sbinet / fits Goto Github PK

Copyright 2014 Shahriar Iravanian ([email protected]). All rights reserved. Use of this source code is governed by a MIT license that can be found in the LICENSE file.

Package fits reads and processes FITS files. It is written in pure golang and is not a wrapper around another library or a direct translation of another library to golang. The main purpose is to provide a native golang solution to reading FITS file and to assess the suitability of golang for scientific and numerical applications.

FITS is a common open-source format for storage and transmission of astronomical images and data. This package is based on version 3.0 of the FITS standard.

The following features are supported in the current version:

1. Images with all six different data format (byte, int16, int32, int64, float32, and float64)
2. Text and binary tables with atomic and fixed-size array elements

The following features are not yet implemented:

1. Automatic application of BSCALE/BZERO
2. Random group structure
3. Variable length arrays in binary tables
4. World coordinate system

Also note that currently this package provides only read capability and does not write/generate a FITS file.

The basic usage of the package is by calling Open function. It accepts a reader that should provide a valid FITS file. The output is a []*fits.Unit, where Unit represents a Header/Data Unit (i.e. a header with the corresponding data). Unit provides a set of variables and functions to access the HDU data.

Let 'test.fits' be a FITS file with two HDU. The first one is of type SIMPLE and contains a single two-dimensional image with the following parameters:

BITPIX  =  -32
NAXIS   =  2
NAXIS1  =  512
NAXIS2  =  256

The second HDU contains a binary table (XTENSION=BINTABLE):

BITPIX  =  8
NAXIS   =  2
NAXIS1  =  100
NASIX2  =  5
TFIELDS =  10
TFORM1  =  E
TTYPE   =  FLUX
TDISP1  =  F10.4

To read this file, we first call

units := fits.Open("test.fits")

Now, units[0] points to the first HDU. We can access the header keys by using units.Keys map. For example, units[0].Keys["BITPIX"].(int) returns -32. Note that Keys stores interface{} and appropriate type-assertion needs to be done. Unit.Naxis returns a slice of integers ([]int) containing all NAXIS data. For example, units[0].Naxis is equal to [512, 256]. We can access the image data points by using one of the three accessor functions: Unit.At, Unit.IntAt and Unit.FloatAt. Each function accepts NAXIS integer arguments and returns the pixel value at that location. Unit.At returns an interface{} and needs to be type-asserted before use. Unit.IntAt and Unit.FloatAt return int64 and float64, respectively.

For table data, we use two other accessor functions: Field and Format. Field accepts one argument, col, that define a field. It can be 0-based int or a string. For example, units[1].Field(0) and units[1].Field("FLUX") both points to the same column. The return value of Field is another function, which is the actual accessor function and accepts one int argument representing a row. For example, units[1].Field("Flux")(1) returns the value of column "FLUX" in the second row of the table as interface{}. The following code populates a slice of float with the value of the FLUX column (note that Naxis[0] is the number of bytes in a row and Naxis[1] is the number of rows):

fn := units[1].Field("FLUX")
x := make([]float32, units[1].Naxis[1])     
for row := range x {
    x[row] = fn(row).(float32)
}

Format function on the hand accepts two arguments, col (same as Field) and row and return a string formatted according to TDISP for the field. For example, if units[1].Field("Flux")(1) is equal to 987.654321, then units[1].Format("Flux", 1) returns "987.6543".