TART telescope tools

Containerized tools for using the TART radio telescope. This repository creates a TART tools container. It is recommended that this container be used using apptainer (or singularity) which requires Linux or recent versions of Windows. They can also be used with docker on all platforms (but some filesystem mounting is required).

The tools available are:

• tart2ms: Create and manipulate measurement sets from TART data
• disko: Discrete Sky Operator imaging. Ideal for working with wide-field TART data.
• disko_draw: Draw and visualize discrete fields of view
• spotless: Model fitting imaging direct from visibilities. Create a sky model direct from visibilities.
• All of the tart_tools python package.
• • tart_download_data
• • tart_upload_antenna_positions etc.

Build The Tools

Currently the tools require apptainer/singularity to be installed, do this on your system, and then

make build

Now each of the tools can be run by prefixing the command with apptainer run tart_tools.sif ie:

apptainer run tart_tools.sif tart_download_data --help

Create an alias called tart_tools

alias tart_tools='apptainer run /home/tim/tart-telescope/tools/tart_tools.sif'

The tart_tools alias allows each tool to be run quite simply, just prefix each command with tart_tools. ie.

tart_tools tart_download_data --help
tart_tools tart2ms --help

Imaging using the TART telescope tools

The tart_tools alias allows each tool to be run quite simply. See the following example:

First download some data

tart_tools tart_download_data --vis --api https://api.elec.ac.nz/tart/mu-udm --n 1

Then convert that data to a measurement set (tart2ms)

tart_tools tart2ms --hdf vis_2025-03-05_13_04_41.765325.hdf --single-field --rephase obs-midpoint --ms udm.ms

Then image the data using l1 regularization (disko)

tart_tools disko --ms udm.ms --nvis 10000 --healpix --fov 170deg --res 30arcmin  --lasso --alpha 0.005 --HDF tart_udm_image.h5

Finally draw the known sources on the image and write it out as an SVG:

tart_tools disko_draw --show-sources --SVG tart_udm_image.svg tart_udm_image.h5

Then convert the image to a JPG.

convert -resize 25% tart_udm_image.svg img/tart_udm_image.jpg

Using Docker

The docker image for TART tools is public (the apptainer image is based on it). You can run it directly in docker

docker run -it ghcr.io/tart-telescope/tools tart2ms --help 

to run this on local files, you will need to mount the necessary directories.

ToDo

• Stimela2 definitions for commands

Changelog

• March 2025. Initial commit.
• 8 March. Move to public container on ghcr.io
• 1.1.3. Add tart-gnss-acquire (github.com/tmolteno/tart-gnss).
• 1.1.4. Update to fix a bug in tart-gnss-acquire with FFT buffer allocation on Beidou
• 1.1.7. Update to latest tart2ms