fitstoolz

Python libraries and command-line tools for interfacing with FITS data in an intuitive way.

FITS numbers its axes from one and writes them fastest-varying-first; numpy does neither. fitstoolz reads the WCS once, gives every axis a name, and lets you use that name everywhere — on the data, on the coordinates, and on the command line — so you never have to work out which convention applies where.

Note

Early beta. The interfaces documented here are real and tested, but they may still change between 0.x releases.

Features

  • Image and coordinate data indexing is consistent

  • Simple API for adding, transposing and expanding axes (or dimensions)

  • Celestial, spectral and Stokes axes evaluated through the WCS rather than approximated by stepping CDELT

  • Xarray support

  • Zarr support

  • A fitstoolz command line whose subcommands double as stimela-ninja pipeline steps

In[1]: from fitstoolz.reader import FitsData

In[2]: myfits = FitsData("example-image.fits")
       myfits.coord_names

Out[2]: ['STOKES', 'FREQ', 'DEC', 'RA']

In[3]: myfits.dshape
Out[3]: (1, 504, 100, 100) # these dimensions match the labels above

In[4]: myfits.coords

Out[4]:
Coordinates:
STOKES   (stokes) int32 4B dask.array<chunksize=(1,), meta=np.ndarray>
FREQ     (spectral) float64 4kB 8.803e+08 8.804e+08 ... 9.328e+08 9.329e+08
RA       (celestial.ra) float64 800B 53.16 53.16 53.16 ... 53.1 53.1 53.1
DEC      (celestial.dec) float64 800B -28.16 -28.16 ... -28.11 -28.11

$ fitstoolz stats image.fits --show

Getting started

Using fitstoolz

Reference

Project

Indices