[AnVIL DX] Add AnVIL datasets to Google Datasets catalog

[NoopDog]

Add AnVIL datasets to the Google Dataset Search catalog so they're discoverable from Google. This is done by embedding schema.org Dataset JSON-LD on dataset detail pages — Google's crawler picks it up.

Companion to galaxyproject/brc-analytics#1264 and #4806.

Reference implementation

NCPI Dataset Catalog has already shipped this for studies. Mirror their pattern:

• Builder + types: app/utils/schemaOrg.ts — SchemaDataset interface and buildStudyJsonLd() factory.
• Render component: app/components/Detail/components/StudyJsonLd/studyJsonLd.tsx — wraps the JSON-LD in a <script type="application/ld+json"> inside next/head, with HTML escaping to prevent script injection.
• Page integration: pages/[entityListType]/[...params].tsx — mounted on the detail route only.
• Tests: app/utils/schemaOrg.test.ts — covers required fields, truncation, and conditional fields.

Google Dataset required + recommended fields

Per Google's Dataset structured data guidelines:

Required

• name — descriptive title
• description — 50–5000 characters

Recommended

• identifier, url, sameAs
• creator, funder, license
• distribution (with contentUrl, encodingFormat)
• keywords, variableMeasured, measurementTechnique
• spatialCoverage, temporalCoverage
• includedInDataCatalog, isAccessibleForFree, version, citation

Initial mapping — AnVIL dataset (DatasetEntity) → Dataset

Source entity at app/apis/azul/anvil-cmg/common/entities.ts (DatasetEntity).

schema.org field	Source / value
@context	"https://schema.org"
@type	"Dataset"
name	title (fall back to dataset_id)
description	description — strip HTML, truncate to 5000 chars (pad to ≥50 chars from title/consortium when needed)
identifier	[dataset_id, ...registered_identifier] (e.g. dbGaP phs accessions)
url	${browserURL}/datasets/${dataset_id}
sameAs	dbGaP and other registered identifier URLs derived from registered_identifier
includedInDataCatalog	{ "@type": "DataCatalog", name: "AnVIL Data Explorer", url: browserURL }
isAccessibleForFree	Map from accessible / data use restrictions (datasets behind controlled access → false)
keywords	Union of consortium, data_modality, phenotypic_sex, reported_ethnicity, species, disease (from biosamples/donors), library prep
creator	{ "@type": "Organization", name: consortium }
funder	TBD — confirm whether AnVIL-side funder mapping is available
distribution	DataDownload[] derived from manifest/curl/Terra export endpoints (mark as access-controlled where applicable)
variableMeasured	Optional PropertyValue[] from dataset summary counts (donors, biosamples, files, libraries)
license	TBD — confirm with team (likely DUO-derived terms)

Open questions for funder / license / how to express controlled access in isAccessibleForFree and distribution should be resolved before merge.

Implementation steps

1. Add app/utils/schemaOrg.ts (or AnVIL-namespaced equivalent) with SchemaDataset types and buildDatasetJsonLd(dataset, browserURL).
2. Add a DatasetJsonLd component that renders the JSON-LD via next/head with the same HTML-escape helper as NCPI.
3. Mount the component on the dataset detail page (pages/datasets/[entityId]).
4. Unit-test the builder: required fields present, description truncation, conditional fields omitted when source is null, controlled-access handling.
5. Validate output against Google's Rich Results Test and Schema Markup Validator for representative datasets (open access, controlled access, multi-consortium).
6. Once shipped, request indexing via Google Search Console and confirm dataset pages start appearing in Google Dataset Search.

Out of scope (follow-ups)

• JSON-LD on biosamples/donors/files/libraries detail pages.
• Sitemap entries for dataset detail pages if not already complete.
• Mirroring this on the AnVIL Catalog explorer (separate ticket).