interp_images in util_misc.py returning NaNs about rm-tools HOT 4 CLOSED

Hi Emma!

I think you're the first person to try out that particular feature in RM-Tools, so thanks for the bug report! Any chance you could send me the cubes you're using? That'll make it easier to track down why the interpolation is messing up.

Using the continuum-averaged I... might be possible. I'd have to think about it. I think it might not play well with normalizing the Q/U spectra.

Cheers,
Cameron

from rm-tools.

Thanks — will email you them now.

from rm-tools.

Hi Emma:

I found the problem. Looks like it was trying to do quadratic interpolation instead of linear (which failed because it was only getting the two nearest Stokes I planes). I've fixed it to do linear interpolation. There's a new version on Github and on pip.

You might also consider playing around with the do_fitIcube.py script inside RMtools_3D/. In theory it's supposed to fit and output a pixel-by-pixel Stokes I model to the data, which can be more reliable than using channel values for Stokes I if the S:N is not really high. I don't know if it's been tested though, so you're welcome to be a guinea pig if you have time. :) If it's broken, I was planning on looking at the 3D toolkit again sometime in September/October-ish in order to add some features needed for the POSSUM pipeline.

Cheers,
Cameron

from rm-tools.

Thanks Cameron!

from rm-tools.