Update release notes

1 files changed, 172 insertions, 0 deletions

diff --git a/source/release_notes.rst b/source/release_notes.rst
index 769a53f..d67613e 100644
--- a/source/release_notes.rst
+++ b/source/release_notes.rst
@@ -33,6 +33,178 @@ The following changes were made since the last release in April 2014:
 
 -  Added Python 3 support (untested).
 
+CALCOS
+------
+
+v3.1.3
+~~~~~~
+
+*None*
+
+Most of the additions to the code were to make calcos run under Python
+3.
+
+v3.1
+~~~~
+
+*None*
+
+Calcos 3.1 includes the following bug fixes and features:
+
+Allow flux extraction from 0 to 100%
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The latest version of calcos (3.0) throws an IndexError when the range
+of the 2zxtab is 0 to 100%. Handle this case separately, always
+integrating over the whole height of the extraction box to duplicate the
+behavior of the boxcar
+
+option.
+
+Ignore DQ flag DQ\_GAIN\_SAG\_HOLE in background regions in profile alignment step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+CALCOS uses DQ flags in the profile alignment step, omitting columns
+that have a pixel DQ value that contains any value in SDQFLAGS (read
+from science data header). The flagging of large areas with gain sag
+holes after the move to LP3 has made it very difficult to find
+background regions that won't cause whole columns be omitted from the
+alignment step, possibly resulting in not enough good columns to do the
+alignment correction.
+
+To alleviate this, pixels in the background region should be able to
+test against a different DQ value (e.g. one without the
+DQ\_GAIN\_SAG\_HOLE value) from pixels in the target region.
+
+Shift DQ data array the same as data array in TRCECORR and ALGNCORR steps of calcos
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The data arrays are shifted in the Y direction in the TRCECORR and
+ALGNCORR steps of calcos, but the DQ arrays are not currently shifted.
+The DQ arrays should be shifted in the same way that the data arrays are
+in the ALGNCORR and TRCECORR steps.
+
+Fix up x1d output for boxcar option
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+There are a number of shortcomings in the x1d output of calcos if the
+XTRCTALG keyword is set to 'BOXCAR'.
+
+1. The BACKGROUND\_PER\_ROW column is calculated incorrectly (total
+   background is divided by the wrong number of rows) and should be
+   renamed to "BACKGROUND PER PIXEL"
+
+2. The NUM\_EXTRACT\_ROWS column is calculated incorrectly (off by 1).
+
+3. The Y\_LOWER\_OUTER, Y\_LOWER\_INNER, Y\_UPPER\_INNER and
+   Y\_UPPER\_OUTER columns are calculated incorrectly - they are
+   constants whereas they should follow the linear relation using the
+   SLOPE and INTERCEPT from the reference file.
+
+Restore display format for x1d columns
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+In the 3.0 update to CALCOS, the display formats for the columns in the
+.x1d output table were removed. They should be restored.
+
+CalCOS should print warnings for certain conditions of HOTSPOT calibration
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The new code to support hot spot flagging with the associated reference
+file will need to output warnings to the user in a couple select cases.
+
+1. If a DQ bit in SDQOUTER is not also in SDQFLAGS, a warning should be
+   printed.
+
+2. If the PHAMIN/PHAMAX header keywords of the applied SPOTTAB are more
+   inclusive that what is applied to the data itself, a warning should
+   be printed.
+
+Change DQ array handling in x1d with hotspot treatment
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+With the CalCOS change to handle the hot spots and the SPOTTAB reference
+file, the DQ array in the output x1d files will change slightly.
+
+1. The data quality value for each pixel should be given by: (DQ\_INNER)
+   \| (DQ\_OUTER & SDQOUTER)
+
+An additional check should be done to verify that the DQ\_WGT in the
+x1dsum is still calculated correctly after this modification.
+
+Events in COS data should now be flagged using the SPOTTAB
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+To handle the hot spots in COS data, a new reference file is being
+delivered (SPOTTAB) that will specify physical regions on the detector
+that should be flagged if they occur during any given dataset. CalCOS
+needs to flag each region given in a SPOTTAB if the time of the event
+(given by START, STOP in each row) overlaps with any good-time-interval
+(GTI) in the observation.
+
+Modules changed:
+
+Setup.cfg:
+''''''''''
+
+Version changed.
+
+lib/calcos/timetag.py
+'''''''''''''''''''''
+
+Added code to shift the DQ arrays by the amount of the trace correction
+and alignment correction before doppler blurring. Added code to pass the
+good time intervals to modules that need it to check for active
+hotspots. Added
+
+code to make sure that events that get shifted into the active area
+after their doppler correction is applied don't cause the minimum and
+maximum X shifts to be drastically wrong.
+
+lib/calcos/cosutil.py
+'''''''''''''''''''''
+
+Added code to check SPOTTAB keywords, and to add active hotspots to the
+DQ flags for events and the DQ arrays. Added code to shift the DQ values
+of events by the amount in the trace and alignment corrections.
+
+lib/calcos/calcos.py
+''''''''''''''''''''
+
+Added code to enable use of SPOTTAB reference file
+
+lib/calcos/concurrent.py
+''''''''''''''''''''''''
+
+Updated interface to cosutil.updateDQArray
+
+lib/calcos/extract.py
+'''''''''''''''''''''
+
+Renamed BACKGROUND\_PER\_ROW table column to BACKGROUND\_PER\_PIXEL.
+Restored display formet to table columns. Treat DQ flags in the outer
+regions if affected by a hotspot by setting DQ\_WGT to 0. DQ\_ALL column
+is now the OR of the DQ flags of all pixels in the extraction region.
+Fixed an indexing bug in how the extraction regions were specified.
+
+lib/calcos/getinfo.py
+'''''''''''''''''''''
+
+Added SPOTTAB capability.
+
+lib/calcos/trace.py
+'''''''''''''''''''
+
+Added code to make sure that gain sag holes don't affect the background
+DQ values.
+
+lib/calcos/x1d.py
+'''''''''''''''''
+
+Added code to pass brftab reference file to timetag.getWavecalOffsets so
+that the Active Area can be determined at the time the offsets are
+calculated.
+
 DRIZZLEPAC
 ----------