WFCAM Software Changes
2007 March 12
perllib/cir_dither, wfcam/process
Modified these two files to change the way bad input is reported in the output products. The IAMFUD keyword that Marco inserts during the meffing stage has been used for some time to flag input files that should not for example be used in creating sky files. These changes have extended the use if IAMFUD in the following way:- For single observation frames, Marco's keywords are left unchanged
- For superframes IAMFUD is inserted into the extension header of any FITS image where at least one of the input frames is missing from the microstep sequence. Although a superframe with a missing member is useless by itself, it can still be useful as input to a stack. For that reason a keyword OK2DITH is inserted into an extension header and set to T if there is at least 1 good input contribution. The stacking routine, then ignores IAMFUD and looks for OK2DITH when stacking superframes. If all the input frames for a superframe are FUD, then OK2DITH is set to F and the stacking algorithm will ignore this superframe at all. If at least one extension in the superframe has the IAMFUD keyword set, then it is also set in the primary header.
- The stacking algorithm looks for OK2DITH in the header of superframes to spot bad input. If there is no interleaving then the IAMFUD in the simple frames flags bad data. IAMFUD is set in the extension header of the stacked product if all the input frames are bad. If at least one extension if FUD then the primary is also flagged with IAMFUD.
2007 March 19
perllib/cir_dither
Bug introduced in 12 March update meant that initial WCS offsets weren't being calculated properly. This caused many dither sets to have 'failed matches' when doing the object offsets. In some cases this caused all shifts to be set to zero and consequently all objects reject out in the combine phase.
2007 March 22
src/cir_imcombine.c
Modified cir_interleaf so that it forces the fractional part of the offsets to be the exact amount that they should be. This is because offsets that are not quite right because of the funny autoguider setup they have at UKIRT can sometimes cause the routine to skip a pixel during the interleaving stage.
2007 May 10
src/cir_imcombine_lite.c
Modified the scaling/biassing algorithm so that the factors are calculated relative to the ensemble median rather than the ensemble mean. This should help to trap dodgy frames in the situation where all the images are expected to have roughly the same background value
2007 May 14
wfcam/darkcombine
Modified to recognise IAMFUD flags
2007 May 17
wfcam/process
Fixed small bug routine splitgroups that causes extra groups to appear in the situation where a repeat observation is done on the last frame in a group
2007 August 24
wfcam/process
New paw sky routine added to do iterative masking for sky estimations. These are called 'level 0' skies.
2007 September 28
wfcam/process
Fixed jitter offset zeropoint bug in paw sky routine.
2007 October 18
wfcam/preview
Fixed a bug which caused a whole pawprint to be removed from the sky analysis, just because it had a single bad frame. This caused the pawprint to be ignored by the tilesky routine and then it went on to be assigned the wrong sky frame.
2007 October 24
wfcam/preview, wfcam/process
Modified to write the version of preview into the instructions.dat file. This is subsequently written to the primary header of any of the fits files that are processed and propogated to the products.
2007 November 29
wfcam/preview
Added a line to pawsky_mask routine to make sure that the correct sky is chosen for each input file
2007 December 12
wfcam/preview
Fixed bug in atilesky so that tiles with no good frames are dealt with properly
2007 December 18
src/catalogue/imcore_background.c
Modified background level iteration parameters
2008 January 21
Astro::FITS::CFITSIO
Upgraded to version 1.05cfitsio
Upgraded to version 3.06
2008 February 11
src/cir_qblkmed.c
Added background tracking capability
2008 February 12
src/cir_imcombine.c
Fixed bug in imstack where replacement value was being calculated incorrectly
2008 March 07
cir_startup.tmpl, Cirdr/Fileinfo/Fileinfo.pm
Now has CIR_NOXML environment variablesrc/cir_getstds.c
Changed target column names for VizieR lookupwfcam/process
Rewrote skycomb_mask routine to improve speed and results
2008 April 09
src/cir_compress_fits.c
Modified cir_compress_mef_f2i to change the way the scaling is done. This is because the underlying cfitsio routine has been fixed to work correctly.
2008 April 10
src/cir_compress_fits.c
Modified to fix the memory allocation in cir_uncompress. This takes into account that long integers on 64 bit platforms are 8 bytes.
2008 April 15
various imocre routinesModified several routines to change the way that data flagging is done.
2008 April 21
wfcam/preview, wfcam/process
Modified the structure and content of instructions.dat greatly. Old versions will now be incompatible. Also modified the sky combination section to include a first cut at sky flipping and to allow for the rejection of single FITS images rather than an entire FITS file.
2008 April 30
wfcam/preview, wfcam/process
Consolidated work on sky flipping algorithm to a final version.
2008 September 15
src/cir_imcombine.cModified median algorithm so that so long as there are sufficient data points, then the median is formed from a weighted average of the 3 points around the median position
src/cir_imdither.c
Background correction is now worked out in the overlapping regions
wfcam/preview, wfcam/process
Tile and mask sky subtraction algorithms now have min/max frame specifiers.
2008 October 08
wfcam/previewFixed a number of small bugs in the paw masking sky estimation algorithm. Also lowered the galactic latitude cutoff used for determining which images are too close to the galactic plane for use in sky modelling. This is now 2degrees.
2008 October 21
src/catalogue/create_table_2.cFixed indexing bug where wrong fluxes were being passed to special (petrosian, etc) radius routines.
2008 October 24
wfcam/previewCriterion for membership to a tile is opened up to 15 arcminutes difference between pawprints
2009 January 22
various imcoreFixed another pixel flagging bug that crept in as a result of 15 April 2008 mods.
2009 January 29
src/catalogue/imcore_radii.c, src/catalogue/create_table.c, wfcam/preview, wfcam/processModified Hall, Petrosian and Kron radius routines to be consistent with Mike's latestversion. Modified imcore_flux which calculates H,P,K fluxes so that if a blend is given then the flux is done as an interplation of the aperture fluxes.
Every catalogue extension header now gets a HISTORY card with the imcore version.
Pipeline now removes the parquet pattern
2009 February 09
include/cirdr.h and various othersThe global macro CIRDR_VERSION is defined in cirdr.h and exported. This is now written to all extension headers as CIR_VERS.
src/catalogue/imcore_version.hA version string is defined in this header file and is written to the extensions of every catalogue as a HISTORY card.
2009 February 23
src/cir_imdither.cStatistics for scaling is now calculated only over the overlapping regions
2009 March 09
wfcam/preview, wfcam/processModified to allow for offset sky observations
Fixed bug in overlap scaling routine
2009 March 16
wfcam/processNow writes sky estimation algorithm into sky file headers
perllib/cir_ditherRewritten so that dither offsets are now calculated using all 4 detectors at once
2009 April 27
BUILDRemoved external packages
2009 JULY 20
wfcam/previewNow traps UDS data so that the sky background is always done with the pawsky_mask algorithm