Overview

ALMASim is a package to generate mock observations of radio sources as observed by the Atacama Large Millimetre/Submillimetre Array (ALMA). ALMASim primary goal is to allow users to generate simulated datasets on which to test deconvolution and source detection models. ALMASim is intended to leverage MPI parallel computing (Dask, Slurm, PBS) on modern HPC clusters to generate thousands of ALMA data cubes, but can also work on laptopts. ALMA database or a prepared catalogue is queried to sample observational metadata such as band, bandwidth, integration time, antenna configurations and so on. ALMASim employs the MARTINI Package (https://github.com/kyleaoman/martini), and the Illustris Python Package (https://github.com/illustristng/illustris_python).

Citing ALMASim

If you use ALMASim in your research, please cite the following paper:

@ARTICLE{10.1093/mnras/stac3314,
author = {Delli Veneri, Michele and Tychoniec, Łukasz and Guglielmetti, Fabrizia and Longo, Giuseppe and Villard, Eric},
title = "{3D Detection and Characterisation of ALMA Sources through Deep Learning}",
journal = {Monthly Notices of the Royal Astronomical Society},
year = {2022},
month = {11},
issn = {0035-8711},
doi = {10.1093/mnras/stac3314},
url = {https://doi.org/10.1093/mnras/stac3314},
note = {stac3314},
eprint = {https://academic.oup.com/mnras/advance-article-pdf/doi/10.1093/mnras/stac3314/47014718/stac3314.pdf}
}

ALMASim entry: https://doi.org/10.1093/mnras/stac3314

Installation Notes

ALMASim works with python3 (version 3.12), and does not support python2. First create a virtual environment with python3 and activate it. Then install the required packages with pip:

• Create the Python Environment:python3.12 -m venv astro-env
• Activate it: source astro-env/bin/activate (in case of your shell is Bash, otherwise check the other activations scripts within the bin folders)

Installing with pip

• pip install almasim

Installing from GitHub

• Clone the ALMASim Repository: git clone https://github.com/MicheleDelliVeneri/ALMASim.git
• Install packages from the requirements file: pip install -e .

Adding Kaggle API

• Login into Kaggle and go to: https://www.kaggle.com/settings
• Click on create new token, this will produce a kaggle.json file which must be saved in your home folder: ~/.kaggle/kaggle.json

Getting started

To run the simulation, just run:

python -c "from almasim import run; run()"

Notes

Cube size will dictate simulation speed and RAM usage. To gauge what you can affort to run, we advice to start with a single simulation of a 256 x 256 x 256 cube.

Parameters and Configuration

• The robustness parameter, often referred to as the Briggs robustness parameter or robfac in some implementations, is typically taken within a specific range that allows for a meaningful balance between sensitivity and resolution. The common range for this parameter is:

  • robust = -2 to robust = +2

Interpretation of Values

• robust = -2:
• This setting heavily favors natural weighting.
• It emphasizes sensitivity, meaning that the weighting scheme gives more emphasis to areas of the UV plane that are more densely sampled.
• The resulting image will generally have lower noise but may have lower resolution and broader synthesized beams.
• robust = 0:
• This is a balanced setting that attempts to compromise between natural and uniform weighting.
• It offers a good balance between resolution and sensitivity.
• This is often considered a default or starting point in many imaging processes.
• robust = +2:
• This setting favors uniform weighting.
• It prioritizes higher resolution by giving more uniform weight across the UV plane, even in less densely sampled areas.
• The resulting image will typically have a higher resolution and a narrower synthesized beam, but with increased noise.

Usage in Practice

• Default Values: Depending on the imaging task, astronomers often start with robust = 0 as a default and then adjust based on the specific needs (e.g., higher resolution or lower noise).
• Adjustment: The choice of robust is influenced by the science goals. For example:
• If detecting faint structures is the priority, a lower robust value (towards -2) might be chosen.
• If resolving fine details in an image is more critical, a higher robust value (towards +2) might be preferred.
• Exploration: It’s common to generate images using several different robust values to see how the balance of resolution and noise affects the final image.