CWL Runner Installation

The recommended CWL runner is cwltool, the reference implementation for the CWL standards.

Tip

• Please explore the cwltool docs for latest installation instructions.
• We also recommend to install a software container engine (e.g. Docker or Podman).

Windows

The installation on Windows can be done following the guide here.

1. Install Windows Subsystem for Linux from the Microsoft Store

   

2. Install Debian from the Microsoft Store

   

3. Set Debian as your default WSL 2 distro: wsl --set-default debian

4. Install Docker Desktop for Windows

   • Start Docker Desktop and Navigate to Settings
   • Select “Use WSL 2 based engine” in the general tab and apply
   • Select “Enable Integration with my default distro” in the resources tab under WSL Integration

5. Start WSL

6. Follow the Instructions for Linux (Debian/Ubuntu)

Linux

For installation on Linux (Debian/Ubuntu):

1. Run sudo apt-get update
2. Install Python 3 if it is not already installed sudo apt install python3
3. Install python virtual environment sudo apt install python3.[your version here]-venv
4. Create a virtual environment python3 -m venv env (named env here, name can vary)
5. Activate the virtual environment source env/bin/activate
6. Install cwltool with pip pip install cwltool

Macos

1. Install conda-forge
2. Install cwltool via conda install -c conda-forge cwltool

cwltool usage

Section titled cwltool usage

• If you are on Windows, start the WSL
• Activate the virtual environment source env/bin/activate
• Run cwltool by specifying the CWL Workflow or CommandLineTool description file path and the (optional) inputs file path (you can use relative or full paths):

cwltool path/to/cwlfile.cwl path/to/jobfile.yml

Minimal example

Section titled Minimal example

Here is a very simplified example to check, that your cwltool installation functions

1. Store the following as echo-tool.cwl

   #!/usr/bin/env cwl-runner
   
   cwlVersion: v1.2
   
   class: CommandLineTool
   
   baseCommand: [echo]
   
   stdout: message.txt
   
   inputs:
     message:
       type: string
       inputBinding:
         position: 1
   
   outputs:
     output:
       type: stdout

2. In the same folder, store the following as job.yml

   message: "I love ARCs and CWL"

3. Now you can execute the tool

   • providing an input directly via CLI:

   cwltool echo-tool.cwl --message "ARCs are great"

   or

   • providing the input via the job.yml:

   cwltool echo-tool.cwl job.yml

4. Both create an output file called message.txt with your specified message.