Welcome to nirHiss, a data reduction routine for the curviest JWST instrument! Here, you'll find all kinds of tricks to extracting high-precision NIRISS light curves (and subsequently transmission spectra)!

For installation, you can grab the latest release from PyPI by doing:

    pip install nirhiss

or you can the latest development version by doing:

    git clone https://github.com/afeinstein20/nirhiss
    cd nirhiss
    python setup.py install

ADF_extracted_stellar_spectra

• F277W filter from the median image, then gaussian filtered to smooth any noise
• Removing cosmic rays and interpolating over
• Using the 2D modeled background to get rid of 1/f noise
• Subtracting the median from each column

ADF_extracted_stellar_spectra_method2

• F277W filter from the median image, but only using the high outliers so it captures the 0th order effects and not any noise in the image
• DQ masked and interpolated pixels
• Using the SUBSTRIP256 model background provided on JDox
• Removing cosmic rays and interpolating over
• Removing bad integrations (5 in total)
• No additional background modeling

ADF_extracted_stellar_spectra_method3

• F277W filter from the median image, then gaussian filtered to smooth any noise
• DQ masked and interpolated pixels
• Using the SUBSTRIP256 model background provided on JDOX
• Removing cosmic rays and interpolating over

ADF_extracted_stellar_spectra_method4

• F277W filter from the median image, but only using the high outliers so it captures the 0th order effects and not any noise in the image
• DQ masked and interpolated pixels
• Using the SUBSTRIP256 model background provided on JDox
• Removing cosmic rays and interpolating over
• Minor additional background correction (only looking at pixels <1.8 sigma), interpolating, and smoothing with a Gaussian filter

WASP-39b data set

ADF_extracted_wasp39_full_v5

• F277W filter from the median image, sigma clipped to isolate the 0th order sources, then gaussian filtered to smooth any noise
  • Scaling to 2 different sources and taking the median
• DQ masked and interpolated pixels
• Scaling the SUBSTRIP256 model background provided on JDox
• 1/f noise correction via Néstor's routine (no mask)
• Removing cosmic rays and interpolating over those pixels
• Minor additional background correction, interpolating, taking a median of all the models, and smoothing with a Gaussian filter
• Decreased box mask size for Orders 1 and 2 to 24 pixels each (diameter)