Matthew Wood

I don't think having element-wise update is critical for us. The use-case would generally be reading/updating a very large number of elements at once. Probably for updating the easiest approach...

Yes this is expected behavior. An upsampled/downsampled map does not preserve the condition on which the DISK geometry is defined (all pixels with centers within X radius of some point...

I think there are two issues here: * We neglected to add the EXPLICIT region geometry to the GADF healpix format. This had been on my todo list but I...

I've uploaded a sample file that illustrates the current serialization format used by fermipy: https://www.dropbox.com/s/oh3lz5o3g1rj8k4/roi.fits?dl=0 Currently there's two parts to the model serialization. There are a set of vector columns...

Sounds reasonable to me although I would go for a tuple rather than a dict argument. With respect to FITS I/O it would be nice if this would support the...

Ah I see you intended that each element would map to a column. I would still prefer something that allows tuple-like access (i.e. by index) so I guess named tuple...

In gamma-ray applications we often have columns which correspond to image planes of a higher dimensional data structure -- for instance a cube where the third axis is energy. In...

I think it would be a great idea to have some HEALPix visualization methods built into `astropy`/`astropy-healpix`. `PolyCollection` is the most flexible approach but doesn't really scale well when you...

A quick addendum... the same issue also arises for all-sky CAR to any WCS projection. Thinking about it a bit more I realize that my proposal to use `mode='wrap'` will...

Sure. Below is a script that reproduces the issue for both WCS and HPX reprojection. In both examples I'm starting from an all-sky CAR map where I've set the pixel...