[Enhancement]: Better FITS header keywords for positions of mask and star

[mperrin]

Instrument or Category

NIRCam Stage 1/2 Pipeline, MIRI Stage 1/2 Pipeline, Image Tools, Analysis Tools

What is your suggestion?

As discussed in slack by @kdlawson @JarronL @juliengirard @kammerje and @mperrin:

• SpaceKLIP currently treats the CRPIX header value as synonymous with "where is the coronagraph mask", and (sometimes, perhaps?) where the star was tried to be positioned during TA. This oversimplifies the operational situation. @JarronL: "Need to disambiguate star and mask centers and keep CRPIX along with distortion header keywords intact"
• @mperrin: "There should be separate explicit header keywords in spaceklip output files for e.g. MASK_CEN, STAR_CEN in addition to CRPIX. Even if the values are identical such that MASK_CEN == CRPIX, this might help reduce confusion in the long run."
• Even better if we added additional keywords for error propagation in how precisely we know those values. @juliengirard suggests "populate MASK_CEN_ERR STAR_CEN_ERR keywords for each of them with comments on how they are determined"

Error traceback output

No response

What operating system are you using?

No response

What version of Python are you running?

No response

What Python packages do you have installed?

No response

[mperrin]

Sub tasks:

• [ ] Add new keyword for mask center, and its uncertainty
• [ ] Add new keyword for star location, and its uncertainty
• [ ] Make sure we obtain and are using the most current values for those. From SIAF for the mask center?
• [ ] Update the image alignment and shift functions to propagate changes to those keywords
• [ ] Update the level 3 image combination functions to propagate changes to those, for derotated and combined images.

[kdlawson]

Regarding the last subtask: we should also keep in mind that a final derotated & combined image will need one "MASK_CEN" value (and uncertainty) per pointing/roll. For Winnie reducs, we include the ImageHDU header for every image used (each containing the equivalent of MASK_CEN), but one could also just add numerical subscripts to combine everything to a single HDU FITS file more similar to current SpaceKLIP outputs.