NKR-vocabs

Repository for managing RDF/SKOS vocabularies used in the NKR/CCMM ecosystem.

This repository provides:

• RDF/SKOS vocabulary storage
• Multi-format serialization support
• Validation workflows
• SKOS normalization/fixing workflows
• GitHub Actions automation

Supported formats:

• Turtle (.ttl)
• RDF/XML (.rdf)
• JSON-LD (.jsonld)
• N-Triples (.nt)
• Source XML (.xml) for conversion sources only

---

Repository Structure

.
├── .github/
│   └── workflows/
│       ├── validate-skos.yml
│       ├── fix-skos.yml
│       └── convert-vocabularies.yml
│
├── scripts/
│   ├── validate_skos.sh
│   ├── fix_skos.sh
│   └── convert_vocabularies.py
│
├── vocabularies/
│   ├── access_rights/
│   ├── resource_types/
│   ├── languages-skos/
│   ├── filetypes-skos/
│   └── ...
│
└── README.md

---

Vocabulary Serialization Strategy

Each vocabulary folder should contain equivalent serializations of the same vocabulary.

Example:

resource_types/
├── resource_types.rdf
├── resource_types.ttl
├── resource_types.jsonld

---

Source Priority Rules

The conversion script uses priority-based source selection.

Priority order:

1. .rdf
2. .ttl
3. .jsonld
4. .nt
5. .xml

Meaning:

• If .rdf exists, it becomes the canonical source
• Otherwise .ttl
• Otherwise .jsonld
• Otherwise .nt
• Otherwise .xml

Generated formats:

Source	Generated
.rdf	.ttl, .jsonld
.ttl	.rdf, .jsonld
.jsonld	.rdf, .ttl
.nt	.rdf, .ttl, .jsonld
.xml	.rdf, .ttl, .jsonld

---

Validation Script

Script:

scripts/validate_skos.sh

Purpose:

• RDF syntax validation using Apache Jena riot
• SKOS validation using skosify

Supported validation formats:

• .ttl
• .rdf
• .jsonld
• .json-ld
• .nt

Regular XML source files are NOT validated.

Example usage

Validate all vocabularies:

./scripts/validate_skos.sh vocabularies

Validate one vocabulary:

./scripts/validate_skos.sh vocabularies/resource_types

---

SKOS Fix Script

Script:

scripts/fix_skos.sh

Purpose:

• Normalize SKOS vocabularies
• Remove redundant hierarchy relations
• Clean labels
• Apply skosify cleanup rules

Example usage

./scripts/fix_skos.sh vocabularies

After fixing:

./scripts/validate_skos.sh vocabularies

---

Vocabulary Conversion Script

Script:

scripts/convert_vocabularies.py

Purpose:

• Convert vocabularies between RDF serializations
• Auto-select canonical source format by priority
• Generate missing formats

Run Conversion

Convert all vocabularies:

python scripts/convert_vocabularies.py vocabularies --overwrite

Convert one vocabulary:

python scripts/convert_vocabularies.py vocabularies/resource_types --overwrite

---

GitHub Actions

Validate SKOS

Workflow:

.github/workflows/validate-skos.yml

Run:

GitHub → Actions → Validate SKOS → Run workflow

---

Fix SKOS

Workflow:

.github/workflows/fix-skos.yml

Run:

GitHub → Actions → Fix SKOS → Run workflow

---

Convert Vocabularies

Workflow:

.github/workflows/convert-vocabularies.yml

Run:

GitHub → Actions → Convert Vocabularies → Run workflow

---

Notes About XML

There are two XML cases.

RDF/XML

Supported RDF serialization:

file.rdf

Handled by:

• validation
• fixing
• conversion

Regular XML Source Files

Example:

iana-media-types.xml

These are NOT RDF/XML.

They are source data and require dedicated XML → SKOS conversion logic.

They are intentionally excluded from validation/fixing workflows.