Welcome to the WDL Runner repository.

This repo contains utilities that facilitate running WDL workflows through the Broad's Cromwell execution engine on the Google Cloud Platform (GCP).

If you are a user and would like some more user-facing documentation, see readthedocs.

The repo has directories containing the following contents.

These are the source pages which feed our readthedocs pages.

Source code for the wdl_runner tool. This is a convenient tool which handles:

• Spinning up a Cromwell server on a GCP VM
• Launching a WDL workflow by submitting it to the newly created Cromwell server.

Additional monitoring tools are stored in the 'monitoring_tools' directory of the wdl-runner repository.

These are scripts that facilitate monitoring the status of WDL workflows run on GCP. The main monitoring script is monitor_wdl_pipeline.sh. It accepts an operation ID for a pipeline, extracts the LOGGING, WORKSPACE, and OUTPUT directories from the operation and then examines these directories to glean some insights into the status of the operation.

Note that if the WORKSPACE and/or OUTPUT directories for the specified operation are already populated (for example by another operation), this script will emit incorrect output.

To run the tests:

# one-time setup only
conda env create --file environment.yml
conda activate wdl-runner
PYTHONPATH=.:./wdl_runner pytest