stpsf.setup_sim_to_match_file yields an error on nirspec MSA data

[eteq]

I ran stpsf.setup_sim_to_match_file(fn) where fn is a fairly vanilla NIRSpec MSA data file. The result was the following:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/var/home/erik/tmp/stpsf/stpsf_src/stpsf/match_data.py", line 52, in setup_sim_to_match_file
    inst.set_position_from_aperture_name(header['APERNAME'])
  File "/var/home/erik/tmp/stpsf/stpsf_src/stpsf/stpsf_core.py", line 1135, in set_position_from_aperture_name
    self.detector_position = ref_in_sci
    ^^^^^^^^^^^^^^^^^^^^^^
  File "/var/home/erik/tmp/stpsf/stpsf_src/stpsf/stpsf_core.py", line 326, in detector_position
    raise ValueError(
ValueError: The maximum allowed detector pixel coordinate value is (X,Y) = (2047, 2047)

While this of course depends on the data file, digging in with the debugger led me to the following snippet (which is not data file-dependent) being the root cause:

nrs = stpsf.NIRSpec()

aperture_name = 'NRS_FULL_MSA'
ap = nrs.siaf[aperture_name]
detname = aperture_name.split('_')[0]
ref_in_tel = ap.V2Ref, ap.V3Ref
nrs_full_aperture = nrs.siaf[nrs.detector + '_FULL']
ref_in_sci = nrs_full_aperture.tel_to_sci(*ref_in_tel)
nrs.detector_position = ref_in_sci

You'll note that seems to mean that any MSA observations will fail for setup_sim_to_match_file right now...

[marcio-melendez]

Thanks for letting us know!
Currently setup_sim_to_match_file doesn't have any specific handling of NIRspec observations, this is something that we can add.
I think the issue comes from pysiaf for this particular aperture, NRS_FULL_MSA. Trying any simple transformation just gives an error:

import pysiaf
instrument = 'NIRSPEC'
siaf = pysiaf.Siaf(instrument) 
nrs_apert = siaf['NRS_FULL_MSA']
nrs_apert.sci_to_tel(200,200)

`RuntimeError: Transformation not supported for this aperture: NRS_FULL_MSA.`

The aperture name defines the position on the detector but in term of the PSF, the default aperture 'NRS1_FULL' for NRS1 (or the equivalent for NRS2), plus the position argument should give you the same result. In other words, performing the manual setup instead of using 'setup_sim_to_match_file'

[eteq]

Gotcha, thanks @marcio-melendez. It would definitely be helpful for me, although I admit caring about the PSF in MSA mode is a pretty niche situation and probably not a ton of people other than me care 🤷 .

I can work around this by manually setting up the instrument as you say, but the one thing I'm hazy on is how/if I should be setting stpsf up to use OPD files from close to the date of the observation. I can probably get away with just copying and pasting from that function, but if you have any guidance about that which would help with that, it would be appreciated!

[marcio-melendez]

Once you do the manual setup for the instrument, you can load an OPD from the time of the observation via load_wss_opd_by_date. It will grab the closest OPD available for that time.

https://stpsf.readthedocs.io/en/latest/jwst_measured_opds.html

[mperrin]

@eteq agreed it's a niche use case but still would be nice to have it not error out. Can you give us an example filename of the file you're having this issue with? Is it a cal file or x1d?

The NIRSpec apertures in SIAF are defined pretty differently from the other instruments. I know how to do the work around for this but it just hasn't come up before for MSA data. Worth fixing though.

[mperrin]

@eteq following up on this from a while ago. Can you give us an example filename of a NIRSpec file you were having this issue with?