buthanoid / easyfits.jl

Using FITS files made easier for Julia

Easy reading/writing of FITS files in Julia

EasyFITS is a Julia package designed to make it easier to read and write data in FITS format without sacrificing performances, flexibility, or readability.

A few examples

The full documentation is available on-line.

The Flexible Image Transport System (or FITS for short) is a file format widely used in Astronomy to store many kinds of data (images, tables, etc.) and metadata. FITS files consist in a concatenation of Header Data Units (HDUs) which each have a header part followed by a data part.

The following example demonstrates how to write a FITS file with 2 HDUs, an Image Extension and a Table Extension:

using Dates, EasyFITS
arr = rand(Float32, (3,4,5));
nrows = 20;
inds = 1:nrows;
speed = rand(Float64, nrows);
mass = rand(Float32, nrows);
position = rand(Float32, 3, nrows);
phase = (1:7) .// 3;
amplitude = exp.(-1:-1:-7);
x = amplitude.*cos.(phase);
y = amplitude.*sin.(phase);
writefits(filename,
          #-----------------------------------------------------------------
          # First HDU must be a FITS "image", but data may be empty.
          #
          # Header part as a vector of `key=>val` or `key=>(val,com)` pairs:
          ["DATE"    => (now(), "date of creation"),
           "HISTORY" => "This file has been produced by EasyFITS",
           "USER"    => ENV["USER"]],
          # Data part as an array:
          arr,
          #-----------------------------------------------------------------
          # Second HDU, here a FITS "table".
          #
          # Header part of 2nd HDU as a tuple of pairs:
          ("EXTNAME" => ("MY-EXTENSION", "Name of this extension"),
           "EXTVER"  => (1, "Version of this extension")),
          # Data part is a table in the form of a named tuple:
          (Speed    = (speed, "km/s"),  # this column has units
           Indices  = inds,             # not this one
           Mass     = (mass, "kg"),
           Position = (position, "cm")),
          #-----------------------------------------------------------------
          # Third HDU, another FITS "table".
          #
          # Header part of 3rd HDU as a named tuple (note that keywords must
          # be in uppercase letters):
          (EXTNAME = ("MY-OTHER-EXTENSION", "Name of this other extension"),
           EXTVER  = (1, "Version of this other extension"),
           COMMENT = "This is an interesting comment"),
          # Data part is a table in the form of a vector of pairs (colum names
          # can be strings or symbols but not a mixture):
          [:phase => ((180/π).*phase, "deg"),
           :amplitude => (amplitude, "V"),
           :xy => (hcat(x,y)', "V")])

Each HDU has a the header part (the metadata) and a data part which is reflected by the pairs of arguments after the name of the file filename in the above call to writefits. The headers are provided by collections (a vector for the 1st one, a tuple for the 2nd) of pairs associating a keyword with a value and a comment (both optional). The data in a FITS Image Extension is any real-valued Julia array. The data part in a FITS Table Extension is provided by a collection of column names associated with columns values and optional units. The columns in a FITS table must have the same trailing dimension (interpreted as the rows of the table) but may have different leading dimensions corresponding to the sizes of the column cells. In the above example, the "Position" column has 3 values per cell (presumably the 3D coordinates), while other columns have a single value per cell.

To read the headers of the 1st and 2nd HDU of the file:

hdr1 = read(FitsHeader, filename)
hdr2 = read(FitsHeader, filename, ext=2)

yield two instance of FitsHeader. Reading the data parts is very easy:

dat1 = readfits(filename)
dat2 = readfits(filename, ext=2)

will yield an array dat1 equal to arr and a dictionary dat2 indexed by the column names (in uppercase letters by default). For example:

dat2["SPEED"] == speed

should hold.

Installation

The easiest way to install EasyFITS is via Julia registry EmmtRegistry:

using Pkg
pkg"registry add General"
pkg"registry add https://github.com/emmt/EmmtRegistry"
pkg"add EasyFITS"

Adding the General registry (2nd line of the above example) is mandatory to have access to the official Julia packages if you never have used the package manager before.

Related projects

The FITSIO package is another alternative to read/write FITS files. EasyFITS is no longer based on FITSIO and now exploits Clang.jl to directly call the functions of the CFITSIO library and BaseFITS to parse metadata (FITS header cards).