Enable a yaml template to reuse in several config files

[arunkannawadi]

No description provided.

[sidneymau]

there are a lot of paths that we probably want to change in here configs

[arunkannawadi]

We probably want to use ProcessAllTemplates here.

[arunkannawadi]

I have followed the structure adopted in LSSTDESC/imSim to register a template, and find absolute paths more robustly so these can be run from anywhere. @aguinot - for your PR, you just need to update the noise in the template.

[arunkannawadi]

End-to-end testing logs: https://github.com/DukeCosmology/roman_imsim/actions/runs/22410753779/job/64883268287

[aguinot]

Hey,
I am sorry I think I missed some of the discussions here. Could you briefly explain why there are a _template.py and _meta_data.py files?

[arunkannawadi]

The _templates.py can be used to give a handy name to the full path of the template (config/default.yaml). Otherwise, you need to invoke galsim only from the root directory and will fail if your current directory is different. The _templates.py together with _meta_data.py makes it robust. These are also there in LSSTDESC/imSim, but I have a leading underscore because this is not something that a user will ever need to use directly.

[aguinot]

Sorry for the basic questions:

• How do I run the default config? I tried galsim config/default.yaml and I get an error.
• How do I use the registered template?

[arunkannawadi]

I should change the name of the default.yaml. it is incomplete intentionally and is not runnable. You can see how it is used in was.yaml and even in hack.yaml to see how to use a registered template.

[sidneymau]

In the config_template branch of roman_imsim_testdata, the updated hack.yaml uses the template. You can run that like normal (e.g., galsim hack.yaml).

https://github.com/DukeCosmology/roman_imsim_testdata/blob/config_template/hack.yaml

[aguinot]

Suggested change

	sky_subtract: False
	# noise:
	# # These are all by default turned on, but you can turn any of them off if desired:
	# type: RomanNoise
	# mjd: { type: ObSeqData, field: mjd }
	# stray_light: True
	# thermal_background: True
	# reciprocity_failure: True
	# dark_current: True
	# nonlinearity: True
	# ipc: True
	# read_noise: True
	# sky_subtract: True
	noise:
	type: NoNoise

Update to the new NoiseBuilder

[aguinot]

Hmm, I am not sure my suggestion will replace all lines 🤔

[sidneymau]

I am confused what's happening here. I thought the suggestion from earlier was to remove all of the comments?

[aguinot]

Well if you remove the commented stuff you don't see how to specify the noise. If you set the RomanNoise with all other keywords to False you still get background noise. If you want no noise you either have to not set any noise at all or use the explicit NoNoise builder

[aguinot]

I think you will probably need to rebase before merging.

[arunkannawadi]

Waiting on CI to be fixed and rebase on main since #66 went in first

[sidneymau]

the CI is in a working state now -- I think we can take a look at this again once rebased.

Because there isn't a branch of the same name on roman_imsim_testdata, the "Make Reference Image" job might not work in the current implementation... we should verify that and, if that is the case, figure out how to make it work for this type of change... (or maybe we can run the "End-to-end Test" workflow here and it should work ok as long as the submodule is advanced to the right commit on this branch)