Replaced global pipeline task dictionary with a task registration scheme. Task registration and a brief description is now registered in the task classes themselves
Work on spw mapping for imaging multiple Ms(s) resumed
Ongoing ALMA Pipeline / Archive discussions via email
A document describing the current pipeline manifest needs to be produced. Longer term the old pipeline results schema should be reviewed
VLASS / VLA
VLASS requirements review took last week
Some pipeline requirements need better definition, e.g. how to automate spectral index fitting.
Some requirements need more data flow analysis, e.g. how does input data flow from a previous pipeline recipe to the next, etc
Excessive memory usage in hifv_importdata. Probably pipeline issue but the code seems to be swapping which is puzzling.
Slow memory leak in repeated hif_editimlist calls. Workaround is to generate standard lists internally in one call. Needs investigation.
Concern expressed on the issue of pipeline release patches for different pipeline at CCC meeting. Pipeline does not see this as a major issue as telescope specific code is maintained in different pipeline "packages". Patches are to be expected from eiter CASA or Pipeline or both.
Pipeline is concerned about the CASA release schedule and bottlenecks there given the current manpower situation.
CAS-10697: "uvcontsub does not work in parallel mode with 'combine=spw'" - ready for validation
Pipeline tests with 5.2.0-9, varying number of cores for selected "big" datasets ongoing, somewhat delayed as the cluster machines had to be rebooted
All of the Mac packaging is now done with XCode 9. There is no rush for the developers to upgrade.
Ongoing ALMA Pipeline / Archive interface discussions
Participated in VLASs single epoch imaging requirements review
Test new task registration capability
CAS-11058: plotting CalTables (done) and associated features in plotms: locate (done), selection (in progress), summary (not started)
Continued work on CalLibrary, mainly exclusivity logic that discerns absent-but-expected from merely absent calibration (per obs/fld/intent/spw) for purposes of jointly supporting per-caltable agnosticism (merely absent) and applymode='*strict' (expected). Almost done!
Some polarization science that had been perpetually languishing...
Federico M Pouzols
CAS-10584 "mstransform regridding in fftshift interpolation creates all zeros" fixed in cvel2/mstransform.
CAS-10340 "mstransform + tclean drops a channel in cube", CAS-10446 "Velocity not transformed to LSRK" - looking into them.
CAS-10794 "Slowdown of pipeline tasks in parallel mode".
CAS-10946 : Resolved : Redid restoring beam fixes from the 5.2 branch onto the 5.3 master. Verification test added. Merged to master (as requested by JM).
CAS-2961 : Ready to Validate : Enable Pseudo-I stokes imaging as a special case via tclean(stokes = "pseudoI"). Verification tests added. Work needed to complete other use cases added to CASR-275 (by JO).
Sent Klara S. some casa images and a script to make images of different dimensions/shapes to get her started with exploring the viewer. Sent Tak T. a description of one way to implement N-sigma thresholds. RH7 issues related to CASA install/build/use all seem to be resolved now (or will soon be once buiild/test turn on RH7 build packaging).