What's in the WFC3 pipeline right now:
- OPUS build 2012.3b
- calwf3 v2.7
- AstroDrizzle 1.01
These are the major changes for calwf3 from v 2.4.1 to v2.7
The calwf3 IR ramp fitting routine (cridcalc step) adds up the total integration time of all unrejected samples for each pixel and stores that value in the output FLT file TIME extension. In some unusual circumstances involving the rejection of either the 1st or 2nd sample as bad or containing a CR hit, the total integration time was not coming out correctly, due to either a single valid sample in the 1st read or the valid part of the ramp starting in the 2nd read. The error was such that the reported total time is an overestimate, because the time associated with the rejected samples is not being removed properly. The fitsamps routine in cridcalc was modified to correctly initialize the total integration time counter with the sample time of the first unrejected sample.
When calwf3 was executed for an association that contained repeat-obs (RPT) members it ignored the setting of the RPTCORR calibration switch in the image headers and always applies RPTCORR processing, which consists of calling the wf3rej task to combine the images. This made it impossible to control the application of RPTCORR processing via the OPUS keyword rules that set the RPTCORR switch. The high-level logic of calwf3 association processing was modified so that it only applies the wf3rej step when RPTCORR=PERFORM in the image headers. In association with this, the calwf3 routines that build the association-level trailer file have been modified so that when a sub-product has not been created (i.e. wf3rej is not run) they use the trailer files of the individual association members, instead of trying to use the trailer file(s) of the crj sub-product.
The wf3rej task was upgraded to read and save the BUNIT keyword value from the [sci,1] extension of each input image and to then check those values each time input sci data arrays are loaded in the various subroutines. If the input data are in units of rates (either DN/sec or electrons/sec), the data are immediately scaled up by the exposure time of the particular image to turn them back into units of counts, which is the expected working units within wf3rej.
Negative cosmic rays are now accounted for in IR WFC3 images during cosmic ray rejection.
The zero-read pixel value is now used for ramp-fitting when the pixel is already saturated in the zeroth read of the ramp. The behavior has been modified such that pixel which are saturated in the zeroth read or in the first read are handled the same way. In both cases the output value is set to the value in the zeroth read. In the case where the zeroth read value is already saturated, a dq=256 flag will be added to the DQ array of the flt file. In the case where saturation does not occur until the first read, no saturation flag will be added to the flt DQ array. This eliminates blank pixels in the output FLT image.
The IR cridcalc handling of pixels detected to contain spikes was modified so that pixels flagged as bad from the static DQ mask still go through the process of CR detection, in order to produce cleaner output FLT images.
New option added "-r" to calwf3 to print the current verison and exit.
aXe - stsdas.analysis.slitless
The test of the width parameter in the OAF file was updated to specify that it's the half-width and not the width. Similarly, the AWIDTH1A BWIFTH1A and AORIENT1A are copied of the Sextractor parameters A_IMAGE, B_IMAGE and THETA_IMAGE. They therefore refer to the semi major axis, semi minor axis and orientation to the semi major axis wrt to the horizontal.
Modified 08/10/2012 MJD