Monday Morning Meeting July 6

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

Attendance

  • Socorro: Jeff, Tak, Lindsey, Sanjay, Urvashi, Juergen, Martin, Kumar, George
  • CV: Andy, Akeem, Ville, Dave, Darrell
  • Garching: Justo, Sandra, Dirk, Julian
  • SCO:

News / Meetings / Visitors

  • PMD Eduardo Visit
    • CV: 7/16 and 7/17
    • Soc: 7/22
  • CUC: Telcon this Thur.

Build, Release, Testing

  • Able to incorporate CASA core today.
  • Git, stash, and bamboo

Architecture

  • Github/Casacore integration finishing soon. It builds in my sandbox and after a commit of two files should be able to build for everyone (the commit has to happen at the same time as we cut over to github casacore).

HPC

  • Should we turn off parallel go?
    • Drop support when we migrate to ipython.
  • First telecon with NAOJ today: minutes are in https://ictwiki.alma.cl/twiki/bin/view/DataProc/SDPipeParaMeeting20150706
  • The creation of an MMS inside importasdm needs a simple modification to accept an MS with a FLOAT_DATA column. This has to be automatic inside the task because a datacolumn parameter doesn't exist. In progress.
  • Fixed the creation of the .flagversions file inside applycal when working on an MMS. It was creating individual flagbackups per subMS.

Development

  • Removal of obsolete code:
    • Twiki on obsolete code to be produced.

Pipeline

  • Support for tier0 plotms and calibrator imaging ported to pipeline trunk
  • Import data tasks modified to support createmms option
  • Refactored flagging code testing
  • Image code cell size heuristics improvements, same cell size for all receiver band / spws rounded to 0.01arcsecr
  • Some concerns about ephemeris table support for ALMA data in CASA and pipeline

