Skip to content

Latest commit

 

History

History
327 lines (244 loc) · 11.1 KB

README.md

File metadata and controls

327 lines (244 loc) · 11.1 KB

@zazuko/vocabularies -- Zazuko's Default Ontologies & Prefixes

GitHub Workflow Status Codecov npm

This package contains a distribution of the most commonly used RDF ontologies (schema/vocab, whatever you call it) including their default prefixes, together with a set of utility functions to work with prefixes.

It is extending RDFa Core Initial Context and contains what we consider commonly used prefixes. Some popular prefixes do not resolve to dereferencable RDF and are thus skipped.

The package is built for use in Node.js projects. We ship N-Quads files of the vocabularies so it could be useful for other programming languages as well as you do not have to take care of downloading the ontologies yourself.

Installation

This is a new release of the previous monolithic package which grew quite a lot.

It is recommended to install the prefixes package and select vocabulary packages individually to reduce the size of node_modules.

$ npm install @zazuko/prefixes @vocabulary/rdf @vocabulary/schema ...

You may still choose to install a one package to get them all.

$ npm install @zazuko/vocabularies

It remains pretty much compatible with the previous @zazuko/rdf-vocabularies package. The main difference is that now it is ESM-only.

Usage

(Read below and take a look at some examples.)

Dataset-as-code modules

All vocabularies published from this package are also exported as JS modules so that then can be imported synchronously (no parsing required) and without additional dependencies when in web app setting (see the raw-loader instructions below).

Modules @rdf-vocabularies/datasets exports factories which returns an array of quads Quad and take RDF/JS DataFactory as parameter.

import $rdf from 'rdf-ext'
import schema from '@vocabulary/schema'

const dataset = $rdf.dataset(schema({ factory: $rdf }))

Vocabularies Metadata

See _index.nq.

vocabularies()

Exported from @zazuko/vocabularies', the vocabularies() function accepts an optional options object:

  • options.only: Array?, default: undefined, a subset of all available prefixes, will only load these.
  • options.factory: RDF/JS DatasetFactory, default: rdf-ext, a dataset factory abiding by the RDF/JS Dataset Specification, used to create the returned datasets.
  • options.stream: Boolean, default: false, whether to return a RDF/JS quad stream instead of regular objects/datasets.

Loading all Vocabularies as Datasets

In browser environment this will cause a request for each individual dataset. It is thus recommended to always only load the needed ontologies to reduce the unnecessary traffic and save bandwidth.

import { vocabularies } from '@zazuko/vocabularies'

vocabularies()
  .then((datasets) => {
    /* `datasets` is:
    {
      "csvw": Dataset,
      "sd": Dataset,
      "ldp": Dataset,
      "schema": Dataset,
      "owl": Dataset,
      "void": Dataset,
      "sioc": Dataset,
      "foaf": Dataset,
      "time": Dataset,
      "dcat": Dataset,
      "oa": Dataset,
      "gr": Dataset,
      "rdf": Dataset,
      "cc": Dataset,
      "ssn": Dataset,
      "rr": Dataset,
      "rdfa": Dataset,
      "org": Dataset,
      "sosa": Dataset,
      "dc11": Dataset,
      "skos": Dataset,
      "dqv": Dataset,
      "prov": Dataset,
      "og": Dataset,
      "qb": Dataset,
      "rdfs": Dataset,
      "dc": Dataset,
      "ma": Dataset,
      "vcard": Dataset,
      "grddl": Dataset,
      "dcterms": Dataset,
      "skosxl": Dataset,
      "wgs": Dataset,
      "dbo": Dataset,
      "dbpedia": Dataset,
      "dbpprop": Dataset,
      "rss": Dataset,
      "cnt": Dataset,
      "vs": Dataset,
      "hydra": Dataset,
      "gn": Dataset,
      "gtfs": Dataset,
      "geo": Dataset,
      "geof": Dataset,
      "geor": Dataset
    }
    */
  })

Loading only some Vocabularies as Datasets

import { vocabularies } from '@zazuko/vocabularies'

vocabularies({ only: ['rdfs', 'owl', 'skos'] })
  .then((datasets) => {
    /* `datasets` is:
    {
      "owl": Dataset,
      "skos": Dataset,
      "rdfs": Dataset
    }
    */
  })

Getting a Readable Stream (Quad Stream)

import { vocabularies } from '@zazuko/vocabularies'
const stream = await vocabularies({ stream: true, only: ['rdfs', 'owl', 'skos'] })

Using vocabularies function in browser

The preferred usage in browser projects is to avoid importing from @zazuko/vocabularies because that will require additional bundling of dynamic n-quads modules.

Instead, import from the partial modules (without even installing the big package, if possible):

  • import { expand } from '@zazuko/prefixes/expand'
  • import { prefixes } from '@zazuko/prefixes/prefixes'
  • import { shrink } from '@zazuko/prefixes/shrink'

The module @zazuko/vocabularies/expandWithCheck requires rdf-ext and parses datasets. See the instructions below for examples how to configure the application.

