Checklist ​

This page is for people new to the Sprocket project who want to understand how to adopt it on their own project. Work through the items below to get your development environment, CI pipeline, and workflow execution set up.

Development environment ​

• Learn WDL (if needed). If you're new to the Workflow Description Language, the OpenWDL Getting started tutorial is a fast-paced introduction to the language. For an in-depth course on building, testing, and releasing production pipelines, see the Production guide.

• Understand how a run works. Skim the execution model to see what Sprocket evaluates locally and what it dispatches to a backend.

• Install Sprocket. Get the sprocket CLI on your machine via Homebrew, direct download, or from source. See the Installation page.

• Set up your editor. Install the VSCode extension or the Neovim plugin for inline diagnostics and syntax highlighting. See the VSCode guide for details.

• Generate shell completions. Enable tab completion for commands and arguments in your shell. See Shell completions.

• Autocomplete your configuration. Run sprocket config schema and point your editor at the result to get completion and inline documentation for sprocket.toml. See the config command and the configuration reference.

Code quality ​

• Lint your WDL documents. Run sprocket lint to catch validation errors and style issues early. See the Guided Tour for a walkthrough.

• Decide which rules you want. Browse the lint rules reference (or run sprocket explain) and record the exceptions your project needs in sprocket.toml. See check/lint configuration.

• Format your WDL documents. Run sprocket format overwrite to maintain consistent style across your project. See the format command reference.

• Baseline existing diagnostics (optional). If your codebase has pre-existing lint or validation diagnostics that you cannot address all at once, run sprocket lint --generate-baseline to capture them in a sprocket-baseline.toml file. Subsequent runs of sprocket check will ignore baselined diagnostics, so CI catches new issues while the existing backlog is cleaned up on your own schedule. See Baselines for details.

• Add the GitHub Action. Add the Sprocket GitHub Action to your CI pipeline so that linting and formatting are checked on every pull request.

• Validate inputs in CI. Run sprocket validate against the input files you ship so that missing, mistyped, or missing-on-disk inputs fail fast instead of failing part-way through a run.

Running workflows ​

• Run your first workflow. Execute a task or workflow with sprocket run. The Guided Tour walks through a complete example.

• Learn how inputs are resolved. Targets, @ input files, and array inputs are covered on the Inputs and targets page; sprocket inputs can generate a template to start from.

• Pin your container images. Replace mutable tags with digests so runs stay reproducible. See Containers.

• Configure an execution backend. If you are running on an HPC cluster or a TES endpoint, configure a backend in your sprocket.toml. See the Execution Backends overview.

• Turn on the call cache. Set cache = "on" under [run.task] so that iterating on a workflow reuses the tasks that already succeeded. See Call Caching.

• Know where your results land. Runs, logs, task attempts, and the provenance database live under the output directory; --index-on gives results a stable path. See Provenance Tracking.

Community ​

• Join the conversation. The #sprocket channel on the OpenWDL Slack is the best place for questions, feedback, and discussion. See Community and Support for issue trackers, discussions, and how to contribute.