Making a Workflow Testing RO-Crate

In order to have your workflow monitored by LifeMonitor, you need to package it as a Workflow Testing RO-Crate (WTROC). If you have a Galaxy, Snakemake or Nextflow workflow that follows community best practices, the easiest way to generate a WTROC for it is to use repo2rocrate. Example:

git clone
repo2rocrate -r rnaseq/ -o

See repo2rocrate’s documentation at the above link for more information.

Using ro-crate-py

If your workflow type is not supported by repo2rocrate, or you can’t conform to community best practices for some reason, you can generate a WTROC with ro-crate-py. Set up a Python virtual environment and install ro-crate-py:

python3 -m venv venv
source venv/bin/activate
pip install --upgrade pip
pip install rocrate

Now move into the directory that contains your workflow and associated assets. This can be, for instance, a copy of a repository in a version control provider such as GitHub. As an example, we’re going to use the FAIR CRCC - send data workflow.

git clone
cd fair-crcc-send-data/

Initialize the crate, ignoring the .git and .github directories:

rocrate init --exclude .git,.github

The above command creates an ro-crate-metadata.json file at the top level that simply lists all files as "File" and directories as "Dataset". This includes the main workflow file:

    "@id": "workflow/Snakefile",
    "@type": "File"

To register it as a computational workflow, run the following command:

rocrate add workflow -l snakemake workflow/Snakefile

If you check the JSON file now, you’ll see that the entry for workflow/Snakefile is more articulate: for instance, it has a more specific @type and it links to an entity representing the Snakemake language. Other changes have also been made to the crate so that it conforms to the Workflow RO-Crate spec. If your workflow is written in a different language, specify it using the -l option as shown above. To get a list of supported options, run:

rocrate add workflow --help

Now we need to point the crate to a test instance for the workflow. LifeMonitor supports monitoring test executions that run on Travis CI, Jenkins and GitHub Actions. In this case, the (scientific) workflow is tested by a GitHub Actions workflow defined by .github/workflows/main.yml.

First, create a test suite:

rocrate add test-suite -i suite_1

Then, add a test instance that points to the CI workflow:

rocrate add test-instance suite_1 -s github \
  -r repos/crs4/fair-crcc-send-data/actions/workflows/main.yml \
  -i first_test

Where is the service URL, while the argument of the -r option is a reference to the CI workflow in the form:

repos/<OWNER>/<REPO NAME>/actions/workflows/<YAML FILE NAME>

That’s it! If you want, you can check ro-crate-metadata.json again to see how it changed in response to the above commands. All that’s left to do is to zip the crate in the format accepted by WorkflowHub:

rocrate write-zip /tmp/