toasty tile-healpix

The tile-healpix command transforms all-sky FITS images in the HEALPix format into WWT’s FITS TOAST format.


toasty tile-healpix
   [--outdir DIR]
   [--parallelism FACTOR] [-j FACTOR]

The FITS-PATH argument gives the filename of the input image, which should be a single FITS file storing data in the HEALPix format.

The TOAST-DEPTH argument specifies the resolution level of the TOAST pixelization that will be generated. A depth of 0 means that the input will be sampled onto a single 256×256 tile, a depth of 1 means that the input will be sampled onto four tiles for a total resolution of 512×512, and so on. When converting from HEALPix, the appropriate TOAST depth can be derived from the HEALPix N_side parameter as approximately:

depth = log2(N_side) - 6.2

One should typically round up to the next larger integer to err on the side of higher resolution, although the best policy will depend on the sampling of the input HEALPix map.

The --outdir DIR option specifies where the output data should be written. If unspecified, the data root will be the current directory.

The --galactic option forces the tiling to assume that the input map is in Galactic coordinates, even if the FITS headers do not specify that. This has been observed in some data sets in the wild.

The --parallelism FACTOR argument (or -j FACTOR) specifies the level of parallism to use. On operating systems that support parallel processing, the default is to use all CPUs. To disable parallel processing, explicitly specify a factor of 1.


This command requires that the healpy Python module is installed.

This command will create FITS files for the highest-resolution tile layer, corresponding to the DEPTH argument, and emit an index_rel.wtml file containing projection information and template metadata.

The WWT rendering engine can also display datasets in the hierarchical HiPS FITS format, which addresses about the same use case as TOAST. If a HiPS FITS dataset already exists, it is probably a wiser choice to use it rather than creating a new TOAST pyramid. However, TOAST provides noticeably better performance for the initial data display and avoiding positional shifts as the user zooms in.

Currently, parallel processing is only supported on the Linux operating system, because fork()-based multiprocessing is required. MacOS should support this, but there is currently (as of Python 3.8) a bug preventing that.

See Also