STScI Logo

COS
CALCOS v2.15.4 Release Notes

This version has been installed with STSDAS Version 3.14 Released on August 2011
This has not been installed in OPUS yet.

---------------------CALCOS v2.15.4--------------------- 

This describes changes from calcos version 2.13.6 to 2.15.4.

Relevant Trac tickets and PRs:
Trac   PR
----  -----
 465  63864  Calcos should copy G140L wavecal info from A to B.
 616  66158  Calcos should be modified to use Sphinx for documentation.
 628  66626  Calcos should use images for pulse height limits.
 635  66908  Calcos should write a 3-D csum file for any FUV TIME-TAG file.
 639  67114   Change calcos procedure for adding WCSAXES keyword.
 640  67115   Calcos can leave a gap at the top of an NUV DQ image.
 645  67204  Add a TIMELINE extension to calcos corrtag files.
 657  67439  Modify the calcos command-line arguments for the csum file.
 661  67473  Calcos should not always re-do wavecorr for corrtag input.
 670  67553   Calcos uses malloc/free in C interface in some functions.
      67675  COS csum header information will only be correct if used
             with the --only_csum option
 702  68209  A "walk" correction needs to be added to calcos.
 707  68355  For COS FUV, splittag and x1dcorr use wrong EXPTIME keyword.
 717  68529  Calcos should compute the FWHM when finding a spectrum.
 718  68538  Update deprecated Python syntax in calcos.
 719  68545  Calcos needs an option to use the default location for the
             target spectrum.
 720  68550  Calcos should accept data taken with different cenwaves.
 735  68712  Change calcos import statements for boxcar.
 736  68713  Background smoothing in calcos includes some bad regions.

A bug in background smoothing during spectral extraction has been fixed.
The smoothing was being affected by zero values of the background near the
endpoints, where all pixels were flagged as bad in the DQ extension.

A bug that prevented using a corrtag file as input to calcos has been
fixed.

When the input to calcos is a tagflash corrtag file, the wavelength
calibration will only be re-done if keyword wavecorr is set to perform.
Previously, this step would be repeated if wavecorr was either perform or
complete, but this would be inappropriate if the corrtag file had been
split by time, e.g. by running splittag.

For FUV input data with WAVECORR = COMPLETE (e.g. corrtag input), if
very verbose trailer info was specified (-v), a warning message was
printed regarding the lampflash tables.  This warning was not necessary
and could be misleading; it should no longer be printed for this case.

The source files for tasks splittag and x1dcorr were modified to use
segment-specific keywords for exposure time, for FUV data.  This affects
columns ERROR and GCOUNTS in the x1d table.

A new TIMELINE extension will be added to corrtag files, giving the
altitude of the Sun, the longitude and latitude of HST, and other
information at one-second intervals throughout the exposure.

For FUV data, the pixel coordinates of a detected photon differ from the
actual location at which the photon hit the detector (this is referred to
as "walk").  The offset is a function of the pulse height amplitude (PHA).
The changes to calcos allow corrections to the XCORR and YCORR pixel
coordinates to be computed as polynomial functions of XCORR and YCORR.
This is applied after the thermal and geometric corrections.

For G140L data, the wavecal shifts are normally copied from segment A to
segment B.  If there is no segment A data, however, calcos used to attempt
to find the wavecal shift.  There is no significant wavecal signal for
segment B, so the attempt always failed and sometimes gave a shift that
was quite far off.  Now if there is no segment A data, calcos will set
the wavecal shifts to zero for segment B.

When extracting the spectrum, the FWHM of the cross-dispersion (Y) profile
and the error estimate of the spectrum location in the Y direction are now
printed, in addition to the expected and found locations.

For FUV TIME-TAG data, the lower and upper thresholds for pulse height
may now be specified by images instead of a table, allowing different
limits at different points one the detector.  The new reference file has
keyword PHAFILE.  (However, no such reference file has been created yet.)
If PHAFILE is present in the primary header, and if its value is not N/A,
the specified file will be used instead of the PHATAB table to do the
PHACORR.  For ACCUM data the PHATAB must still be used.

Most of the doc strings have been modified to use Sphinx syntax.

If a csum file is to be created, and if raw pixel coordinates should
be used instead of corrected coordinates (relevant only for FUV),
some of the calibration switch keywords were set to COMPLETE in the
csum header, which is misleading because it would not be true for the
csum file.  These keywords will now be set to OMIT if the option to
use raw coordinates (--raw) was specified.