jfoster17 / pyfits2itk

Python module to convert FITS to NRRD (for Slicer3d). Replaces fits2itk.

pyfits2itk

pyfits2itk is a Python replacement for fits2itk from the IIC/Astronomical Medicine Project, which is no longer maintained. This module uses astropy to read in a FITS cube or image, performs some simple manipulations useful for converting astronomy data (such as flipping the RA/longitude axis) and then calls the pynrrd module (included) to write out a nrrd file for inport into 3D programs such as Slicer3D.

Since Slicer3D does not understand astronomy coordinates, you must preregister and congrid all cubes externally before converting them to nrrd.

There is limited support for converting astronomy coordinates into units Slicer3D does understand. See Advanced Usage.

Dependencies

astropy

If you have astropy (which requires numpy) then you're good to go.

Optional Dependencies

scipy

Scipy is used to do the interpolation regridding required if one wants to rescale the velocity axis.

Installation

python setup.py install

Example usage

import fits2itk

	# convert a FITS file using the default parameters
infile = "ngc1333_co.fits"
outfile = "ngc1333_co.nrrd"

fits2itk.convert(infile,outfile)

	#Or rescale the data and velocity scale relative to spatial
fits2itk.convert("13co10_done.fits","output.nrrd",data_scale=10,vel_scale=2.)


	# orconvert a FITS file using parameters defined
	# in an external file

fits2itk.convert(infile,outfile,vel_scale=1,use_conv="ngc1333_conv")

	# read in the nrrd file to examine it
readdata, options = fits2itk.read(filename)
print readdata.shape
print options

Advanced usage

If all your datasubes are fairly homogeneous, you can put them all on the same scale. Create a dictionary for this transformation and store it in a file, then call convert with use_conv='name of file'. See ngc1333_conv.py for an example. Note that the defaults here are sensible for the test cube, but probably note for your dataset.

In the following there are two cubes with different spatial/spectral resolution and only partially overlapping in space. With the use_conv option the dimensions are rescaled so that they will properly register when imported into Slicer3D (other 3D software not tested). The c18o32 file has velocities in km/s, while the 13co10 file has velocities in m/s, so vel_scale is used to account for this.

import fits2itk
fits2itk.convert("ngc1333_c18o32.fits","c18o32.nrrd",vel_scale=1000.,use_conv="ngc1333_conv")
fits2itk.convert("ngc1333_13co10.fits","13co10.nrrd",vel_scale=1.,use_conv="ngc1333_conv")

License

See LICENSE.