Known Issues

nifsPipeline.py

nifsSort.py

Object and Sky frame differentiation

If the sorting script does not create a skylist in the object or telluric observation directories this means that the offsets between sky frames and object frames were smaller than expected. A skyframelist can be manually created and saved in the appropriate directory, or the limit placed on the offset can be changed. To lower the limit manually do a search in nifsSort.py for “skyframelist.append(entry)” and change the <= <value_in_arcseconds> to something new.

nifsBaselineCalibration.py

  • Long data file path names are not fed to IRAF tasks. It seems IRAF task parameters

must be 99 characters or less. Nifty’s data files are stored in the Astroconda environment packages directory; for example, on my system it is “/Users/nat/miniconda2/envs/niftypip/lib/python2.7/site-packages/nifty/pipeline/”. If you have a long username, for example, this can cause the path name to be too long to be parsed by iraf. Temporary fix: I have made the names of all the data files short enough for it to work okay on my system. Please let me know if this seems to be causing you issues and I can come up with a better fix.

nifsReduce.py

  • z-band data is not capable of a flux calibration (yet!).
  • Seems to be missing the first peak of the ronchi slit when iraf.nfsdist() is run interactively. This does not seem to be a problem.
  • We noticed that iraf.nfsdist() produced different results when run interactively and non-interactively. To test check the log for iraf.nfsdist() output and make sure it is identifying 8/8 or 9/9 peaks every time.
  • iraf.nftelluric() was not built to be run automatically. A lightly tested modified version that allows an automatic telluric correction is included in the extras directory but more testing is needed. For now applyTelluricPython() is recommended.

nifsMerge.py

nifsUtils.py

General Issues

  • A longstanding bug (see `astropy<https://github.com/astropy/astropy/pull/960>`_ ) in astropy has made it difficult to build Nifty as a binary executable.
  • The conversion of print statements to logging.info() statements was messy. Some of these may still not be properly converted and will throw nasty tracebacks. However these seem to have no effect on the functionioning of the code.
  • Logging is still not perfect. One or two iraf tasks are sending their log files to “nifs.log” instead of “Nifty.log”.