Monday Morning Meeting DATE

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

Attendance

  • Socorro: Rob, George, Sanjay, Urvashi, Tak, Pam, Juergen
  • CV: Mark, Dave, Andy, Darrell, Dirk
  • Garching: Sandra
  • SCO:

  • Apologies: _Jeff at ALMA ICT. Justo at OpenMPI course this week.

News / Meetings / Visitors

  • Masaya Kuniyoshi: Joined CASA development team at NAOJ
    • Previously a post-doc at NRAO NM, then to LOFAR, now NAOJ.

Build, Release, Testing

  • 4.3 branching today.
  • 4.3 pre-release packages
    • Branched from stable version, can be used for testing immediately. Stables for EL5 and EL6, but need OSX.
  • 4.3 testing
    • CV has RPMs, have four machines setup for user testing.
    • Sent Steve Myers the links to the tarballs. AOC testing requires stable roll out. ACTION: Rob to roll out in AOC.
    • Mark also working with Kana to get the latest stable and start testing.

  • 4.4 test packages started last night. Failed due NFS, but Darrell restarted this morning.
  • Avoid merge trouble - Developers to please commit fixes to trunk and branch.

  • New package types (test, stable, monthly), naming and numbering
    • Wiki materials to be circulated to team and users - ACTION: Mark to circulate material after the wiki.
    • Implementation of changes over next few weeks, assuming no major issues identified by team.
    • Juergen will have to update user facing pages once finalized/adopted.

  • JIRA workflows & screens
    • Now implemented for the CASA project.
    • Wiki link
    • George may have found a bug with older tickets. Couldn't move straight to RTT. May be reviewed.
    • ACTION: Review old tickets, reassign if team member no longer here.

Architecture

  • Technical Forums:
    • CASACore path forward (TBD)
    • B&T Requirements
    • B&T System Redesign
    • TClean
    • Imaging Pipeline
    • Structure of flagging classes and flagging algorithms
    • Vi/VB2 implementation (already presented)
    • CARTA Architecture (when ready for discussion)
    • ADMIT Project
    • VLBI
    • MSTransform heuristics & MMS - Nov 13?

  • Casacore Unification
    • OK to commit changes now to trunk.

  • Statistics Framework
    • Dave: Caller's responsibility to ensure all data is available to framework. Framework should not have to cache data.
    • Not a consensus on this approach. Need to handle large data sets.
    • Users of statistics framework to discuss with Dave.

Development

  • HPC meeting on Thursday
    • Justo's revised MPI document
    • Share information with Juergen.

  • 4.4 Planning
    • Have input from CASA team, some collaborators, ALMA Pipeline, VLA Pipeline, CSSC (via CUC presentations), previously flagged items in JIRA, etc.
    • Team review of the target list at MMM on 10/27.
      • Jeff will distribute in advance.

