Monday Morning Meeting 07 MARCH 2016

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

Attendance

  • Socorro: Juergen, Jeff, Jim, Susan, Pam, Tak, Martin, George, Kumar, Urvashi
  • CV: Andy, jen, Ville, Darrell, Akeem, David
  • Garching: Dirk, Justo
  • SCO:

News / Meetings / Visitors

  • Daylight Savings Next Week
  • Mark Kettenis - Jive
  • Rob Latham - Argon National Lab
  • CUC Meeting Last

Build, Release

  • Branch today:
    • Copy ephemeris into Cal tables
    • Sandra's fix to FlagData
    • Dirk: Fix Clouds
    • Urvashi: Unit test failing on wide field imaging

Verification Testing

  • Build and test team discussion on using containers for automated testing (as opposed to virtual machines)
  • Imhistory is the only failing test (both on RH and OSX) - recently added to the accepted test list
  • Spreadsheet of current tests has been compiled

Validation Testing

  • Tall Poles:
    • DONE CAS-8196
    • ALERT! CAS-7924
    • ALERT! CAS-8313

Architecture

  • Technical Forums:

HPC

  • We had a telecon last Thursday where we mainly discussed the integration of MPI in C++ in CASA. We have decided to try the alternative approach to passing a communicator from the python controller process to C++, create the communicator via the MPI3 non-collective communicator creation functions which should allow to use it without adding any extra logic to the controller process. For that Julian will create and test MPI 3 rpms.

Development

  • CASA-4.7 Planning Mostly Final

Pipeline

  • Waiting for input from CV team

AOB

Developer Reports

Monday Meeting
  • Sanjay Bhatnagar
  • Sandra Castro
    • I am unable to attend the telecon today as we are having the ESO overview this week.
    • Implemented an overwrite sub-parameter in task flagcmd when savepars=True. By default, it will overwrite an existing file when saving the parameters. CAS-7852.
    • On the same subject, we should align the behaviour of flagdata and flagcmd when savepars=True. Currently, it will append the parameters of the task to a file if it already exist. The new behaviour will not append to the file. I need to know if this change is fine with the pipelines. See CAS-8391 created today.
    • Improved the inline help of flagdata for the quack mode. CAS-7565.
  • Lindsey Davis
  • Pam Ford
    • CAS-8337 RTT (plotms problem with multiple observation IDs) - plotms spw averaging bug
    • Investigated CAS-8363 (QObject/QApplication messages when running plotms from tarball casa session) - cannot reproduce except on own machine. Will leave ticket unscheduled and see if any new reports come in. Have fix ready if needed.
    • Investigated CAS-8270 (Implement remaining plotcal functionality in plotms) and created subtasks
    • CAS-2656 (plot cal table SNR in plotms) - subtask to CAS-8270 - ready to commit and test in 4.7
  • Kumar Golap * Finished FT Machine move to VI/VB2 * Ready for Sanjay
  • Justo González Villalba
    • Work in progress migrating mstransform re-gridding code to an independent TVI
    • Investigation of differences between clean and tclean in ALMA M100
  • Jim Jacobs
    • Looking at Kana's growing MS problem.
  • Jeff Kern
    • 25 Meetings last week
  • Susan Loveland
    • Viewer: Spent most of the week dealing with spectral lines having Doppler shifts > 1 with help & advice from Kumar.
    • CARTA: Finished compiling layer group code reorganization, started working on runtime problems.
  • David Mehringer
    • (CAS-6896) immath should give control over file used for header generation: Updated docs as requested. Cannot use mustexist attribute for imagemd parameter in order to color parameter red in inputs if image does not exist since mustexist does not allow an empty string which is the default value.
    • (CAS-8365) imstat Logic Error: Resolved. Not a bug. This no longer appears to be an issue as it was likely serrendipitously fixed at some point after the version in question that was used to discover it.
    • (CAS-8357) remove LCEllipsoid constraint that center must be inside lattice: RTT. Fixed. Developed algorithm for computing masked pixels when the center of the specified ellipse lies outside the lattice.
    • (CAS-8368) add history for imcollapse(). Resolved. History added.
    • (CAS-8369) write history records for ia.imagecalc(). Resolved. Writing of history added.
    • (CAS-8372) enable ia.imageconcat() to write history to new image: Resolved. Writing of history added.
    • (CAS-8373) enable ia.fromarray() to write history: Resolved. Writing of history added.
    • (CAS-8367) add history to image for ia.addnoise() call: Resolved. Writing of history added.
    • (CAS-8382) ia.fromshape() should write history: Resolved. History writing updated to use standard interface.
    • (CAS-8378) ia.fromfits() to write history: Resolved. Writing of history added.
    • (CAS-8379) ia.fromimage() to write history records. Resolved. Writing of history added.
    • (CAS-8376) ia.toascii() should write history: Resolved. Writing of history added.
    • (CAS-8139) non-closing files in exportuvfits: RTT. Communicated with Jen about reproducing the original issue in a version earlier than my fix.
  • George Moellenbrock
    • CAS-8313:
  • Dirk Petry
  • Martin Pokorny
    • Updating tClean2 prepared to check it in. Also bring in Julian's patches for the communicator.
  • Urvashi Rao
    • Enabled redrawing of automasks after every major cycle
    • Re-tested some widefield uniitests that are sometimes producing NaNs - it's a heisenbug.
    • Other non-casa work
  • Darrell Schiebel
    • Looking into distributed testing
    • Need to port CASA QT to QWT-6
  • Ville Suoranta
    • Working on Bamboo build and test mock-up, will distribute issue via e-mail
    • Looking at QWT compatibility issues
  • Julian Taylor
    • Updated openmpi rpms to 1.10.2 and mpi4py to 2.0.0, regression tests work but an issue in an error-case still needs solving.
    • my patches to the SD pipeline from last year have been applied by Kana, pipeline is 30% faster now.
  • Takahiro Tsutsumi
    • Fixed a Bocker ticket, CAS-8196 - found a bug in the old Imager code, imagecoordinates2() for reusing internally stored the telescope position this caused an error in clean in the default width and when trying to use tclean image with (the old imager tool based ) makepb.
    • Added a description of the current outframe behavior in tclean help
    • Various tclean related discussions
    • Add Sun's ephemeris table in the data repositiory per Bryan Butler's request
Friday NAOJ Meeting
  • Kanako Sugimoto
  • Wataru Kawasaki
  • Masaya Kuniyoshi
  • Takeshi Nakazato
  • Renaud Miel

-- PamFord - 2016-03-04

This topic: Software/CASA > Software > WebHome > CASAMeetingPage > Mar0716
Topic revision: 2016-03-07, 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