at that right now.)

[josePhoenix]

Closed by #83.