Fix MIRI LRS spec3 regtest setup

jwst/regtest/test_miri_lrs_slit_spec3.py

@@ -20,13 +20,31 @@ def run_pipeline(rtdata_module, request):
     Run the calwebb_spec3 pipeline on an ASN of nodded MIRI LRS
     fixed-slit exposures using different options for the WCS and output
     image shape for the resample step.
+
+    The first iteration ("default_wcs") creates an output WCS for the combined
+    product based on the built-in WCS's in the inputs.
+
+    The second iteration ("user_wcs") creates a user-supplied WCS input file
+    using the WCS from the default product, just to prove that you get the
+    identical result when using a user-supplied WCS.
+
+    The third iteration ("user_wcs+shape") uses the same user-supplied WCS and also
+    specifies the output 2D file shape, using a shape that is identical to the
+    default. Hence all of the first 3 iterations should produce identical results
+    and therefore are compared to a single set of truth files.
+
+    The fourth iteration ("user_wcs+shape1") uses the same user-specified WCS and
+    specifies an output 2D file shape that is 1 pixel larger than the default in
+    both axes. Hence the resulting s2d product needs a separate (larger) truth file.
+    Meanwhile, the x1d product from this iteration should still be identical to
+    the first 3, because the extra row and column of the 2D data file are ignored
+    during extraction.
     """
     rtdata = rtdata_module
 
     # Get the spec3 ASN and its members
     rtdata.get_asn("miri/lrs/jw01530-o005_20221202t204827_spec3_00001_asn.json")
     root_file = "jw01530-o005_t004_miri_p750l_"
-    rtdata.get_truth("truth/test_miri_lrs_slit_spec3/jw01530-o005_t004_miri_p750l_s2d.fits")
 
     args = [
         "calwebb_spec3",
@@ -36,9 +54,12 @@ def run_pipeline(rtdata_module, request):
     rtdata.custom_wcs_mode = request.param
 
     if request.param != "default_wcs":
-        dm = datamodels.open(rtdata.truth)
+        # Get the s2d product that was just created using "default_wcs"
+        default_s2d = root_file + "s2d.fits"
+        dm = datamodels.open(default_s2d)
+        # Create a user-supplied WCS file that is identical to the default WCS
         af = asdf.AsdfFile({"wcs": dm.meta.wcs})
-        wcs_file = rtdata.truth[:-8] + 'wcs.asdf'
+        wcs_file = default_s2d[:-8] + 'wcs.asdf'
         af.write_to(wcs_file)
         args.append(f"--steps.resample_spec.output_wcs={wcs_file}")