iraf.net

I routinely use imstat to get statistics on images. One of the more useful aspects of imstat is the ability to do the sigma clipping and remove deviant pixels. I use it a lot to look at detector characterization data and get the mean/standard deviation of the main set of pixels with the hot/bad pixels removed. However, I am now forced to re-create the imstat results in Python since it has been decreed that we must use Python. I have tried a few Python routines that are supposed to do sigma-clipping of images, but they give much different results than imstat does (and I think they are simply wrong given the way histograms of the pixel values look). I wish to duplicate imstat exactly, or at least very closely, because I have a long history of use of IRAF and do not want to have to adapt to another algorithm. And, I want to keep using IRAF myself whether or not the powers that be prefer Python.

Is there some documentation somewhere of the exact algorithms used in imstat? Or can I get the source code to try to adapt it to Python to get the same output?

The values computed by IMSTAT are described in the task help page, although it's probably been a while since we checked that the code hasn't been updated slightly since these were first written. The source itself is in the file iraf$pkg/images/imutil/src/t_imstat.x and should be fairly straight forward to translate but be sure to note which values in the IRAF code explicitly use double precision.