ICAL
====

ICAL provides a Python-based interface that wraps the `Rapthor <https://rapthor.readthedocs.io/>`_
pipeline. Rapthor itself is an iterative self-calibration pipeline designed for direction‑dependent
(DD) calibration (using DP3) and high‑resolution imaging (using WSClean) of radio interferometric
data. It was originally developed for LOFAR HBA, and is now also being developed for SKA‑Low. It
automates calibration and imaging using a sequence of modular operations driven by CWL workflow
definitions.

The flowchart below provides an overview of the main steps in the ICAL (Rapthor) pipeline, including
calibration, prediction, and imaging stages. It takes pre-processed target visibilities (generated
by the BPP pipeline) as input and produces gain tables and facet regions (used downstream by the
CIMG pipeline) as well as images after each iteration of self-calibration.

.. mermaid::

    flowchart TB
        sky_model@{ shape: disk, label: "Initial Sky Model"} --> tessellate["Group: Define DDE calibrators"]
        input_ms@{ shape: disk, label: "Pre-processed <br> Visibilities <br> (Target)"}

        bpp(("Batch Pre-Processing<br>(BPP)")) --> input_ms

        subgraph Predict["Predict"]
            dppp["Predict model visibilities"]
            model_vis@{ shape: disk, label: "Model visibility data"}
            subtract["Subtract model from data"]
            image_vis@{ shape: disk, label: "Imaging visibility data"}
        end

        subgraph Image["Image"]
            dppp_image["Phase shift visibilities"]
            image["Image using facets"]
            facet_regions@{ shape: disk, label: "Facet Regions"}
            images@{ shape: disk, label: "Images"}
            sky_models@{ shape: disk, label: "Updated Sky Model"}
        end

        subgraph Calibrate["Calibrate"]
            ddecal_fast["Calibrate fast phases"]
            ddecal_slow_separate["Calibrate slow gains"]
            process_gains["Process gains: smooth and normalise"]
            solutions_cal@{ shape: disk, label: "Gain Table (DDE)"}
        end

        tessellate & input_ms --> ddecal_fast

        ddecal_fast --> ddecal_slow_separate
        ddecal_slow_separate --> process_gains
        process_gains --> solutions_cal

        solutions_cal --> dppp
        dppp --> model_vis
        model_vis --> subtract
        subtract --> image_vis

        image_vis --> dppp_image
        dppp_image --> image
        image --> images & facet_regions & sky_models

        sky_models --> tessellate

        facet_regions & solutions_cal --> cimg(("Continuum Imaging <br> (CIMG)"))

Useful Links
------------

- **Repository**: https://gitlab.com/ska-telescope/sdp/science-pipeline-workflows/ska-sdp-ical
- **Documentation**: https://developer.skao.int/projects/ska-sdp-ical/en/latest/
- **Quickstart**: https://developer.skao.int/projects/ska-sdp-ical/en/latest/README.html
- **Status**: In Development
- **Contact**: Team GECKO
- **Where to get help**: `#gecko-selfcal-workflow <https://skao.slack.com/archives/C056R29JTA8>`_