Don't fail on uncritical nonexisting item

[teutoburg]

This error occurred in AstarVienna/irdb#166 and is fixed by this simple patch. It should be fine (at least for now) to skip any items that cause trouble here...

[codecov]

Codecov Report

Attention: Patch coverage is 57.14286% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 74.90%. Comparing base (6057f73) to head (e488076).

Files	Patch %	Lines
scopesim/effects/fits_headers.py	57.14%	3 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff               @@
##           dev_master     #395      +/-   ##
==============================================
- Coverage       74.92%   74.90%   -0.03%     
==============================================
  Files              56       56              
  Lines            7777     7782       +5     
==============================================
+ Hits             5827     5829       +2     
- Misses           1950     1953       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[hugobuddel]

Yeah.. I deliberately not added such a try/except block before, because if the value cannot be found, then there is a problem (bug) that should be fixed so I rather have it complain loudly.

I think the proper fix would be to have more fine grained fits keywords yaml files, for example per mode-yaml file. That is, there should be an METIS-ELT-fits-keywords yaml files that defines the keywords that AstarVienna/irdb#166 trips over.

That yaml file should then only be included if the ELT itself is included. The yaml file should be in the METIS irdb package and not in the ELT package, because other ELT instruments might define different headers.

I YAGNI'd that yaml refactoring; that is, I postponed it until we needed it. It seems like a single fits keywords yaml file isn't sufficient anymore with AstarVienna/irdb#166, so I think it would be better to do that refactoring now in the context of that PR instead of postponing it further with this PR. Also, this PR will add "unavailable" to the keyword values, which is wrong.

Nevertheless, I'll let you decide on how to proceed, so I'll approve this workaround.