nextflow-summary-to-galaxy-reference-data

SKILL.md

# nextflow-summary-to-galaxy-reference-data

Follow the procedure below and use the artifact/reference sections as the runtime contract.

## When To Use

- Decide the Galaxy-side shape of external reference data declared by a Nextflow pipeline.

## Inputs

- Read artifact `summary-nextflow`. Schema: summary-nextflow. Produced by `summarize-nextflow`. Structured Nextflow pipeline summary emitted by summarize-nextflow; the source for reference-data param evidence.
- Read artifact `open-requirements-ledger`. Produced by `advance-galaxy-draft-step`, `apply-galaxy-workflow-changeset`, `compare-against-iwc-exemplar`, `cwl-summary-to-galaxy-data-flow`, `cwl-summary-to-galaxy-interface`, `cwl-summary-to-galaxy-template`, `freeform-summary-to-galaxy-data-flow`, `freeform-summary-to-galaxy-interface`, `freeform-summary-to-galaxy-template`, `implement-galaxy-tool-step`, `interview-to-galaxy-workflow-changeset`, `nextflow-summary-to-galaxy-data-flow`, `nextflow-summary-to-galaxy-interface`, `nextflow-summary-to-galaxy-reference-data`, `nextflow-summary-to-galaxy-template`, `repair-galaxy-draft-topology`. Carried obligations ledger open-requirements-ledger: the run's open, resolved, and surrendered entries with their provenance. Absent on the first Mold of a run; start an empty one.

## Outputs

- Write artifact `nextflow-galaxy-reference-data` as `nextflow-galaxy-reference-data.md`. Format: `markdown`. Reviewable Markdown brief: per-asset reference inputs, requiredness, rebuild-on-absence behavior, iGenomes key disposition, open questions.
- Write artifact `open-requirements-ledger` as `open-requirements.ledger.yml`. Format: `yaml`. Carried obligations ledger re-emitted by this step: entries it appended or closed updated, every other entry passed through with its provenance intact.

## Required Tools

- None declared. Procedure should not assume external CLIs are present.

## Load Upfront

- `references/notes/nextflow-reference-data-classification.md`: Research note copied verbatim into the bundle. Classify the source pipeline's reference-data usage against the eight Nextflow-side classifications (none, reference-producing, single asset, coordinated bundle, key-expanded bundle, compute-if-missing, multi-DB pick-list, parallel bundles plus cohort data).
- `references/notes/nextflow-to-galaxy-reference-data-mapping.md`: Research note copied verbatim into the bundle. Apply the v1 Galaxy posture (explicit optional inputs + in-tool rebuild) and datatype defaults when translating each classified asset into the brief.
- `references/notes/open-requirements-ledger.md`: Research note copied verbatim into the bundle. Inherit open entries rather than re-deriving them, close the ones a settled reference-data shape discharges, and append reference obligations the pipeline leaves open — a genome key with no Galaxy data table, a reference the pipeline builds at runtime, a build the corpus doesn't carry.
- `references/schemas/summary-nextflow.schema.json`: Schema file copied verbatim into the bundle. Read `reference_assets[]` (curated path-typed inputs with `used_by` attribution to consuming subworkflows), `reference_rebuilds[]` (compute-if-missing rebuild rules with builder process and guard), and `params[]` provenance (`source_kind`, including `getGenomeAttribute` for nf-core key-expanded bundles) to determine which reference assets the source pipeline consumes and how it rebuilds them when absent.

## Load On Demand

- `references/notes/nextflow-params-to-galaxy-inputs.md`: Research note copied verbatim into the bundle. Cross-check reference selectors against the broader params translation rules — curated string enums, do-not-introduce-new-data-tables, runtime-control exclusion. Use when: a reference param looks like a curated selector (closed enum, .loc-keyed) rather than a data-bearing path.

## Validation

- None declared.

## Procedure

Classify the ways the Nextflow pipeline uses reference data using nextflow-reference-data-classification, then use nextflow-to-galaxy-reference-data-mapping to determine and describe how each classified asset maps onto the target Galaxy workflow. Output is a reviewable Markdown brief consumed by nextflow-summary-to-galaxy-interface, nextflow-summary-to-galaxy-data-flow, and nextflow-summary-to-galaxy-template before they pin workflow inputs, data-flow edges, and the gxformat2 skeleton.

### Reading the summary

The summary-nextflow schema surfaces reference-data evidence directly — read these fields before falling back to raw `workflow.conditionals[]` or guard-text inference:

- **`reference_assets[]`** — the curated asset inventory. Each entry has `param` (FK into `params[]`), `asset_kind` (`fasta`, `fasta_index`, `sequence_dictionary`, `bwa_index`, `tabix_index`, `gtf`, `gff`, `bed`, `vcf`, `database`, `other`), `required`, `source_kind`, `used_by` (consuming subworkflow/process names), and `evidence.confidence`. Use this as the primary input to classification — it already filters out non-asset path params and attributes callers.
- **`reference_rebuilds[]`** — detected compute-if-missing branches. Each entry binds `asset_param` to a `builder` process plus its `builder_outputs`, the verbatim `guard`, resolved `guard_params`, and `fallback_for` (the `<asset>_in`-style take when paired). When `reference_rebuilds[]` is non-empty, the *compute-if-missing* classification applies to those `asset_param` values. Empty `reference_rebuilds[]` does NOT mean "no compute-if-missing" — some pipelines use positive-form idioms (`if (<asset>) { unpack } else { build }`) that the detector does not yet match (see galaxyproject/foundry#229 follow-ups).
- **`params[].source_kind == "getGenomeAttribute"`** plus `params[].source_expression` — the *key-expanded bundle* signal. A pipeline that synthesizes many path params via `getGenomeAttribute('attr')` is using the iGenomes pattern; classify the bundle accordingly.
- **`subworkflows[].invocations[]`** — caller-side argument binding onto `take:` names. Use when `reference_assets[].used_by` is ambiguous or when verifying that the same param threads into multiple subworkflows.

If a needed signal is missing from the summary (e.g. RNA-seq's positive-form rebuild idiom is not yet detected), say so in the brief and flag the asset as *rebuild-unverified* rather than re-parsing source files.

## Feedback Mode

- Feedback mode is off unless the caller explicitly enables `--feedback` or supplies a feedback-ledger path.
- When enabled, read `_feedback.md` before doing the work and use its registered `foundry-feedback.ledger.yml` protocol.
- Preserve harness-owned run and phase state. Append only concrete observations about a canonical Foundry source asset or a related project that this run showed to be at fault; do not put ordinary workflow requirements in this ledger.
- Before reporting completion, make one explicit pass over the work you just did. Do not ask yourself whether anything was unclear — recall what happened: where you guessed at something the instructions should have settled, needed information this bundle does not carry, hit an instruction that contradicted another or contradicted the artifacts in front of you, used a packaged reference that did not cover your case, or did something the procedure never describes.
- Append an entry for each such event that clears the protocol's bar. If none do, append nothing and report `no feedback` explicitly. Silence and a clean pass are not the same thing, and nothing downstream can tell them apart unless you say which one it was.
- Pass the same ledger path to any subagent used for this work, and merge updates serially so one writer cannot overwrite another.

## Runtime Notes

- Do not read Foundry source files at runtime; use only files packaged in this skill bundle and user-supplied artifacts.
- Preserve declared artifact filenames unless the user or harness supplies explicit paths.
- Carry unresolved assumptions into the output artifact instead of silently inventing missing source evidence.