v1.0.1

What's Changed

• Also mention the notebook tools in terra-notebooks-playground. by @deflaux in #44
• Switch to only manual dispatch for 'Publish to Terra' GitHub action by @deflaux in #45
• Add more workflow metadata and documentation for notebook_workflow.wdl. Also export values for GOOGLE_PROJECT and PET_SA_EMAIL when running on Google Cloud. By @deflaux in #46

Full Changelog: v1.0.0...v1.0.1

terra-examples v1.0.0

Includes the initial release of a Cromwell workflow that executes a Jupyter notebook from start to finish. It also includes several collections of demonstration Jupyter notebooks that use public data as input. A tutorial and cheatsheet for using version control on app.terra.bio and the All of Us Researcher workbench are also provided.

The notebook_workflow.wdl workflow will:

• Optionally install a list of Python packages before executing the notebook.
• Optionally pass parameters to the notebook via Papermill to change its behavior. See also https://papermill.readthedocs.io/.
• Execute the notebook from start to finish.
• Save the executed ipynb file (containing cell outputs) as a result of the workflow.
• Also save an HTML copy of the executed ipynb file as a result of the workflow. This allows the notebook outputs to be read in the cloud console immediately.
• If the notebook created any output files or directories underneath the current working directory, they will also be included in a tar output file.

The notebook is executed on a new, clean VM (as opposed to where you normally run notebooks interactively). This is useful not only for reproducibility and provenance, but to specifically confirm that the notebook does not depend on any local dependencies (e.g., files or Python/R packages) installed where you normally use Jupyter interactively.

NOTE: If an error occurs during notebook execution, the resulting ipynb and html files are still saved, but you will need to go look for them in the execution directory of the workflow run.