Pipeline Diagram API Reference

The kreview.pipeline_diagram module declares the Nextflow DAG once — nodes, the processes each stands for, and the routing between them — and emits it as static SVG at build time. Layouts are computed here rather than in the browser, so the geometry that ships is the geometry that was verified, and the same declaration produces both the report's embedded diagrams and the standalone pages in docs/diagrams/.

tests/test_pipeline_diagram.py asserts the declaration against the workflow in both directions, so a stage added to the DAG cannot leave the diagram describing an older topology.

For conceptual explanations, see:

• Nextflow
• Dashboard Guide

---

kreview.pipeline_diagram

The Nextflow DAG as a drawn object. One declaration of nodes, their constituent processes and the routing between them; layouts are computed here at build time and emitted as static SVG, so the page never positions a node at runtime. Two densities ship in the report (overview for Methods, standard for Run diagnostics); the fuller expansion lives in docs/diagrams/.

Docs: https://msk-access.github.io/kreview/pipeline_diagram.html.md

mini_svg(cluster)

Sub-topology of one compressed node, for the drill-down panel.

pipeline_svg(level, *, chips=True)

Static SVG for one detail level.

chips=False omits the status slots entirely — the Methods copy explains the pipeline and has no run to report, so it should not carry empty slots that look like missing data.

mini_store()

All sub-topologies in one hidden block; the page clones them into the panel.

node_process_map()

{node key: [Nextflow process, ...]} — the page joins the trace on this.

declared_processes()

Every Nextflow process the diagram claims to cover.