Montage Workflow - CWL Implementation

This repository contains a Common Workflow Language (CWL) implementation of the Montage Workflow, initially implemented for the Pegasus v5.0 workflow management system.

Requirements

Before the workflow can be executed, the environment must meet the following requirements:

• CWL Runner: A CWL runner is required to execute the workflow. The reference implementation, cwltool, can be installed via pip:

python3 -m venv venv
source venv/bin/activate
pip install cwlref-runner

• Docker: Docker is required, as several steps rely on container images containing essential tools (e.g., AstroPy).

• Montage Tools: The Montage toolkit must be installed, and the binaries must be available in the system path. Download instructions are available on the official website.

Workflow Overview

The repository contains two primary workflows:

• cwl/main_get_dataset.cwl: Downloads and prepares the necessary dataset.
• cwl/main_montage.cwl: The core Montage analysis workflow.

Usage

Dataset Generation

First, the main_get_dataset.cwl workflow is executed. This process generates the required data and creates the config_montage.json file, which is necessary for running the main analysis.

mkdir data
cwl-runner --outdir data cwl/main_get_dataset.cwl cwl/config_get_dataset.json

The dataset can be customized by modifying the config_get_dataset.json file. The current configuration is based on the example provided in the Pegasus Montage workflow implementation.

Montage Workflow Execution

Once the data is generated and dependencies are installed, the main workflow can be launched using the generated configuration file:

cwl-runner cwl/main_montage.cwl data/config_montage.json