@ontologies/core

Like definitely typed, but for ontologies

Usage no npm install needed!

<script type="module">
  import ontologiesCore from 'https://cdn.skypack.dev/@ontologies/core';
</script>

README

Ontologies

Never manage a namespace object map again, scrap typo's for well-known ontologies. Like DefinitelyTyped, but for ontologies.

Usage

@ontologies/core

When working with RDF (linked data) ontologies are very important, but being able to quickly create and work with the fundamental building blocks of RDF is equally important. @ontologies/core exports an object (data factory) which aims to to just that.

Assuming the following import in the examples:

import rdf from "@ontologies/core"

Create literals

// Strings
rdf.literal("Hello world!") 
// { termType: "Literal", value: "Hello World!", datatype: { termType: "NamedNode", value: "http://www.w3.org/2001/XMLSchema#string" } }
// Numbers
rdf.literal(9001) 
// { termType: "Literal", value: "9001", datatype: { termType: "NamedNode", value: "http://www.w3.org/2001/XMLSchema#integer" } }

Most JS literals will be mapped to their RDF (xsd) counterparts, passing the datatype explicitly is also possible. Please note that data types in RDF must be IRIs.

rdf.literal("(5,2)", rdf.namedNode("http://example.com/types/myCoordinate")) 
// { termType: "Literal", value: "(5,2)", datatype: { termType: "NamedNode", value: "http://example.com/types/myCoordinate" } }

Create links

Use namedNode to create links to other resources which have been named, meaning an authority has given the resource a fixed identifier on their domain. This can be a resource on the web (e.g. http:, https:) but also in the internet (e.g. ftp: or magnet:) or elsewhere (e.g. urn:isbn: or doi:). Note that choosing schemes which are widely deployed (e.g. https:) will allow others easier access to find, access, and share your data.

rdf.namedNode("https://schema.org/Thing") 
// { termType: "NamedNode", value: "https://schema.org/Thing" }

rdf.namedNode("https://example.com/myDocument#paragraph") 
// { termType: "NamedNode", value: "https://example.com/myDocument#paragraph" }

rdf.namedNode("urn:isbn:978-0-201-61622-4") 
// { termType: "NamedNode", value: "urn:isbn:978-0-201-61622-4" }

Cool iris don't change, but designing systems is difficult, so there will be resources which don't have their own name or are too expensive to assign a name. This is where blank nodes can be used, these are resources as well, but with "an identifier yet to be assigned". This is a bit of a tricky way of saying "it thing has a name, but I don't know it yet".

// Create blank nodes (links which have no place in the web yet)
rdf.blankNode() 
// { termType: "BlankNode", value: "b0" }

rdf.blankNode("fixed") 
// { termType: "BlankNode", value: "fixed" }

Note that most of the time blank nodes can be replaced with fragment (#) IRIs without much trouble.

{
  "@id": "http://example.com/myCollection",
  "members": {
    "@id": "http://example.com/myCollection#members", // Append `#members` to the base rather than use a blank node
    "@type": "http://www.w3.org/1999/02/22-rdf-syntax-ns#Seq",
    "http://www.w3.org/1999/02/22-rdf-syntax-ns#_0": "First item",
    "http://www.w3.org/1999/02/22-rdf-syntax-ns#_1": "Second item",
    "http://www.w3.org/1999/02/22-rdf-syntax-ns#_2": "Third item",
  }
}

Create statements

// Create quads (Statements about some thing in the world)
rdf.quad(s, p, o) 
// { subject: <s>, predicate: <p>, object: <o>, graph: <defaultGraph> }
// Compare rdf objects
import { Thing } from "@ontologies/schema";
console.log(rdf.equals(rdf.namedNode("https://schema.org/Thing"), Thing)) 
// true


// Serialize to n-triples/quads
console.log(rdf.toNQ())

Other exports

Overview of the exports of @ontologies/core;

  • A default export which is a proxy to the currently assigned global Data Factory.
  • A named export globalSymbol which is a symbol to identify the Data Factory used by the other @ontologies/ packages to create rdf objects with.
  • A named export setup which binds the PlainFactory to the global scope under the globalSymbol identifier if it was previously undefined.
  • A named export globalFactory which should be a reference to the last .
  • A named export PlainFactory which implements the Data Factory interface (with slight adjustments) in a functional way (e.g. no instance methods, but provides an equals function on the factory itself).
  • A named export createNS which you can use to create namespaces with which ease NamedNode creation using the global factory.
  • A small set of types useful for working with RDF.

@ontologies/*

The other packages are generated from their respective ontologies, providing client applications with importable symbols and a named export ns with which custom additional terms can be created within the given namespace.

import { name } from '@ontologies/schema'

console.log(name) // http://schema.org/name

All terms

import * as schema from '@ontologies/schema'

console.log(schema.name) // http://schema.org/name

Custom terms

import { ns } from '@ontologies/schema'

console.log(ns('extension')) // http://schema.org/extension

Use .value for the string representation

import { name } from '@ontologies/schema'

console.log(name.value) // "http://schema.org/name"

Overriding the default factory

The default factory used is the PlainFactory from this package. This is a factory which should suffice most needs, but certain JS RDF libraries expect more methods to be available on the factory. It is possible to override the factory with a custom one.

Initialize a custom factory by calling setup() from @ontologies/core

import { setup } from "@ontologies/core";
import myFactory from "./myFactory";

setup(new myFactory);

Library authors who want to provide an alternate default than the PlainFactory but who don't want to override the end-users setting can soft-override the factory;

import { setup } from "@ontologies/core";
import LibFactory from "./LibFactory";

setup(new LibFactory(), false); // Passing false will override the default but not a user-set factory.

Help, my factory isn't loaded!

Chances are you have called setup too late in the module initialization cycle. Be sure to;

  1. Move the setup call in a separate file.
  2. Don't import any package using the default export from @ontologies/core in that file.
  3. Import that file before any other import which uses default export from @ontologies/core in that file.

Non-js symbols

Dashes in term names are replaced with underscores. The default export contains both the verbatim and the underscored values.

import dcterms, { ISO639_2 } from '@ontologies'

console.log(ISO639_2) // NamedNode(http://purl.org/dc/terms/ISO639-2)
console.log(dcterms.ISO639_2) // NamedNode(http://purl.org/dc/terms/ISO639-2)
console.log(dcterms['ISO639-2']) // NamedNode(http://purl.org/dc/terms/ISO639-2)

Collisions with ES reserved keywords

If a term collides with an ES 5/6 reserved keyword or certain built-in classes, the term is prepended with the symbol of the ontology:

import { name, schemayield } from '@ontologies/schema'

// 'name' is not a JS reserved keyword
console.log(name.value) // "http://schema.org/name"
// 'yield' is a reserved keyword, so the package name is prepended to the js identifier.
console.log(schemayield.value) // "http://schema.org/yield"