Monday Morning Meeting 8/27/2018

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

Attendance

  • Socorro:
  • CV:
  • Garching:
  • SCO:

News / Meetings / Visitors

  • SRDP busy week in Socorro, wave 1 requirements/design discussion
  • CASA Architect interviews underway
  • CASA SW Engineer posting going up today
  • No meeting September 3 - NRAO holiday

Development

  • CASA 5.5 Development planning
    • tclean maturity
    • HPC automated test suite
    • VI2/ mstransform / cvel / uvcontsub completion

Pipeline

  • Development team started to work on Pipeline release notes
  • E2E and Acceptance tests continued. Issues still on work
    • pipeline
      • A strange weblog plot in hifa_tsysflag ( CAS-7920): under investigation
      • use of by field flagging in hifa_tsysflag ( CAS-8459): PWG testing existing capability
    • casa
      • strange noise pattern in mosaic='gridder' in tclean ( CAS-11816 Blocker): Kumar introduced mosweight parameter but it has issue in parallel run.
      • ia.statistics crashed in a pipeline run ( CAS-11828 Blocker): only robust=True shows the issue. Dave will look into it in early Sep.
  • Two _pg interface issues/behaviors found: what happens to parallel task interface in CASA 6.0?
    • subparameter does not respect main parameter dependent defaut parameter value definition
    • task execution parameters are not print in logger (intentional?)

Users

Build, Release

Verification Testing

  • Pipeline test is currently failing - once Brian K. returns to the office, we will check in with him about a fix
  • CAS-11539: For the regression tests we are getting false positives even though there may be tests that fail. Currently being investigated

Validation Testing

CASA 6

HPC

  • task_cli vs. task_pg.
    • The ALMA pipeline has reported a few issues coming from the use of _pg in their code. A bit of background about the _pg interfaces of tasks. "pg" stands for "parallel_go, which was the old parallelisation framework used in CASA many years ago. The normal interfaces of CASA tasks are defined in the _cli.py files. The _pg was introduced in the past so that *_pg wrappers would skip things which only mattered in the interactive environment like writing *.last files, or in general relied on the existence of an ipython shell (which is not loaded for the MPI servers). The automatic creation of the _cli and _pg tasks are defined in two files: casa2pycli.xsl and casa2pypg.xsl.
    • We have noticed some differences between the 2 templates that carry potential risk for trouble. In the _cli template, the defaults() function is different from the one in the _pg template. The code between lines 298-314 in casa2pycli.xsl sets the defaults of some sub-parameters according to the value specified in the parent parameter. This does not happen in the casa2pypg.xsl template. https://open-bitbucket.nrao.edu/projects/CASA/repos/casa/browse/gcwrap/install/casa2pycli.xsl#298-325. The result of this is something like shown below for task_sdcal for example.
    • The funtcion itsdefault() inside sdcal_cli.py defines the value of the sub-parameters differently depending on the value of the parameter calmode. The sdcal_pg.py does not define this!
if self.parameters['calmode'] == 'otfraster':
    a['fraction'] = '10%'
    a['noff'] = -1
    a['outfile'] = ''
    a['overwrite'] = False
    a['intent'] = 'OBSERVE_TARGET#ON_SOURCE'
    • We already created a ticket about a year ago ( CAS-9871) to unify the two templates (cli and pg) so that only cli is used. Lately the pipeline has moved to import many tasks from the _pg interfaces, which it seems they should not have done. We need to clarify what are the requirements for the tasks that run in the MPI Servers: should they use the list of globals from the Client?; should the .last be written in the Servers?; should the parameters list be shown, etc.?

CARTA

Developer Reports

Monday Meeting
  • Sandra Castro
  • Kana Sugimoto
  • Bjorn Emonts
  • Pam Ford
  • Enrique Garcia
  • Bob Garwood
  • Kumar Golap
  • David Mehringer
  • George Moellenbrock
  • Dirk Petry
  • Federico M Pouzols
  • Darrell Schiebel
  • Ville Suoranta
  • Neal Schweighart
  • Takahiro Tsutsumi
Friday NAOJ Meeting (8/23/2018)
  • Wataru Kawasaki
  • Takeshi Nakazato
    • reviewed existing tickets for 5.5 development
    • CASR-295 ephemeris handling: work on Solar simulted data
    • Assigned CAS-11818 to Sumi for his training on JIRA workflow
  • Renaud Miel
  • Suminori Nishie
    • CAS-11427 Unit Test (UT) for PointingDirectionCalculator
      • coding Google-Test Programme
      • working on Moving Source calicuation. (a set of 4 APIs.)
      • ciscovered a bug by this Unit Test, CAS-11818 was created. Sumi is assigned
      • MS table manupilation mostly OK, good level to cover exceptional cases originated from MS.
    • CAS-11818 Moving Source calculation may cause crashes in specific API sequence
      • Fix this bug until Resolved in JIRA. Acuire JIRA / git / verification operation.

-- TakeshiNakazato - 2018-08-21
Topic revision: r18 - 2018-08-27, PamFord
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback