Agenda

Minutes in bold

CASA/JIVE VLBI Coordination Meeting

Attendance: George, Mark, Des, Ryan, Justin, Neal

2021 Feb 24 @ 1500UTC

Logistics

  • Future Meeting dates (usually last Wed of the month):
    • 2021 Mar 31 @ 1400 UTC (NB: DST start March 14 & 28 in USA & Netherlands, respectively)
    • 2021 Apr 28?

General Admin

  • General Status w.r.t. Coronavirus measures
    • NRAO/ALMA:
      • NRAO in North America continues work@home for foreseeable future. Infections rates declining, vaccinations ramping up.
      • ALMA progressing with array restart plans; 10+ antennas science ~ready as of Feb 11; decisions expected ~this week on starting "Initial + Basic Science Operations Phase"
    • ASTRON/JIVE
      • Science/computing staff continuing work-at-home for now
  • CASA Release Schedule
    • CASA 5.8/6.2: Development finished. Validation wrapping up. VLBI-specific tickets reviewed below...
    • CASA 5.9/6.3: "Bug-fix" release ~June?
    • CASA 6.2.1: ALMA-specific (PL) bug-fixes ~September
    • CASA 6.4: End of year?
  • Requirements collation: 2020 May 07 document / VLBA Costing/Prioritization (2020 May 21)
    • CASA 5.9/6.3 (and 6.2.1) Planning underway
  • (Ryan) CNGI/ngCASA Status:
    • Working toward demo 'release' in March, marking end of prototyping phase, and beginning transition to science-driven development, and including collaboration with community
    • FYI: CNGI Development instructions
  • Admin issues
    • CASA Docs now hosted on readthedocs platform: https://github.com/casangi/casadocs
      • Edit permissions for JIVE provided by Bjorn/Ryan
      • (Mark) How does xml ingestion from code base work? (Ryan) A github 'manual' operation triggers it. On master, to be initiated at will, generally. On release branches, with some care/coordination.
    • DIY casa5 tarball build instructions can be found here
    • casa6 build issues: formal instructions for plotms, CASAshell, casa5/6 co-exist, c++ unit tests in casa6.
  • Meetings/workshops of relevance
    • Ilse has submitted URSI GASS (J11) abstract (Rome, 28 Aug - 4 Sep)
    • CASA Internal Virtual F2F meeting Feb 1-5
      • (Ryan) Socialization for old and new core CASA team members; planning to have a broader brief similar meeting with external contributing groups, including JIVE, ARDG, PL, CARTA, etc.
    • Jan-Willem gave great summary of CNGI/ngCASA to EHT SDC group Feb 11. Well-received, some followup expected...
    • Mark to give short talk at EAVN VLBI meeting (Thailand, building 40m dish that will also be used for VLBI), 2021 Mar 5
    • RCI seminars: https://www.radiocamera.io/seminars. (Current series wrapped up 19 Feb, but slides/video for many talks are available.)
  • VLBI Integration Branch: CAS-13295
  • Other admin?
    • (Mark) Sandra has switched to new testing DR. Something of a surprise to external developers like JIVE.
    • (George) Should casa-staff email list be expanded to include external groups (and thus shower them with all manner of daily trivia)? Or should there be a distinct email exploder expressly for important announcements relevant to the broader CASA contribution community?