The package's main module can also be used in browser albeit it needs a bundler such as webpack and additional steps to configure it:

The package can be used in browser albeit it needs a bundler such as webpack and additional steps to configure it:

  • Enable dynamic imports. In webpack it is done with @babel/plugin-syntax-dynamic-import

  • Extend the bundler setup to have it load the contents of vocabulary files (all n-triples). In In webpack it can be done with raw-loader:

    module: {
      rules: [
        {
          test: /\.nq$/,
          use: ['raw-loader']
        }
      ]
    }
    
  • Be careful with prefetching chunks. Some applications may generate prefetch links for dynamically loaded chunks. Some of the ontology files are quite large and their number will grow over time. Hence, it may be desired to exclude certain chunks from the being eagerly loaded. Check the wiki for examples.

Expanding a Prefix

expanding means: 'xsd:dateTime' → 'http://www.w3.org/2001/XMLSchema#dateTime'. It is the opposite of shrinking:
expand(shrink('http://www.w3.org/2001/XMLSchema#dateTime')) === 'http://www.w3.org/2001/XMLSchema#dateTime'

There are two ways of expanding a prefix:

  • vocabularies.expand(prefixedTerm: String): String synchronous

    Expand without checks. It is similar to prefix.cc in the sense that prefix.cc would expand schema:ImNotInSchemaDotOrg to http://schema.org/ImNotInSchemaDotOrg.

  • vocabularies.expand(prefixedTerm: String, types: Array<String|NamedNode>): Promise<String> asynchronous

    Expand with type checks. types is an array of strings or NamedNodes. See this example:

    const { expand } = require('@zazuko/rdf-vocabularies')
    const Class = expand('rdfs:Class')
    const Property = expand('rdf:Property')
    
    // Will return <schema:person> expanded to `http://schema.org/Person`
    // iff the dataset contains either:
    //   <schema:Person> <rdf:type> <rdfs:Class>
    // or
    //   <schema:Person> <rdf:type> <rdf:Property>
    await expand('schema:Person', [Class, Property])

Shrinking an IRI

shrinking means: 'http://www.w3.org/2001/XMLSchema#dateTime' → 'xsd:dateTime'. It is the opposite of expanding:
shrink(expand('xsd:dateTime')) === 'xsd:dateTime'

  • vocabularies.shrink(iri: String): String

    Note: returns empty string when there is no corresponding prefix. Always check the output when using shrink with user-provided strings.

    import assert from 'assert'
    import { shrink } from '@zazuko/prefixes'
    
    assert(shrink('http://www.w3.org/2001/XMLSchema#dateTime') === 'xsd:dateTime')
    assert(shrink('http://example.com#nothing') === '')
    
    const iri = 'http://example.com#nothing'
    const stringToDisplay = shrink(iri) || iri
    console.log(stringToDisplay) // 'http://example.com#nothing'

Accessing Prefixes: vocabularies.prefixes

Getting an object with prefixes and their base URI:
(Returns this object.)

import { prefixes } from '@zazuko/prefixes'

console.log(prefixes)
/*
 {
  v: 'http://rdf.data-vocabulary.org/#',
  csvw: 'http://www.w3.org/ns/csvw#',
  sd: 'http://www.w3.org/ns/sparql-service-description#',

}
*/

Accessing Data Files from the Package

Accessing the N-Quads files:

import path from 'path'
import module from 'module'

const { resolve } = module.createRequire(import.meta.url)

console.log(path.resolve(resolve('@vocabulary/skos/skos.nq')))

Command line

The package also includes a simple command line interface which forwards the vocabulary datasets to standard output. It can be used in two ways.

By prefix:

rdf-vocab prefix foaf

By namespace URI:

rdf-vocab prefix http://schema.org/

Versioning Scheme

The packages follow semver but that is to indicate library code changes rather than vocabulary changes. Usually, that is. If we identify significant changes to the source RDF for any given vocabulary, we may decide to bump a major version.

Adding new prefixes

New prefixes can be added by opening a pull request on Github. For new requests, first check if the creator/owner of the namespace defined a prefix. If not check prefix.cc. In case prefix.cc is ambiguous a discussion should be raised before the pull-requests gets integrated. Last thing to check are the predefined namespaces in the DBpedia SPARQL endpoint or other popular RDF resources like LOV. If you find one please refer to it in the pull request.

Steps to add a prefix

  1. Add a new directory under ontologies with a package.json. For a minimal example see ACL vocabulary
  2. If necessary, add overrides to the vocabulary key, similar to the others
    • for the file option, a file: scheme IRI can be used, with path relative to the repository root
  3. Run npm run fetch in the vocabulary's dir to process the triples.
  4. Add a dependency to vocabularies meta package
  5. Commit changes and submit a PR

Project-specific prefixes

It is also possible to add prefix within a project so that it can be used with the functions expand and shrink.

import { prefixes, expand, shrink } from '@zazuko/prefixes'

prefixes['foo'] = 'http://example.com/'

// 'http://example.com/bar'
const foobar = expand('foo:bar')

// 'foo:bar'
const prefixed = shrink(foobar)