Bug in OCRREJECT CR-growth parameter

[sean-lockwood]

@stscirij -

Setting stistools.ocrreject.ocrreject() parameter crthresh to 0 does not force all surrounding pixels to be flagged within a radius defined by crradius, and instead does not grow the CR region at all.

From the docstring:

If crthresh = 0 then all adjacent pixels (see crradius) will be rejected.

Setting it to zero doesn't actually do this:

from stistools.ocrreject import ocrreject

filenames = ['oeki09cvq_flt.fits', 'oeki09cxq_flt.fits', 'oeki09czq_flt.fits',
             'oeki09d2q_flt.fits', 'oeki09d4q_flt.fits']

ocrreject(', '.join(filenames)),
          output='combined.fits',
          all=True,
          crthresh=0,
          crradius=3.,)

However, setting it to a slight negative value seems to do the trick.

ocrreject(', '.join(filenames)),
          output='combined.fits',
          all=True,
          crthresh=-0.001,
          crradius=3.,)

stistools calls cs2.e with the -thresh parameter set to the provided crthresh value.
This gets inserted into the par->psigma parameter in hstcal, and gets used here:
https://github.com/spacetelescope/hstcal/blob/main/pkg/stis/lib/cs2/crrej_loop.c#L92

[sean-lockwood]

Aha!

stistools doesn't set the flag because if 0 doesn't evaulate to True:

if crthresh:
    optional_args.append("-thresh")
    optional_args.append("%.10g" % crthresh)

We should change this to if crthresh is not None.