Monday Morning Meeting: February 6, 2017

• DIAL-IN NUMBERS & PASSCODES:
• IP: 192.33.117.12##8110
• Phone: (434) 817-6524

Attendance

• Socorro: Jeff, Lindsey, Martin, George, Pam, Juergen, Kumar
• CV: Andy, Akeem, Anand, Ville, Darrell, Jen, Dave, Bob
• Garching: Ferndando, Enrique, Sandra
• SCO:

News / Meetings / Visitors

• CASA Liason talk Feb 15 at noon in Socorro.

Build, Release

• CASA 5.0 final release containing ASAP.
  • Will add deprecation notice to all ASAP tools for 5.0. After 5.0 release all ASAP tools will be removed.
• ic2233 regression failing on all platforms. EL7 had 5 additional regression test failures: https://open-bamboo.nrao.edu/browse/CASA-MT-TE-156/test . Failures started with this build: https://open-bamboo.nrao.edu/browse/CASA-PM-230/commit
  • Tclean is now using vb2. There is an environment variable that can be set to use previous algorithms. Kumar will add information to ticket on how to use non-vb2.

Verification Testing

• Continued work on getting new build and test hardware - seems new charge codes have management in a state of confusion
• Plone test framework documentation (CAS-9381) set to "Ready to Verify" I believe for review by the style folks
• Got input from Anand on suggested devs to create task level automated tests
  • Input from Jeff still needed

Validation Testing

• Plone updates

Architecture

• Warnings Summary

HPC

• mpi4casa
  • (CAS-9628) Still verifying a report from Akeem that test_mpi4casa fails when max user processes is set too low (at 1024), depending on the number of cores used for the MPI servers. I have not been able to reproduce this yet. What I see is that one test fails if I use more cores than the machine has. This is a test that launches 2 plotms simultaneously. IN PROGRESS.
  • Darrell asked us about the use of the *_pg.py files in the MPI framework. They are used by MPICommandServer because they are lighter versions of the *_cli.py files. These files do not startup a few things in the paralell servers such as writing to the .last files and they do not rely on an ipython environment. If the wish is to remove this approach, please let us know and we will plan a refactoring of this part in the framework.
• ALMA pipeline testing in parallel
  • Not much progress in the last week due to many build issues with prerelase tarballs.
• EVLA pipeline testing in parallel
  • Brian fixed the issue with unknown intents in applycal. Both my serial and parallel tests run successfully. I have sent the logs to Juergen for verification. This was a band K dataset. Will try further with a VLASS dataset.

Development

• 5.1 mstool iterator routines
• Calibration & VI/VB2:
  • GM: Got combined modes working. Cleaning up details.
  • GM: out tomorrow

Pipeline

• ongoing development of hif_editimlist task for VLASS survey
• CASA 4.7.1 / pipeline C4R2.1 bug fixes
• CASA 5.0 integration issues
  • buildmytasks, startup sequence, 3rd party library, logging, version handling, packaging changes
• Jeff waiting until VLA and ALMA test results are returned before rolling final tarball.

AOB

Developer Reports

Monday Meeting

• Sanjay Bhatnagar
• Sandra Castro
  • HPC tasks as above.
• Lindsey Davis
  • hif_editimlist task review, cleanup
  • Ongoing exportdata task cleanup
  • CASA 5.0. integration testing
• Pam Ford
  • CAS-9482 plotting two vis axes/datacolumns ready for commit and verification
• Enrique Garcia
• Bob Garwood
• Kumar Golap
• Justo Gonzalez
• Jim Jacobs
• Jeff Kern
  • Response to CUC. Will circulate before sending.
  • Planning meetings for VLASS
• Susan Loveland
• David Mehringer
  • Working on RMS images with Amy.
• George Moellenbrock
  • Finally sorted out combine modes for VI2
  • Plone docs for calibration
• Dirk Petry
• Martin Pokorny
  • Initial prep. for Asynch I/O. Playing around in vi/vb2. Tried running unit tests. One seems good, but test coverage is not good.
• Federico M Pouzols
  • CAS-9473 (new mode in flagdata to flag all integrations in which an antenna is flagged): ready for commit and verification/validation. It was scheduled for 5.1 but it could be included in 5.0 if still possible.
  • From comments in 'antintflag': maybe we need to fix/clarify FLAG, FLAG_ROW?
  • CAS-5808 (rflag's use of timedev and timedevscale parameters): what is the status of this (very old, but more recently there seemed to be some consensus). Should we schedule it for 5.1 or close it?
• Urvashi Rao
  • Bug fixing and documentation.
• Darrell Schiebel
  • fixed logsink.logfile() problem
  • fixed casa's ipython rc & history location
  • fixed --pipeline 'agg' issue
  • finished up Intel compiler build loose ends
  • curated JIRA tickets
  • workflow discussions
  • Jeff does not want to distribute astropy with CASA. Darrell recommends keeping pyfits even though we cannot update. Jeff recommended adding a comment acknowledging that further updates are not possible and pyfits may be discontinued.
  • Adam ginsburg is working on RadioTools, which is similar to astropy, but focused on radio astronomy. Jeff has not had a chance to talk with Adam about collaboration.
• Ville Suoranta
  • Added OSX1012 to Accepted tests
  • Git workflow discussions
  • Investigating and fixing pipeline problems
  • Testing packages compiled with Intel compiler
• Takahiro Tsutsumi
  • Checked in the remaining set of the modifications to automask for 5.0
  • Added a few more plone contents for masking
  • Looked into the coordinates issue for the parallel continuum tclean (related to CAS-9518)

Friday NAOJ Meeting

• Kanako Sugimoto
  • GIT workflow tests
    • code/gcwrap workflow using CAS-8449 (refactoring of atmosphere tool funcions): completed
    • asap workflow using CAS-9502 (dummy ticket): completed but I see workflow for submodule may end up with no-op depending on timing.
  • CAS-9485 Support additional header tables of NOSTAR in importnro: done
  • CAS-9015 Update of ATM library in CASA 5.0: done
  • support establishing environment to run pipeline assisted QA2 at EA-ARC
• Wataru Kawasaki
• Masaya Kuniyoshi
• Takeshi Nakazato
• Renaud Miel

• Some test failures in bamboo: across platforms - ic2233, ~6 failing on RH 7

• • Input from Jeff still needed