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
    • HPC activities as reported above.
    • Implemented support for spw selection in the online flag commands created by importasdm (CAS-7623). There still seems to be an issue with the spw names created online, where SQLD basebands may not be treated consistently in the flags. This is being discussed in http://ictjira.alma.cl/browse/ICT-1736.
    • Working on adding support for FLOAT_DATA in importasdm for the MMS cases.
  • Lindsey Davis
    • Added PWV plots to web log, testing for all pipelines
    • Added support for antennapos.csv file to hif_antpos
    • Investigated EVLA pipeline export / restore data issues with Brian K.
      • EVLA BDF flags have a format problem
      • importasdm / importevls differences cause problems
  • Pam Ford
  • Kumar Golap
  • Justo Gonz├ílez Villalba
    • Implemented a mechanism to change the log level of mpi4casa logging at run time
    • Prepared a proposal for a generic method to propagate weights for linear frequency axis transformations
    • Investigated and fixed critical plotms bug related with velocity transformations.
    • Analyzed tclean multiple write problem, and concluded that casacore does not support it.
  • Jim Jacobs
  • Jeff Kern
    • Misc. Administrative Tasks
  • Susan Loveland
    • Fixed miscellaneous CARTA bugs.
  • David Mehringer
    • Answered Peter Teuben's questions re "changing shape of an image".
    • Had discussions about whether MSMetaData should support invalid MSes which have no STATE subtable but nonnegative STATE_ID values in the main table. My opinion was that they shouldn't because there were fairly non-expensive ways (in my opinion) to make these types of MSes valid, or possible API modifications, such as object configuration or specialization, that would allow the developer to explicitly say whether or not she wanted this type of support (in fairness I only thought about that possibility after the primary discussion had occurred), but I was essentially overruled.
    • Voluntarily made personal decision to relinquish all my GitHub casacore module coordinator responsibilities. Asked Tammo to remove my admin invitation to GitHub which he did. I'm afraid the current #Thunderdome model of conflict resolution (or whatever it's called when there is no process in place for resolving such disagreements) wasn't going to work well for me.
    • Responded to Martin's question re: stats framework and visstat.
    • Sought lots of help from Julian re: git. Thanks Julian!
    • Told a local user I couldn't immediately look at his imregrid issue and suggested he create a JIRA ticket for it, which he promptly did.
    • Wasted a couple of hours trying to understand why I couldn't compile NRAO svn casacore. Found an email regarding the issue I was having that Julian sent around a few weeks ago which resolved my issue. Thanks Julian!
    • Answered Remy's question about how best to mask whole channels of an image. There is an LEL function which allows one to do this sort of thing.
    • (CAS-7319) MSMD additions: Ger has merged in my changes (along with his changes to support certain types of invalid MSes described above) into GitHub casacore master. I still have gcwrap changes I must commit, but I cannot do that until our code/ tree is able to compile against GitHub casacore.
    • (CAS-7703) imfit and imstat will not work with polygon region if chans parameter is also set: Resolved as Won't Fix. This is exactly how this was designed to behave for the reported combination of box/chans/stokes and region parameters. Granted I made that design decision, because no one supplied me with requirements for this use case. I pointed out that this has been the behavior for years, probably since the advent of CRTF region files several years ago, and that this is the first time I had received any complaints about this behavior. Discussions about various levels of communication regarding software requirements and user needs ensued, as well as on CRTF format and documentation in CASA guides regarding these files. Fortunately, I think all of this discussion was handled well and was very productive, with me acknowledging that I understood the issues raised, but also saying that, in order to resolve those issues, I needed clear and specific requirements that preferably had been agreed upon by a CSSC consensus. This discussion led to the creation of (CAS-7704) Image analysis tasks should have 'chans' and 'stokes' to override par.region specification, which has beautifully detailed requirements. Hooray for requirements!
  • George Moellenbrock
    • (CAS-7287): Continued work on VI2 in solving. Should begin testing later this week.
  • Dirk Petry
    • visited ASTRON
      • discussed further steps on a unified approach to updated the measures tables (CAS-7401)
        • agree to at least for this year keep two sites which serve the measures tables: ASTRON and NRAO
        • agree to set up a job which will regularly (weekly) download the tables from ASTRON and NRAO and confirm that they agree (action item on Dirk, CAS-7710)
        • agree that ASTRON will update the magnetic field information to the latest version (action item on ASTRON)
        • agree to fix the data type problem in the DE ephemeris tables (dMJD should be float, not int) (action item on Dirk, CAS-7711)
        • agree to provide the user with a script to (a) get the data and (b) test that it is correctly installed (Dirk, CAS-7712)
        • agree to put information about how to get and install the data on the github casacore web page (ASTRON)
        • agree to have a script which makes it easier (for CASA personnel) to add entries in the Observatories table (Dirk, CAS-7713)
      • learned about github usage
      • Met with Mark Kettenis on requirements for importfitsidi which have become apparent from the JIVE work on the fringe fitter
        • importfitsidi needs two new optional parameters to control obsid and scanid settings (CAS-7714)
    • started to look into CAS-7709, a fixplanets problem from Todd
    • resolved (CAS-7570) Make wvrgcal produce identical output under RHEL and OSX
  • Martin Pokorny
  • Urvashi Rao
    • Started coordinating tclean testing and responding to user queries/bugs (clark/hogbom, mosaic, interface, pblimit fix...)
    • Started inline docs for the tclean task
  • Darrell Schiebel
  • Ville Suoranta
  • Julian Taylor
  • Takahiro Tsutsumi
    • Worked for 2 days... Added 'channel selection flags' to handle gaps within a spectral window as well as strides and enabled 'tuneselectdata' in tclean. There seems to be a bug in one of underlying functions causing channel selections to be shifted (still investigating).
Friday NAOJ Meeting
  • Kanako Sugimoto
  • Wataru Kawasaki
  • Masaya Kuniyoshi
  • Takeshi Nakazato
  • Renaud Miel
-- JeffKern - 2015-07-02-- JeffKern - 2015-07-02
Topic revision: r18 - 2015-07-06, JeffKern
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