columnjump is an additional step that can be applied as part of the JWST DETECTOR1 pipeline for data from the NIRISS instrument. The step should be called after dark current subtraction and before jump detection. The columnjump step removes random jumps in the levels of some columns (~50 columns per Ng=100 NISRAPID ramp) that cause increased noise along those columns. Note the term columns here refers to the original detector coordinates and these are actually rows in the DMS orientation, i.e. they are orthogonal and distinct from the well-known stripes of 1/f noise.

Installation
columnjump is available at pypi:

pip install columnjump

Usage
A typical calling sequence is:

from columnjump import  ColumnJumpStep  
columnjump = ColumnJumpStep()  
#Manually set any desired non-default parameter values  
columnjump.nsigma1jump  = 5.00        \# sigma rejection threshold for one jump in the ramp  
columnjump.nsigma2jumps = 3.00        \# sigma rejection threshold for two jumps in the ramp  
columnjump.outputdiagnostics = True    \#  output a table of the columns corrected?  
columnjump.output_dir = out_dir  
columnjump.output_file = out_file  
#Run the step  
result = columnjump(inputfile)

image1overf.py performs a correction for 1/f readout noise on NIRISS or NIRCam imaging data. The operation is performed on the calibrated level 2 image. It includes the effect of a variable background (by temporarily removing it before determining the 1/f correction and adding it back afterwards) and masking pixels containing sources. Only two subarrays are supported: FULL and SUB256. Use with care and inspect the results for any unintended consequences.

Usage
A typical calling sequence on a single level2 image named cal2file is:

from image1overf import sub1fimaging
cal21overffile = cal2file.replace('_cal.fits','_cal_1overf.fits')
print ('Running 1/f correction on {} to produce {}'.format(cal2file,cal21overffile))
with fits.open(cal2file) as cal2hdulist:
    if cal2hdulist['PRIMARY'].header['SUBARRAY']=='FULL' or cal2hdulist['PRIMARY'].header['SUBARRAY']=='SUB256':
        sigma_bgmask = 3.0
        sigma_1fmask = 2.0
        splitamps = False   #Set to True only in a sparse field so each amplifier will be fit separately. 
		usesegmask = True   #Recommend set to True in most cases
        correcteddata = sub1fimaging(cal2hdulist,sigma_bgmask,sigma_1fmask,splitamps,usesegmask)
        if cal2hdulist['PRIMARY'].header['SUBARRAY']=='FULL':
            cal2hdulist['SCI'].data[4:2044,4:2044] = correcteddata  
        elif cal2hdulist['PRIMARY'].header['SUBARRAY']=='SUB256':
            cal2hdulist['SCI'].data[:252,:252] = correcteddata
        cal2hdulist.writeto(cal21overffile, overwrite=True)

dosnowballflags.py flags pixels in snowballs - expand saturated ring and diffuse halo jump ring. The GROUPDQ array will be flagged with the SATURATED and JUMP_DET flags. Saturating snowballs early in short ramps can have unflagged central pixels that jump in the previous group. This is called after the regular jump step. The output file is overwritten. You need a working installation of WebbPSF. Works for NIRISS, NIRCam and NIRSpec. Requires checkifstar.py for differentiating between saturated stars and snowballs if imagingmode is True.

Usage
A typical calling sequence is:

from dosnowballflags import snowballflags
jumpdirfile = './jw01345001001_02201_00001_nrca1_jump.fits'
filtername = 'F115W'
npixfind = 40
satpixradius = 3
halofactorradius = 2.0
imagingmode = True
snowballflags(jumpdirfile,filtername,npixfind,satpixradius,halofactorradius,imagingmode)

checkifstar.py builds a WebbPSF model and then compares an image cutout with the model to determine if that cutout corresponds to a star or not. This use the diffraction spikes and PSF asymmetry. Currently works for NIRISS and NIRCam. Will fail on extremely saturated stars, but good on moderately saturated ones.

dopersistflags.py flags pixels that reached a very high number of counts in the previous integration. The GROUPDQ array will be flagged with the JUMP_DET flag (this is called after the jump step and snowball flagging so won't interfere with that). Only groups within timeconstant after end of previous integration are flagged. The input file is overwritten.

Usage
A typical calling sequence is:

from dopersistflags import persistflags
prevrawdirfile = 'jw01222002001_03104_00002_nrs1_uncal.fits' #note can be empty string for first exposure in visit
rawdirfile =    'jw01222002001_03104_00003_nrs1_uncal.fits'
jumpdirfile =  'jw01222002001_03104_00003_nrs1_jump.fits'
countlimit=50000
timeconstant=1000.0
persistflags(prevrawdirfile,rawdirfile,jumpdirfile,countlimit,timeconstant)

makebpmreadnoise.py generates three bad pixel mask files and readnoise reference file for the NIRISS detector. The three masks have different dark noise thresholds appropriate to various types of calibration or science data. The routine is robust to high rates of cosmic rays and separates cosmic ray hits from noisy pixels. makebpmreadnoise.py calls makedarknoisefilesgdq.py. Inputs come from the config file, an example is bpmreadnoise_nis006_20220620.cfg.

makedarknoisefilesgdq.py makes dark current and noise images from darks, optionally using GDQ flags to mark locations of cosmic ray jumps.

getipc.py determines 5x5 and 3x3 IPC kernel images based on spread of charge from hot pixels.

makenirissimagingflats.py generates NIRISS imaging flat fields for all filters from one or more integrations per filter that have passed through level 1 pipeline processing.

makenirissgrismflats.py takes in direct (for all filters) and dispersed (currently only F115W and F150W are required) NIRISS WFSS flat field images to identify features of low transmission on the NIRISS pick-off mirror (POM), including the coronagraphic spots. The outputs per filter are

1. An oversized image of POM transmission including the measured POM outline and features of low transmission.
2. Grism flat field reference files. These grism flats are equal to the imaging flats over most of the detector, but with all the POM features corrected leaving only the detector response.

For both types, files are generated for each of the GR150C and GR150R grisms. The contents of the files are identical, but separate copies are required for automatic CRDS selection for use with data from each grism.