Quickstart¶
This quickstart uses the runnable bioinformatics example in examples/bioinfo.
1. Move Into The Example¶
cd examples/bioinfo
The example includes:
a
ginkgo.tomlconfig filesmall FASTQ inputs
a workflow package under
workflow/a task-specific Pixi environment under
workflow/envs/bioinfo_tools/
2. Inspect The Workflow Entry Point¶
The canonical flow lives in examples/bioinfo/workflow/flow.py.
At a high level it:
filters each FASTQ in a Pixi environment
computes per-sample QC tables in the same environment
counts reads in a container-backed shell task
merges those outputs in a local Python task
3. Preview With A Dry Run¶
Before running for real, confirm the workflow is wired correctly:
ginkgo run --dry-run
This validates the workflow without executing any task body. Ginkgo resolves the expression tree, checks environments and secrets, computes cache keys for every task, and reports which tasks would run, which would serve from cache, and which resources they declare.
A project may also keep validation workflows under tests/workflows/. Those are
workflows too, so they run by path: ginkgo run tests/workflows/smoke.py, with
--dry-run to validate instead of executing. See
Validation workflows.
4. Run The Workflow¶
ginkgo run
When you run from the example root, Ginkgo auto-discovers the canonical
workflow/flow.py entrypoint.
If you prefer to be explicit:
ginkgo run workflow/flow.py
5. Inspect Outputs¶
After a successful run, look at:
results/filtered/for filtered FASTQsresults/qc/for per-sample statsresults/read_counts/for container-generated read-count tablesresults/summary.csvfor the merged report.ginkgo/runs/<run_id>/for run provenance and logs
6. Re-Run To See Cache Reuse¶
Run the workflow again without changing inputs:
ginkgo run
Ginkgo should reuse cached results where the cache key still matches. The cache identity includes task source, resolved inputs, and execution-environment identity where relevant.