Fix downloads in notebooks

[teutoburg]

Following AstarVienna/ScopeSim#446.

For the PSF download in MICADO, we might eventually expand the download facilities in ScopeSim, once pooch has proven to be (more) stable. For now, I used pooch's "script mode" to move AstarVienna/ScopeSim#332 forward.

[teutoburg]

I also fixed the naming of various sim.download_packages(...) along the way. @oczoske , I only saw afterwards that you did the same for at least some notebooks in oc/issue_173, so you might want to rebase your branch once this is merged. Sorry for the complications. In fact, I think I already did the same fixes in #175, which was never quite completed...

[hugobuddel]

Approved, but two comments:

• I don't think it is worthwhile to actually run the MICADO notebook and update the output. The output has not meaningfully changed, so this will add a megabyte or two to the repository for very little benefit. I think it would be better to just change the cells and commit those. Then test them, but discard those changes. (Also the MICADO notebook is broken.. #161.)
• These changes highlight another problem about deprecating the list argument, because now the interface is rather ugly in some cases.

[teutoburg]

• I don't think it is worthwhile to actually run the MICADO notebook and update the output. The output has not meaningfully changed, so this will add a megabyte or two to the repository for very little benefit. I think it would be better to just change the cells and commit those. Then test them, but discard those changes.

Agree. I'm not quite sure why the output would even change at all with the same random seeds. Anyway, I now force-pushed those large but useless "changes" away.

• These changes highlight another problem about deprecating the list argument, because now the interface is rather ugly in some cases.

Disagree on that one. In all cases except the "LSS-YSO" notebook, the only resulting change was the square brackets, which in my personal opinion actually reads clearer without those, but that's more a matter of taste. Either way I think plus or minus a square bracket is negligible.

In the case of the "LSS-YSO" notebook (which I assume was the one you were referring to anyway), the change resulting from the list -> *args modification was from

sim.download_example_data(list(fitsfiles.values()))

to

sim.download_example_data(*fitsfiles.values())

Where the original list(...) was actually redundant, because the previous implementation of download_example_data checked for Iterable (as it should), which also works for dict_values. So the only "real" change is the addition of the *.

The remaining "wrap" of

fitsfiles = dict(zip(fitsfiles, sim.download_example_data(*fitsfiles.values())))

is some "iterable magic" to get the locations of the cached files back into the fitsfiles dict, which would look exactly the same when still passing a list (as the format of the output, list[Path], didn't change).

[hugobuddel]

Good, thank you for fixing yet another one of these!