Monday Morning Meeting June 2nd

• DIAL-IN NUMBERS & PASSCODES:
• USA Toll Free Number: 866-901-8266
• USA Toll Number: +1-203-566-3863
• PARTICIPANT PASSCODE: 725760

Attendance

• Socorro:
• CV:
• Garching:
• SCO:

News / Meetings / Visitors

• Users Committee / ANASAC Meeting
• Jira System e-mail

Developer Reports

Monday Meeting

• Sanjay Bhatnagar
  • Debugging normalization in mosaic imaging with the new framework
• Sandra Castro
  • CAS-6543: Working on a fix for the case when an input is an MMS in mstransform. Currently, it treats the MMS as a monolithic MS, but in some cases it should run in parallel.
  • CAS-5833: Support to SYSPOWER flagging in flagdata. Asked Steve about selections in the SYSPOWER table. No answer from him. This implementation is stalled with lack of requirements.
  • CAS-5852: mstransform needs to have the keepflags=True/False capability of split. Tested that this feature is included in the casapy-test-43.0.29326-001-1-64b package. All mstransform tests pass, including the new tests that check for this new feature.
  • CAS-6472: When splitting part of the channels out of ms, all visibility amplitude with FLAG=True become 0+0j. Tested that this problem doesn't exist in split2 or mstransform.
  • Gave a talk on MPI basics in the weekly ALMA coffee.
  • 2 days of vacation.
• Lindsey Davis
• Kumar Golap
• Justo González Villalba
  • Released CASA 4.3 MPI parallel processing framework documentation (Installation and advance user guide)
  • Changed MPI finalization procedure to mpi4casa to avoid blocking situations in case of server timeout
  • Successfully tested SWIG's thread option to disable Python's GIL in the components call
  • Investigated a mpirun mode to change the MPIClient convention to rank #0
  • Tested TAQL strategies to apply flag commands in 1TB EVLA sky survey MS
  • Discussed with Dirk implementation details of the new WEIGHT/SIGMA handling in the cvel algorithm
• Jim Jacobs
• Jeff Kern
  • PEP Process
  • ANASAC Meeting
  • Users Committee Meeting
  • Jira Cleanup
• Susan Loveland: Fixed viewer bug, merged plotms to 4.2.2, worked on CARTA UI prototype.
• David Mehringer, moved ImageBeamSet description method from ImageSummary to ImageBeamSet
  Did some immath research for a JIRA issue
  implemented msmd.fieldsforsource()
  Fixed a label in listobs
  implemented support for flux density calculation in imstat for images with units of K
• George Moellenbrock
  • Fixed time_then_chan_avg regression to follow 4.2.2-related split changes
  • Studied VI2/VB2/mstransform internals re new weights schemes
  • Diagnosed problem with new Tsys weight scale calculation that was detected in 4.2.2 pipeline testing
  • Discussions with Jeff and Crystal re ALMA long-baseline strategies
• Dirk Petrey
• Scott Rankin
• Urvashi Rao
  • Added ms-clean and clark-clean to new imaging framework.
  • Many discussions with various people about algorithms, code, bugs, and some user support related to the workshop.
• Darrell Schiebel -- much of this week was spent with third-party packages for OSX 10.8, with gcc 4.8, SIGABRT on unit tests, now rebuilding with llvm-gcc42
• Rob Selina
  • Worked on a plan for the Build & Test Review with Mark.
  • Working on a long-term schedule for CASA development and infrastructure work (~3 years) with Jeff.
  • Researching JIRA dashboards and workflow.
• Takahiro Tsutsumi
  • Fixed the fluxscale dictionary issue in the 4.2.2 branch (CAS-5699)
  • Commented on CAS-6550 why the Ephemeris tables for the Butler-JPL-Horizons 2010 is not updated
  • More Setjy code modifications for moving away from per-spw process

Friday NAOJ Meeting

• Kanako Sugimoto
  • merged SD developments for CASA 4.2.2 from trunk to release-4_2_2
  • discussing about having JAO/NAOJ monthly skypecon to keep each other up-to-date for EOC (formar CSV) activities and CASA-SD developments.
  • created a couple of JIRA tickets of CASA 4.3 targets and also set fix version and due dates to CASA 4.3 targets.
  • discussions of proper flag handling in single dish tasks (CAS-1756~1758)
  • transfered functional tests of sdstatold to those of sdstat.
• Wataru Kawasaki
  • found a bug in asapplotter.
  • fixed a bug in sd.scantable.sub_baseline()
  • updated unit test codes (done: sdcal2, sdplot, almost: sdmath, remaining: sdgrid)
  • works for very old tickets CAS-175[678]: listed up how flags should be treated in each SD tasks and discussed with other SD developers.
• Takeshi Nakazato
  • added a few new tests to sdfit unittest
  • filltsys discussion and bug fix (CAS-6355, CAS-6558)
  • fixed sdimprocess unit test failure
  • looked at Mac OS X build failure report
  • reported strange clean behavior that the result changes in every run and never reprodicible
• Shinnosuke Kawakami
  • CAS-6461(remove async from task) it was able to remove async parameter by modifing xsl file
  • done CASA4.2.1 regression/unit test for introducing CASA 4.2.1 release tarball to servers at the data analysis center in NAOJ.

Build, Release, Testing

Architecture

• Technical Forums:

Development

• Prospects of modifying Logger filter behaviour to enable compressed views in the GUI, but longer logs in the txt file. Currently, the casalog.filter(...) and the GUI priority text box have different syntax and follow different selection rules, the combination of which does not allow the above ( from UR ).
• Prospect of developer builds running on EL6 machines in Socorro ? Cannot release some algos without testing on VLA data large enough to need the more than single desktop. Can build EL6 CASA on a few (one?) machine(s), but cannot run it on any EL6 compute node in the building unless code is a packaged release. What should we tell early testers/users currently in the building with data expecting to test new algorithms over the summer ? (From UR, but also brought up before by KG and SB).

Pipeline

AOB