Foundry
Home Schema

Galaxy workflow summary

JSON Schema for the structured summary emitted by the summarize-galaxy-workflow Mold.

draft schemasource/galaxy
Revised
2026-07-01
Rev
1
Related
schema summary-galaxy-workflow @galaxy-foundry/foundry upstream ↗
No schema registered for summary-galaxy-workflow. Add an entry to site/src/lib/schema-registry.ts.

This page points to the JSON Schema authored in this repo and shipped as part of @galaxy-foundry/foundry (orphan schema — no TypeScript producer Mold owns it; the producer is the LLM-run summarize-galaxy-workflow Mold). The schema is intentionally lighter than summary-nextflow because gxformat2 already carries typed workflow structure, mirroring summary-cwl on the Galaxy-as-source side.

Source-of-truth chain:

  1. packages/foundry/src/schemas/summary-galaxy-workflow/summary-galaxy-workflow.schema.json — canonical JSON, hand-edited as part of the Mold/cast loop.
  2. packages/foundry/scripts/sync-schema.mjs regenerates the typed summary-galaxy-workflow.schema.generated.ts const wrapper at prebuild.
  3. Published as part of @galaxy-foundry/foundry, exporting summaryGalaxyWorkflowSchema and the foundry validate-summary-galaxy-workflow subcommand.

Generated skills should validate emitted summaries with:

foundry validate-summary-galaxy-workflow summary-galaxy-workflow.json

Why This Is Lighter Than Nextflow

Nextflow summarization infers process graphs, channel shapes, sample-sheet semantics, and containers from DSL2 plus ecosystem conventions. A Galaxy workflow is already an explicit graph: workflow inputs, workflow outputs, per-step tool_id / tool_version / tool_state, and named input connections are all first-class in gxformat2. The summary therefore records validated, normalized structure and adds only the annotations the update pipeline’s interview and change-set steps need.

Top-Level Shape

The v1 summary contains:

  • source — ecosystem: galaxy, workflow name, slug, format (always gxformat2), original_format (gxformat2 or ga), release, license, annotation, URL/pin.
  • documents — entrypoint, optional converted_path (populated when a legacy .ga was converted via gxwf convert), and validation diagnostics.
  • workflow_inputs / workflow_outputs — ids/labels, input class (data / collection / parameter), collection type, optionality, defaults, formats; outputs record the promoted step/output source.
  • steps — step class, tool_id, tool_version, pinned tool_shed_repository, verbatim tool_state, named input connections with their sources, declared outputs, and when guards.
  • graph — source-to-sink edges with via markers for shape-affecting Galaxy features (map-over, batch, collection reduction).
  • tests — existing workflow tests (sibling *-tests.yml) recorded as the regression baseline, only when discoverable.
  • warnings — unsupported extensions, lossy conversions, unresolved references.

Intentional Non-Goals

  • No target translation or editing. The change-set (interview-driven) and its application to the workflow belong to interview-to-galaxy-workflow-changeset and apply-galaxy-workflow-changeset.
  • No tool discovery or wrapper authoring. tool_id / tool_shed_repository are recorded as-found; resolution of newly introduced tools is the per-step Galaxy loop’s job.
  • No runtime execution. This Mold summarizes and validates structure; run-workflow-test owns execution.

Incoming References (5)

  • Validate Summary Galaxy Workflowrelated note— AJV gate for summarize-galaxy-workflow JSON documents.
  • changeset-to-galaxy-test-planrelated note— Carry an existing Galaxy workflow's tests forward as a regression baseline and augment them for a change-set's deltas, emitting a Galaxy test plan.
  • interview-to-galaxy-workflow-changesetrelated note— Interview a user against an existing Galaxy workflow summary and emit a reviewable, step-anchored change-set.
  • summarize-galaxy-workflowrelated note— Read an existing Galaxy gxformat2 (or .ga) workflow and emit a structured summary for interview and change-set steps.
  • summarize-galaxy-workflowschema of artifact output— Read an existing Galaxy gxformat2 (or .ga) workflow and emit a structured summary for interview and change-set steps.