L1 reffile usage

[braingram]

I'm opening this as a romancal issue but it extends to romanisim.

Looking at the L1 files for B21 they contain rather extensive meta.ref_file information:

│ │ ├─ref_file (dict) # Reference File Information
│ │ │ ├─apcorr (str): ? # Aperture Correction Reference File Schema
│ │ │ ├─area (str): ? # Pixel Area Reference File Information
│ │ │ ├─crds (dict) # Calibration Reference Data System (CRDS) Information
│ │ │ │ ├─context (str): roman_0046.pmap # CRDS Context
│ │ │ │ └─version (str): 13.1.3 # CRDS Software Version
│ │ │ ├─dark (str): crds://roman_wfi_dark_0364.asdf # Dark Reference File Information
│ │ │ ├─darkdecaysignal (str): crds://roman_wfi_darkdecaysignal_0001.asdf
│ │ │ ├─distortion (str): crds://roman_wfi_distortion_0014.asdf # Distortion Reference File Information
│ │ │ ├─epsf (str): ? # ePSF Reference File Information
│ │ │ ├─flat (str): crds://roman_wfi_flat_0166.asdf # Flat Reference File Information
│ │ │ ├─gain (str): crds://roman_wfi_gain_0022.asdf # Gain Reference Rile Information
│ │ │ ├─integralnonlinearity (str): crds://roman_wfi_integralnonlinearity_0003.asdf
│ │ │ ├─inverse_linearity (str): ? # Inverse Linearity Reference File Information
│ │ │ ├─inverselinearity (str): crds://roman_wfi_inverselinearity_0042.asdf
│ │ │ ├─linearity (str): crds://roman_wfi_linearity_0039.asdf # Linearity Reference File Information
│ │ │ ├─mask (str): ? # Bad Pixel Mask Reference File Information
│ │ │ ├─photom (str): ? # Photometry Reference File Information
│ │ │ ├─readnoise (str): crds://roman_wfi_readnoise_0024.asdf # Read Noise Reference File Information
│ │ │ ├─refpix (str): ? # Reference Pixel Reference File Information
│ │ │ └─saturation (str): crds://roman_wfi_saturation_0031.asdf # Saturation Reference File Information

The schema describes meta.ref_file.dark as:

      Name of the calibration reference file used to correct
      for dark current signal in the science data.

This is misleading because for instance no dark step has been run on this data. This is further complicated by romancal propagating the L1 meta.ref_file information to the L2 file.

This information is also redundant with romanisim.simulate_reffiles which lists the same dark filename.

@schlafly mentioned that the use of meta.ref_file was a request from RTB. To avoid this confusion I suggest that we either:

• stop populating meta.ref_file in romanisim (and update any downstream code to use romaisim.simulate_reffiles
• stop using meta.ref_file information from L1 files (which is only a partial fix since it avoids the confusion for the L2 files but not for the L1 files).

[braingram]

@stscijgbot-rstdms

[stscijgbot-rstdms]

This issue is tracked on JIRA as RCAL-1334.