Hi folks,

I’ve submitted a YTEP for consideration to create a FITS data standard for yt:

https://github.com/yt-project/ytep/pull/5

yt handles FITS data in two ways:

1. It can take slices, projections, or covering grids and write them to FITS files using the FITSImageData class and its subclasses: 

https://yt-project.org/doc/visualizing/writing_fits_images.html

2. There is a FITS frontend to read in FITS images as datasets:

http://yt-project.org/doc/examining/loading_data.html#fits-data

The FITSDataset class is subclassed also to handle things like spectral cubes and 2D images with sky coordinates, but generic image files can be read in also, though without metadata it will be difficult to set things like units and the like.

The new FITS data standard would be for FITS files written by yt _only_ and would enable slices, projections, etc to be more self-describing in terms of coordinates and units. This metadata would be stored in FITS headers. Then there would be a subclass of FITSDataset, YTFITSDataset, which would look for this metadata and use it to define the dataset in yt.

Please have a look at the YTEP and let me know if you have any comments. Once the YTEP is merged, I’ll be advancing consideration on the corresponding yt PR:

https://github.com/yt-project/yt/pull/2010

Best,

John