Pipeline

AOB

Developer Reports

Monday Meeting

  • Sanjay Bhatnagar
  • Sandra Castro
    • added MMS support to task visstat. Basically only the tests needed changes, not the task. (CAS-4247)
    • Looked at MPI issues when running test_flagdata and test_partition.
    • Verified that split2 is also fixed after the POLARIZATION tables were fixed. (CAS-6178).
    • Verified that the problem with flag commands in importasdm is not a bug. By default it appends the cmds to the given file. Ticket CAS-6805 should be looked at and requirements should be given to solve this. (CAS-7006)
  • Lindsey Davis
  • Pam Ford
    • CAS-6939 - r31286 - plotms grid can go into infinite loop when drawing
    • CAS-6974 - in progress - added buttons to maximize/minimize or close plot, but found existing bug that causes seg fault when plot is closed
  • Kumar Golap
    • wrote a small standalone app to remove undefined spectral frame from an MS
    • Worked mainly to fix an issue with msuvbin where the source peak was down by 10% ..this was due to the PSF being different and the cause was due to a rounding error in uv nearest point while gridding
  • Justo Gonz├ílez Villalba
    • Successfully tested and documented a [build/stand-alone packaging/torque integration] procedure for the OpenMPI -based CASA paralell framework
    • Changed mpi4casa send mode to use a defined buffer in order to fix an issue when sending large flagdata dicts
    • Fixed a mstransform issue when spliting ALMA MSs containing WVR (CAS-6941)
    • Fixed a mstransform issue when merging array columns with variable shape (CAS-6206)
    • Followed up science testing
    • NOTE: Next week (13-17Oct) I'm attending a MPI/OpenMPI course
  • Jim Jacobs
  • Jeff Kern
  • Susan Loveland
  • David Mehringer
    • CAS-6354 Rewrite ImageAnalysis::getFreqProfile(). Passed what looked like a viewer-specific bug along to Susan, and then passed back to Andy for more testing after her fix. Andy tested and confirmed Susan's modification fixed the issue.
    • Responded to query re: tMSMetaData Jenkins failure. All casacore tests continue to pass on trunk in my sandbox, so I couldn't offer any assistance without further information on the Jenkins failure.
    • (CAS-6896) immath should give control over file used for header generation. Responded to Ger and asked Juergen if it would be OK if the user specified which image to use, rather to rely on the first image that appears in the expression as per his requirement.
    • Responded to more of Jim's questions re: stats framework. The requirement my design makes use of is that it be the caller's responsibility to ensure all data sets which are to be used be simultaneously accessible (eg, in the case of a set of data which cannot fit in memory, that it is
      the caller's responsibility to ensure that these data are cached appropriately so that the stats classes have access to all of them, ie it should not be the stats framework's responsibility to cache data). If that requirement is withdrawn then it because much more costly and much more difficult to design and implement the stats framework if it is to be responsible for caching data.
    • CAS-6979. imfit should have an option to report the size of a uniform disk (rather than a Gaussian) that would be consistent with the image fitted size when convolved with the beam: Pointed out that I don't know how simple this is to do because I am not aware that the convolution of a disk with a Gaussian results in a function that can be expressed in a form that can be supported by our fitting framework. This may need input from someone who can design an efficient algorithm.
    • Updated three tickets to v 4.4 and/or "on hold" status as per Jeff's email request.
    • Took new JIRA workflow for a test drive.
    • (CAS-4889) immath leaves open IERS table: fixed by adding call to me.done() in tearDown() method of test class.
    • Responded to Jim's casacore mege email by giving a heads-up that Malte has not completed ASAP modification to remove casacore-components dependency so components is still in a partially migrated state of limbo until that happens. Also suggested that there should be internal written guidelines and processes for our dev team around interactions with casacore and this partnership going forward so everyone is on the same page. I feel the current state of affairs regarding casacore interactions is ambiguous, at best.
  • George Moellenbrock
    • Attempted to turn on WEIGHT_SPECTRUM in solving context for 4.3 via old VB framework, but failed.
    • Out sick part of Wed and Thurs.
    • Fixed a few minor bugs for 4.3 CAS-4047, CAS-6755
  • Dirk Petry
    • wvrgcal integration into code base. Part to stay as a 3rd party package.
    • Tackling wvrgcal tickets.
    • Working with Michel on ephemeris handling in asdm2ms, etc.
  • Urvashi Rao
    • User support & documentation.
    • Bug related to async parameter, fixed by Shinnosuke. Fix should go into 4.3 branch - Urvashi will remind Shinnosuke.
  • Darrell Schiebel
    • ironed out problems with regression test script
    • working on extending wrapper script for local versions in ~/.casa/versions
    • will create branch https://svn.cv.nrao.edu/svn/casa/branches/release-4_3 (assuming that patches will be drawn from this branch as it evolves)
  • Rob Selina
  • Julian Taylor
  • Takahiro Tsutsumi
    • automask - work on design of infrastructure
    • went through some of the open 4.3 tickets
      • CAS-6658: did some tests on optical velocity spectral grid, sent question to Dirk
      • CAS-6915: looked at the scripts for ephemeris data retrieval and format conversion

Friday NAOJ Meeting
  • Kanako Sugimoto
    • review and discussion of CASA 4.3 development cycle. many thanks for valuable discussions and suggestions.
    • defined schedule and assignments of CASA 4.4 developments. Creating JIRA tickets for them.
    • Introduction to CASA developments for Masaya
    • discussing Darrell how to manage bug fixes to ASAP and how to merge it to release branch.
    • input my thoughts on proposal of changes to packaging, deployments, and workflow.
  • Wataru Kawasaki
    • verifying SD unit tests for flag handling implemented by Kana : finished
    • commented on user testing report by Bunyo of SD unit tests for flag handling
    • built standalone ASAP
  • Masaya Kuniyoshi
    • regression tests (ALMA sd data) : ongoing
  • Takeshi Nakazato
    • get back from childcare leave
    • tracked asap build failure report (found that r2986 by me is the origin)
    • responded testers: CAS-6582, CAS-6588
    • commented on NRO bug report CAS-6986
  • Shinnosuke Kawakami
    • tget problem was resolved(reported by Urvashi)
    • GC-casacore unification: Ger will send email about changes to build CASA with GC-casacore

-- JeffKern - 2014-10-06
Topic revision: r19 - 2014-11-06, MasayaKuniyoshi
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