calcos 2.21

Calibration software for COS (Cosmic Origins Spectrograph)

Release notes for 2.21:

• When x1dcorr in costools (x1d.py in calcos) is called to extract a 1-D spectrum, the flt and counts files are first recreated from the corrtag file. These files contained WCS keywords that were copied without change from the input corrtag table, and these table-specific keywords are not permitted in image headers. This was fixed by calling the function to populate the spectroscopic keywords.

• A TEAL interface has been added to the calcos source directory, which allows the user to see and edit the calcos parameters interactively (import calcos; from stsci.tools import teal; teal.teal(“calcos”)).

• The variables used for data quality flags have been renamed to agree with the interpretations of the data quality values for the latest bad pixel table, as described in the Data Handbook.

• A bug was fixed in the code for trapping a missing-row exception. This is relevant primarily when writing a “calcos sum” (csum) file; calcos must trap the exception, write the csum file, and continue processing until all input files have been calibrated at least to the point of writing the csum file.

• CALCOS was modified to use astropy.io.fits in place of the stand-alone PyFITS module.

• The check on whether the aperture name is in the list of recognized values has been extended to include “N/A”, which is valid for dark exposures.

• Calcos now checks the value of wavecorr before updating WAVECALS so that the WAVECALS keyword will only be updated if wavecorr is COMPLETE.

• A new column has been included in the wavecal parameters table, giving a zero-point offset for the search range. Calcos now reads this column (if it is present) in the WCPTAB and adds the offset to the limits of the search range.

• The EXPTIME keyword in NUV x1dsum files has been fixed. When averaging spectra to make an x1dsum table, the exposure time keyword was incremented for each stripe, in addition to each input file, resulting in a value that was three times too large. (The EXPTIME column was already correct.)

• Calcos checks that EXPEND - EXPSTART is not zero when computing the exposure time needed to scale the flat-field when averaging spectra.

• Code now gets DETECTOR from PRIMARY header when including the flat-field in the weight when averaging 1-D spectra.

• The first or last pixel of an x1dsum table sometimes had a data quality value of 0, even when that pixel was actually off the detector and should have been flagged as bad. This has been fixed.

• The exposure time keyword now gets updated before writing the csum file, so that if calcos fails due to a bad value of the APERTURE keyword, the exposure time keyword in the csum file will be correct.

• CALCOS now checks the value of WAVECORR before updating the WAVECALS primary header keyword and only updates if WAVECORR=COMPLETE.

• Code now uses boxcar function from stsci.convolve.