Possible WCS bug (could be header dependent though)

[gully]

Hi there,
I found your code while looking for an existing tool to easily make postage stamp images while preserving WCS information in the header. Your imfuncs.py poststamp_radec() worked for one set of images, but when I tried it on a different image I got predicted (x,y) positions that differed from the visually confirmed ones. I traced the problem back to wcs.sky2pix(). Specifically, I compute different (x,y) values with the CodeCDF wcs and the astropy wcs.

CodeCDF

import wcs
vals = wcs.sky2pix(hdu[0].header, RA, DEC)
print vals

astropy

import astropy.units as u
from astropy.wcs import WCS

w = WCS(hdu[0].header)
x, y = w.wcs_world2pix(RA*u.degree, DEC*u.degree, 0)
print (x.tolist(), y.tolist())

I was reluctant to make an issue at first, since imfuncs worked for one image, but not another. That led me to believe that the header in my image could be malformed or at least non-standard. However, the astropy WCS manages to calculate the correct (x,y), so maybe astropy WCS has a way to catch nonstandard header information. I threw in the towel on trying to get it to work, so unfortunately I can't offer any more detail on what the root of the issue could be, and this could all be a red herring from some strange edge case. But I just wanted to give you a heads up.

[cdfassnacht]

Thanks for pointing this out. I have been meaning to rewrite and/or clean up wcs.py for a long time now, since it was thrown together rather quickly. You should certainly trust the astropy team, who are much more focused on the code than I am. That being said, I will try to track down this bug.