For level 3 coron previews, show TA image and TA ROI

[stscijgbot-jwql]

Issue JWQL-283 was created on JIRA by Bryan Hilbert:

From Jonathan Aguilar :

A significant number of issues can be traced back to problems with target acquisition. For those, we would love to see the TA images with the TA ROI and ROI reference position overplotted, and if possible also the centroid reported by the telescope telemetry.

This is similar to what we started working on with the TA monitor long ago. It might be more useful to have the TA image and info sitting next to the science image anyway (as part of the "preview image". 

[stscijgbot-jwql]

Comment by Jonathan Aguilar on JIRA:

This clarification is pursuant to a conversation wtih Bryan Hilbert on Slack: MIRI coronagraphic data are typically taken in subarray readout mode, which makes plotting the ROI on the data annoying since you can't just use the "det" coordinate system of the ROI itself. Instead, you must translate the "det" coordinates of the ROI corners into the "sci" coordinate system of the aperture named in the SUBARRAY keyword of the PRI fits header. The following code snippet is one such way to do that:

import numpy as np
from astropy.io import fits
import matplotlib as mpl
from matplotlib import pyplot as plt
import pysiaf

fits_file = "MAST_2025-12-04T16_08_33.916Z/MAST_2025-12-04T16_08_33.916Z/JWST/jw09056014001_02101_00002_mirimage_cal.fits"
img = fits.getdata(fits_file, 1)

# use the header information to get the right pySIAF aperture objects
miri_siaf = pysiaf.Siaf("MIRI")
ta_apername = fits.getval(fits_file, 'APERNAME', 0)
readout_name = fits.getval(fits_file, 'SUBARRAY', 0)
roi_aper = miri_siaf[ta_apername]
readout_aper = miri_siaf["MIRIM_" + readout_name]

# convert the corners and reference position to SCI coordinates and subtract 1 for the pixel convention
ref_pt = readout_aper.det_to_sci(*roi_aper.reference_point("det"))
corners = readout_aper.det_to_sci(*roi_aper.corners("det"))
ref_pt = np.array(ref_pt) - 1
corners = np.array(corners) - 1

# for the ROI, plot it as a matplotlib patch
roi_box = mpl.patches.Rectangle((corners[0][0], corners[1][0]), roi_aper.XSciSize, roi_aper.YSciSize, lw=1, ec='w', facecolor='none', alpha=0.75)fig, ax = plt.subplots(1, 1)

# this is usually a good guess for the scaling
vmin, vmax = np.nanquantile(img, [0.01, 0.99])

ax.imshow(img, origin='lower', vmin=vmin, vmax=vmax)
ax.add_patch(roi_box)
ax.scatter(*ref_pt, marker='+', c='w', alpha=0.75)```