Pre-spectral line (PL2023 and earlier):

Prior to the VLA spectral line pipeline, the pipeline was optimized for continuum only and bright spectral lines could be flagged as RFI and downweighted by statwt. The restore workflow had two tasks and output an RFI flagged MS with the weights reset.


  • Pipeline script:
    • hifv_restoredata()
    • hifv_hanning()
    • hifv_statwt()
  • Products
    • calibrated measurementset $SDMID.ms (has RFI flagging applied)

The template_hifv_deliver_ms.xml should run on the $SDMID.ms needs to specify data types for importonly mode datacolumns={'data':'raw',corrected:'regcal_contline_all'}.

The SSA restore workflow should operate as follows:

  • If original CASA version < PL 2024
    • Restore with original CASA (when possible)
    • Run a second CASA session with  template_hifv_deliver_ms.xml, which will produce an additional MS (_targets_cont.ms).
  • If original CASA version >= PL2024
    • Restore with CASA PL2024 (will produce _targets_cont.ms and when applicable _targets.ms)

Post-spectral line (PL2024):

The VLA spectral line pipeline will spawn multiple measurement sets, one with calibrators and RFI flagging applied, which is the same as the previous workflow. To avoid compromising spectral line data, the $SDMID_targets.ms will have RFI flags removed, the $SDMID_targets_cont.ms will have  RFI flags applied and is suitable for continuum-only work.

  • Pipeline script:
    • hifv_restoredata()
    • hifv_hanning()
    • hifv_statwt()
    • hifv_mstransform()
  • Products
    • calibrated measurement set $SDMID.ms (has calibrators, RFI flags applied to target)
    • calibrated $SDMID_targets.ms   (if spectral line, has RFI flags removed)
    • calibrated $SDMID_targets_cont.ms (has RFI flagging applied)

With the two _targets*.ms files already spawned as part of the restore workflow, it's unclear if we should be running the template_hifv_deliver_ms.xml since that will essentially run the hif(v)_mstransform task again and will assume that corrected = regcal_contline_all when in this case the calibrated data will be in data.


  • No labels