Skip to content

Latest commit

 

History

History
98 lines (76 loc) · 3.89 KB

TODO.md

File metadata and controls

98 lines (76 loc) · 3.89 KB
Raw

TODO's and ISSUES

Planned for version 1.2 (September 26, 2024)

  • urgent unfiled issues:

    • combos of e.g. MX-2 sometimes fail (hardware issue) [#46]
    • combos of mosaic doesn't work? they also take VERY long, where is the CPU going?
    • LineCheck sometimes fails for blanking, when driver is ok (eg. 113206)
    • Should oid= allow us to use ${obsnum}__${oid} - this would make it possible to make multiple (line) cubes
    • can we create a different cube for bank0 and bank1 ? should be possible.
    • "restart=1 bank=0" fails, but is ok with meta=0
    • why is RMS/RMS0 not closer to 1 for SEQ (it.s often 1.5--2 or worse)

  • active github issues: https://github.com/astroumd/lmtoy/issues

    • 57: spectral vs. average Tsys
    • 51: SEQ/Ps html now broken for dual IF
    • 48: allow missing roach files Kamal
    • 47: "wtn" has the wrong beam plotted
    • 46: SEQ low power can crash pipeline (MX-2 example mentioned before)
    • 45: RSR not commutative +Min
    • 38: final DataVerse items marc
    • 35: inttime not correct - we need the ON time only,not ON+OFF+CAL

  • python versions via install_anaconda3: we now use version=2023.03-1 # 3.10.14

    Also note that numpy 2.0 is upcoming and could have some impact beyond 2026.

  • on malt we should also do a "last_obsnum" trick we do on unity. malt is becoming slow again and for certain obsnums we will miss one in the last100 could also leave less data in DATA_LMT

  • final spectrum in SEQ (tab_plot.py) should honor location=; ccdspec needs the patch

  • re-install with a new python: make pip install_lmtslr install_dreampy3 install_dvpipe install_maskmoment

  • autorun mode (restart=2) [partially implemented]

  • Webrun

    • verify.py ; obsmode (INSTR/MODE)
    • webrun.sh
    • etc/parameters.txt
    • lmt_web_lite

  • FITS export option for spectra [mostly done, but not checked with dysh]

    • should there be an SDFITS option?
    • RSR works, but SEQ/Bs/Ps still in ascii [no rich header]

  • SDFITS conversion of .nc file [#53]

    • rudimentary gridder exists

  • SEQ/Bs works with dual bank now (some fixes needed for single bank) examples: 2023S1SEQUOIACommissioning VX-Sgr 108756 108762 108766 2021-S1-UM-3/100558 UGCA281 99713 .. 100558

  • SEQ/Ps almost working examples: 2024-S1-UM-3/121219 Core_187 ... 2024S1SEQUOIACommissioning/110416 MonR2 2023S1SEQUOIACommissioning/108764 VX-Sgr 2018-S1-MU-79/90692 GP_J1735 ... 2018-S1-MU-25/84891 Enceladus Titan

  • stage/unstage on unity in case we do the work in /work and rsync to /nese (or /scratch?) this has MAJOR impact on how to run pipeline

  • add the repo counter into the fits header of final output [currently only in lmtoy_repo=]

  • revive maskmoment (has an import issue) [done?]

  • resolve tsys_aver in lmtslr/spec/spec.py as per Heyer

  • RMS0: should we define some ? Currently RMS0 is defined per 100K (as is for RSR)

  • Autoscaling the rather wide HTML pages for SEQ summary

  • split parameters in single obsnum and combo (where only gridding is allowed) - useful for webrun

  • pointing offsets from ifproc in meta-data ?

  • automated generation of the script generator?

  • FITS export option for spectra [sp2sdfits.py now exists] [#53]

  • the never finished manual (flow diagrams w/ mermaid?)

Longer term wishes (>March 2024)

  • SpecFile (netCDF format) to be replaced with an SDFITS file (just only convert)
  • interoperability with dysh (e.g. sp2sdfits)
  • advanced mapping programs (cf. dysh)
  • SDHDF ?
  • docker image for lmtoy

Random

  • SLpipeline
    • SEQ needs to log its commands
    • use "INSTR/MODE" in SLpipeline
    • safe frontends SLpipeline vs. SLpipeline.sh, etc. - see verify
  • script generator for SEQ numbands=2
  • sbatch_lmtoy should learn how to use fewer cpu's [check w/ Grant]
  • create flexible summaries of observations
    • obsnum,date,bad_beam
    • efficiency of observations