hello-world-workflow

A tiny test CWL workflow written to help develop new features for Toil. Also not a bad toy example for introducing yourself to CWL workflows.

Running the workflow with Toil

Theoretically this should work with any CWL runner that supports scattering and sub-workflows; but, assuming you want to use Toil:

1. First install Toil with the CWL extra by running

   pip install 'toil[cwl]'
   

2. Install Docker by following Docker's install docs (or otherwise)

3. Run the main workflow with

   toil-cwl-runner hello_world_scatter_workflow.cwl config.yaml
   

If you want to get fancy by using Toil's advanced features, such as autoscaling for AWS, Azure, or Google Cloud Platform, consult the official Toil documentation.

Workflow breakdown

In the config file, config.yaml, we have two lists: a list of names, and a list of numbers. The main workflow, hello_world_scatter_workflow.cwl, "scatters" (i.e., takes every combination of) the names and numbers, and runs each combination through the subworkflow hello_world_subworkflow.cwl, which runs the following steps:

hello_world

The hello_world step takes a name and number and prints a unique "hello world" message to a file.

duplicate_file

The duplicate_file step takes the file from hello_world and duplicates it into two separate files. This step is (needlessly) run in the node:slim Docker container, to incorporate the Docker feature into the workflow for testing purposes.

make_caps

The make_caps step takes the files from duplicate_file and converts all their alphabetic characters to uppercase and creates a tarball with these converted files.

By default the final output tarballs created by make_caps should appear in the base directory of the repository.