irsfsoftware / Blog: Recent posts

pyIRSF-2.5 released. Fixed an issue that the pipeline crashes when no star is detected in the first frame of the dithering set.

pyIRSF-2.4 uses astropy.io.fits instead of pyfits, which is obsolete.
you need astropy module to use this version.

You can stop sirius.py after sky-bias frames are created (and before sky subtraction) with 'stop=1.5' in sirius.param. Use this when you want to modify sky-bias frames. 'stop=1.5' will force 'keeps=1'

I have not examined if the pipeline is suitable to reduce data with the ND/NB filters.
Especially, if it creates the flat accurately is not clear at all.

• Fixed sirpol.py, sirius.py and dorecom.py to make the IRAF cache directory in each working directory to avoid errors when multiple pipeline processes run simultaneously.

• Some frames have large offsets, more than 3 pixels, even within a cycle of SIRPOL waveplate rotation. In that case, the pipeline failed in making the Q, U, V, I frames for each cycle. This problem was fixed in this version.

Some major changes for sirpol.py have been made.

• Removal of the vertical stripe in the background of the Q, U, and V images
  by setting 'polstripe' = 1 in sirpol.param. The default value is 0.
  In most of the cases, most of the background of the Q, U, and V images are empty, zero
  and flat. In the algorithm, the median of each column of the image is calculated
  and the median value is subtracted from the column. ... read more

• the extended image of counting the number of combined frames had -32 bit. it was changed to 16 bit.
• There was a minor bug in the final file cleaning process. Not found error occured in some cases. It was fixed.
• A parameter 'isigma' was added to recom.param, which is lsigma and hsigma in IRAF's imcombine and affetcs the final combining process of registered images.

• Sigma clipping using M.A.D.(median absolute deviation) is adopted for the combining process of registered images.
• You can select either the conventional IRAF's imcombine or the new combine program with the M.A.D. clipping.
• polmode is automatically selected when you execute sirpol.py set.
• A parameter 'geotran.interpolation' has been changed to 'drizzle' in registration.dotran()
  • It was 'drizzle[0.5]', which could yeild a poorer spatial accuracy < 0.25 pixels at most.
• Two new scripts have been added.
  • twflatview2pdf.py : This script writes the output of twflatview.py to a PDF file with 12 outputs per page.
  • subregionview.py : A selected square region of a combined image for all precombined image is written to a PDF file.

Making badpixel masks failed when only sky-bias frames with low background levels,
~ a few x 10 ADUs, were available. This has been fixed.

• irsf.normalize_and_settozero_median_sky(para.band, skyarr) was replaced
  with sky4mask = irsf.measure_settozero_mediansky(para.band, skyarr)
  in sirius.py and sirpol.py

• irsf.measure_settozero_mediansky() was created.

• irsf.mkbadmask2() was modified. ... read more

The main change is that sigma clipping using M.A.D.(median absolute deviation) is adopted for median sky bias frame construction. You can choose the conventional IRAF's sigclip+mclip median combine as well in sirius.param/sirpol.param.

pyIRSFmanual160212J.pdf and pyIRSFmanual160212E.pdf are also released.

I put sirwcs2.py at Files.
Please replace it with the previous one.
The magnitude in [j|h|k]sirphotwcs.txt can be shifted systematically, when it was non-photometric.

sirwcs2.py had a manor bug that
the magnitude in [j|h|k]sirphotwcs.txt can be shifted when it was non-photometric.
it has been fixed.

a bug fixed for mktwflist.py. it failed when all the frames for an evening/morning set have a higher backgroud level than 'maxgood' .

'sirius.py/sirpol.py comset' had a bug and it is fixed.

The masking process was improved.
There was a redundant process in the pipeline.

some bugs have been fixed
- sirpol.py failed when lines for the sky bias were removed from obslog
- some tasks failed because of the overwritng error of IRAF. The clobber parameter of IRAF is set to 'yes' in the pipeline to avoid the failure.

minor bugs in pyIRSF-2.0 have been fixed
- 'or' was used incorrectly in irsf.checkpol and irsf.polread
- when iraf.clobber is 'no', an overwrite failure happened at registration.finexyxymatch.

Th English manual will be released in a few weeks.

pyIRSF is the latest data reduction package for IRSF.
the software is distributed through this sourceforge site.