Differences between revisions 35 and 36
Revision 35 as of 2014-03-11 09:29:49
Size: 4785
Editor: obelix
Comment:
Revision 36 as of 2014-03-13 10:34:25
Size: 4787
Editor: obelix
Comment:
Deletions are marked like this. Additions are marked like this.
Line 15: Line 15:
  * Occasional crashes in YJ-band.   * Occasionally crashes in YJ-band.

SPARK Known bugs

Check as well the Change Log to see which bugs have already been fixed, but not released yet.
Here we list some of the most unconvenient bugs. The technically more interested user can check the SPARK Bug Tracking (only from MPE) for more detailed information.

v1.3.0

  • kmo_wave_cal

    • Products are missing the filter extension (lcal.fits instead of lcal_YJYJYJ.fits and det_img_wave.fits instead of det_img_wave_YJYJYJ.fits). easySPARK-scripts needing an LCAL-frame will fail. Renaming LCAL solves the issue.
  • kmo_reconstruct, kmo_sci_red, kmo_multi_reconstruct

    • Misleading recipe output regarding xcal-interpolation: When the angle of the science frame to reconstruct lies between 30-60deg and 300-330deg the interpolation is executed correctly, but the recipe output would state that it would NOT have been executed.
      Example: angle of 40deg gives following output
      "Loading all calibration frames with angle 60deg (input: 40deg)"
      instead of:
      "Loaded calibration frames with interpolated angles between 0deg and 60deg (input: 40deg)"

  • kmo_sci_red, kmo_sky_tweak

    • Occasionally crashes in YJ-band.

v1.2.8

  • kmo_sci_red, kmo_sky_tweak

    • Background fitting is weak in some bands (specially in YJ). Background fitting has to be improved and stabilized.
  • kmo_sci_red

    • Providing just one SCIENCE frame and an obj_sky_table indicating to subtract skies from another IFU won't work now (sky isn't subtracted, no notification in the log). Workaround: provide a 2nd SCIENCE frame in the sof-file and update the provided obj_sky_table accordingly (add --no_combione parameter as well).
  • kmo_sky_tweak

    • Recipe fails when inactive IFUs are present.
  • kmo_std_star

    • Sometimes it can happen that in the STAR_SPEC product ranges of NaN values in the middle of the spectrum appear. Furthermore it can happen that the TELLURIC shows some extreme spikes at these locations.
  • kmo_make_image

    • Take into account range-parameter even if no OH-line spectrum is provided. Workaround: supply an OH-line spectrum filled with one everywhere.
    • Fix undescriptive error message when CRVAL1 keyword in OH-line spectrum is an integer instead of a double value.

v1.2.5

On MacOS:

Some users reported following warning in kmo_wave_cal even when the resulting XCAL, YCAL, LCAL have the same number of rotator angles with matching values:

  • [WARNING] : **********************************************************************
    [WARNING] : **********************************************************************
    [WARNING] : *** For XCAL another angle has been picked than for LCAL ***
    [WARNING] : *** (XCAL: 2.68156e+154, LCAL: 6.92925e-310) ***
    [WARNING] : *** ***
    [WARNING] : *** The recipe will be executed, but the ***
    [WARNING] : *** results should be mistrusted !!! ***
    [WARNING] : *** ***
    [WARNING] : *** Please take care to take XCAL, YCAL and LCAL frame ***
    [WARNING] : *** from the same calibration set !!! ***
    [WARNING] : **********************************************************************
    [WARNING] : **********************************************************************

One value is very large and the other very, very close to zero.
Unfortunately this effect isn't reproducible and hence not really debugable and has to stay open for the moment. The calibrations don't seem to be corrupted and can be used anyway. To be more precise: This warning is emitted after having calculated the LCAL frame and during following reconstruction of the ARC_ON frame in order to create the reconstructed detector image, which in turn is used to collect some QC parameters.

v1.2.4

  • SPARKplug with MacOS X

    • Needs as prerequisite the port p5-tk to run, but it fails to compile. The bug has been reported but not fixed yet.

v1.1.0

  • SPARKplug

    • some static calibration frames in the resulting sof-files are tagged wrong

SPARK / SPRED

MacOSX-Users with a working installation of the SPRED pipeline for SINFONI can experience difficulties using it after having installed SPARK. This happens only if they install packages needed for the SPARKplug tool which uses as well python. In this case separate working environemnts are recommended.

KMOS-SPARKwiki: Known Bugs (last edited 2014-05-05 10:20:43 by obelix)