SIMG

The Spectral Line Imaging Pipeline (SIMG) is a CLI-based application designed to process visibilities from MSv4 Zarr files, and generate science-ready Spectral FITS cubes. The MSv4 schema is defined in xradio documentation. This pipeline relies on dask for distribution of work across multiple processes or multiple HPC nodes.

Inputs and Outputs

Inputs:

Pre-processed visibilities (MSv4), YAML configuration file, continuum model images (FITS)

Outputs:

Restored spectral cubes (FITS), one per polarization.

High-Level Stages

The pipeline executes the following sequence:

Polarisation Frame Conversion:

Converts input visibilities to the expected output polarisations.

Visibility Prediction:

Predict model visibilities from input model continuum images.

Continuum Subtraction:

Performs continuum subtraction in the visibility domain.

Imaging:

Produces final spectral cubes through synthesis.

 
        flowchart LR
  cimg(("Continuum Imaging <br> Pipeline <br> (CIMG)")) --> continuum-model-image[("Continuum <br> Model Image")] --> simg
  ical(("Self-Calibration <br> (ICAL)")) --> corrected-vis
  corrected-vis[("Calibrated <br> Visibilities")] --> simg
  subgraph simg ["Spectral Line Imaging Pipeline"]
    direction LR
    pol-frame-conversion["Polarisation Frame <br> Conversion"] --> predict["Visibility Prediction"]
    predict --> contsub["Continuum <br> Subtraction"]
    contsub --> imaging["Imaging"]
    imaging
  end

  simg --> spec-cube[("Restored <br> Spectral Cube")]

Key Dependencies

The pipeline integrates high-performance external libraries for core operations:

Library

Description

ducc

Uses Wgridder for prediction and imaging.

aoflagger

(Optional) Handles visibility flagging.

radler

(Optional) Manages deconvolution operations.