H₂KG Ontology

Release

Domain-specific ontology for hydrogen electrochemical systems, FAIR workflow representation, and external semantic alignment

H₂KG Ontology logoOntology release

Developed through DECODE, the Helmholtz Metadata Collaboration (HMC), and AIMWORKS to support FAIR ontology publication and reusable metadata infrastructure for hydrogen electrochemical systems.

Release Readiness

FAIR signals on this page follow a public-first policy: local build artefacts improve readiness, but full findability and accessibility require publicly established publication infrastructure. External OOPS! and FOOPS! results are reported separately as observability signals so service outages do not masquerade as ontology quality.

The detailed quality and transparency view is published on the Quality Dashboard.

FAIR Snapshot

  • 100/100 Findable
  • 100/100 Accessible
  • 100/100 Interoperable
  • 90/100 Reusable

Release Artifacts

  • Public release artefacts are established and synchronized with the current resolver-backed publication.

TIB Terminology Service submission

Canonical H₂KG 2.3.1 distribution

The full taxonomy is the canonical TIB artifact. The shared core, complete workflow distribution, compatibility graph, examples, profiles, and external-alignment bridge remain separate modules and are not submitted as the primary ontology.

  • Pre-release checks: passed
  • TIB mandatory warnings: 0
  • Classes: 2226 active + 160 deprecated
  • Properties: 23 object, 1 datatype, 9 annotation

Current preflight status

No blocking pre-release findings.

H₂KG 2.3.1 release

Intrinsic taxonomy and contextual workflow use

The recommended complete distribution combines the H₂KG taxonomy, value-free SHACL workflow profiles, and QUDT guidance. Scientific quantities remain Properties or Characteristics; their use as selected inputs is represented through Parameter records and the hasParameter / ofProperty path.

The 2.2 correction ledger records 626 parent and definition decisions across 544 classes. It distinguishes applied corrections, replacements, splits, retained decisions, and source-ambiguous compatibility terms.

The compatibility graph is retained only for migration and may contain legacy class-level workflow assertions. It does not drive canonical outputs or Explore.

Evaluation

Machine checks separated from independent review

The release records 5486 automated registry-consistency and traceability checks. These are machine checks, not reviewer judgments. A focused audit of 134 prespecified application items by two external experts is review pending.

  • Review design: 44 ITC fields, 34 asserted public mappings plus one withheld negative control, 7 diameter/radius occurrences, 24 DECODE-local concepts, and 24 edge projections
  • RQ1: independent semantic-anchor and normalized exact-label queries over one source-output population and the same 24 ITC inputs
  • Comparator: 20 author-curated capability judgments across 4 resources; automation checks graph parsing and cited-IRI presence only

No independent acceptance, precision, recall, retrieval-improvement, or ontology-superiority claim is made while review is pending. DECODE-local concepts remain project-local regardless of application-level acceptance.

Interoperability

External semantic alignment

H₂KG 2.3.1 publishes a separate semantic bridge derived from the frozen TIB snapshot dated 2026-10-14T07:32:20+00:00. The bridge contains 256 asserted mappings covering 168 H₂KG classes and 31 external ontologies.

  • Discovery scope: 3 TIB collections, 143 unique ontologies, 165 collection memberships
  • TIB benchmark assertions: 245 (skos:broadMatch=83, skos:exactMatch=140, skos:narrowMatch=22)
  • Foundational mappings: 11
  • Asserted relations: rdfs:subClassOf=2, rdfs:subPropertyOf=1, skos:broadMatch=83, skos:exactMatch=148, skos:narrowMatch=22
  • Principal benchmark ontologies: afo (36), chmo (32), voc4cat (26), chebi (16), sio (16), kitmtxx (14), bao (11), emmo (11)

The 256 alignment records comprise 245 TIB-derived SKOS mappings and 11 foundational reuse/alignment records. Foundational dependencies and reused vocabularies such as HDO, EMMO, QUDT, ChEBI, PROV-O, and DCTERMS remain distinct from ontologies aligned through the TIB benchmark. Asserted bridge links use the original external ontology IRIs; TIB catalogue URLs are retained only as provenance and discovery links. No NFDI collection is imported wholesale.

ODK Release Artefacts

ODK artefacts are now promoted and serve as the accepted machine release surface for comparison, QC, and downstream tooling.

Base ontology

Primary ODK machine artefact for downstream ontology reuse without import closure.

Generated: 2026-11-02T19:58:23+00:00 | Size: 11325600 bytes

Full ontology

Companion ODK artefact representing the broader asserted or inferred release surface for validation and distribution.

Generated: 2026-11-02T19:58:27+00:00 | Size: 3515345 bytes

Simple ontology

Companion ODK artefact optimized as a simplified machine release for tooling that prefers a smaller asserted view.

Generated: 2026-11-02T19:58:28+00:00 | Size: 3515345 bytes

Asserted Source

Supplementary machine-readable JSON-LD serialization of the asserted H₂KG source, published for inspection and downstream reuse. This download complements the ODK release artefacts.

ODK Release Status

  • Current authority: AIMWORKS pipeline
  • ODK status: promoted
  • Reasoner: ELK
  • ODK version: shadow scaffold
  • Parity: aligned

ODK artefacts are technically aligned, show no IRI drift, and have been explicitly promoted out of shadow mode.

  • Import stability GOOD configured
    Required imports are configured and surfaced through the nested ODK workbench.
  • Parity accepted GOOD aligned
    ODK base artefact currently matches the curated schema and vocabulary baseline closely enough for shadow-mode review.
  • Collected base artefact completeness GOOD complete
    The collected ODK base artefact contains the expected local schema terms.
  • Reproducible CI build GOOD good
    GitHub Actions workflows include the full ODK prepare, build, collect, and release sequence needed to reproduce the shadow build in CI.
  • No IRI drift GOOD none
    No local H₂KG IRI drift was detected between the curated release and the parity comparison target.
  • Promotion readiness GOOD promoted
    ODK promotion has been explicitly approved and the current technical gates are satisfied.

Publication Establishment

  • Overall publication status GOOD published
    This is the canonical release-state signal for whether the current ontology publication is treated as publicly established.
  • Resolver status GOOD established
    The public resolver status tracks whether the ontology namespace is established through the configured persistent identifier service.
  • Documentation publication status GOOD published
    Documentation publication tracks whether the human-readable HTML documentation is publicly established as part of the current release.
  • Machine artefact publication status GOOD published
    Machine artefact publication tracks whether the ontology serializations and release bundle are treated as publicly established release artifacts.

ODK and HDO Integration

  • HDO import status: configured
  • Last refresh: configured (shadow mode)
  • Included in current ODK build: yes
  • Local terms assessed against HDO: 141
  • Aligned to HDO: 8

HDO is the primary alignment source for data, metadata, digital-object, information-profile, identifier, schema, and validation concepts in H₂KG.

Direct HDO term reuse becomes stronger when a real HDO cache file is available at cache/sources/hdo.ttl.