Development

  • CASA 5.8/6.2:
    • JIRA (VLBI) Status
      • Notables:
      • (George) CAS-12735: Frequency meta-info (includes generalized fix for CAS-12553, and several other bug tickets)
        • Progress continues, but now deferred to 6.3....
        • Some difficulty to do with managing implicit spw selection and proper resolution of frequency meta-info in such cases. (In any case, freq meta info will be discerned prior to solving iteration rather than sensed during, as now. Future to include user-specified spw fan-in...)
      • (Des) CAS-12765 (wide-spaced spws): waiting for freq meta fix in CAS-12735
      • (Neal) CAS-13184: accor corrdepflags. DONE.
      • (Mark) CAS-8537: gain curve. DONE.
      • (Neal) CAS-12736 (revised accor test). DONE.
      • (Mark) CAS-12989 (eMerlin crosshands) Validation complete (Anita Richards). Need hard-to-get eMerlin dataset for proper verification test....
      • (Pam Harris) CAS-11318: Fringe-fitting solution plotting has been implemented in plotms! DONE.
      • (CAS-13056, CAS-13003, and possibly CAS-13005 to slip to next cycle)
  • CASA 5.9/6.3 prioritization underway
    • (Justin) Anything to report? Current plan not favorable for (NRAO-based) VLBI requests, so far. Discussions/finalization TBD (meeting today).
      • (George) Plotting and gaincurve already done for 5.8/6.2.
      • (George) Several of the more interesting VLBA-specific feature requests (resolved polarization calibrators, VLBA pulse-cal, VLA SysPower for Tsys, TEC) may indeed be hard to schedule this year.
      • (Mark) Likely to support pulse-cal effort with ingestion definition and mechanics based on recent similar work for Tsys, gaincurve. NB: required algorithmic development will likely borrow heavily from the existing fringefit implementation, including an ambiguity-resoloving step using a fringefit solve on a short scan
      • (George) Scan intent-setting likely a good project for Neal (pushback from PL probably just due diligence queries to stakeholders)
  • Other development topics
    • Nearly ready for validation/merge-to-master for 5.9/6.2 (post-5.8/6.2 branching)
      • (Mark, CAS-13270) Handle NaNs in imported Tsys.
      • (Des, CAS-?) memory management improvements in fringefit
      • (Mark, CAS-8550) Pulsar-binned and single-IF import now not blocked by sys power support in exportuvfits; Need to coordinate casacore merge after 5.8/6.2 branching
    • Ongoing for 5.9/6.3
      • (George, CAS-12735, etc.) Frequency meta-info
      • (Mark, CAS-13287) AveragingTvi2 preavg opt-out for fringefit
      • (Mark, CAS-13266) Per-scan interpolation via overloading per-obs mechanism
      • (Des, CAS-?) Overlapping solution intervals (initial implementation awaiting review by George)
    • Underway for 5.9/6.3
      • (Mark, CAS-13383) Mixed polarization basis conversion for EHT
        • To be implemented via a generic freq-dep fully general (4-element) Jones matrix, derived from JJones class. This Jones matrix will include the (complex) polarization balance solution and the required basis conversion (solved for in python and deposited into an otherwise standard caltable)
        • One complication: The correction is to be applied upstream of FITSIDI formation, but correlation order is currently non-cannonical (PP,QQ,PQ,QP) at this stage. Mark to bug Harro about improving this; in meantime, consider correlation-axis sorting or revised matrix algebra.
      • (Des, CAS-13361) Polarization combination for fringefit
        • Squish correlation in VB2 cd channel-averaging. Does this mean a new VI2 layer?
        • (Mark) For fringefit, should probably generate duplicated solutions into both polarizations of existing fringefit caltable convention (2x4 parameters); cf gaincal('T') which solves for an intrinsically single-pol TJones solution (a whole VisCal specialization differing only trivially from dual-pol GJones)
        • (George) Mechanism will be adopted by TJones. Should formally form Stokes I = (PP+QQ)/2, not weighted average of parallel hands
        • (George) Should also support selection of one or other parallel hand, to support solving for time-dep tropospheric coherence correction to permit an improved estimation of polarization-relative solution.
    • Other/Miscellaneous
      • (Justin) f.d. scaling difference with AIPS? Lucas Hunt at USNO. cf AIPS ACSCL (avoid edge channels in accor)?
      • CAS-12485: OTF weight munging for solve (Can we complete validation and deploy? Des seeking interested users at JIVE)
      • (George/Justin) to coordinate initial testing on CASR-463 (TEC) to better characterize basic requirements (also CAS-13040 re TEC retrieval changes)

AOB

  • n/a
-- GeorgeMoellenbrock - 2021-02-24
Topic revision: r6 - 2021-03-31, GeorgeMoellenbrock
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