digibib/ld-graph

Makes JSON-LD less painfull to work with

LD-graph

LD-graph is a small library intended to make JSON-LD less painfull to work with.

Status

Work in progress. Unstable. API not finalized

Usage

Example input used throughout the documentation:

var JSONLD = {
  "@context": {
        "dc": "http://purl.org/dc/elements/1.1/",
        "ex": "http://example.org/vocab#",
        "xsd": "http://www.w3.org/2001/XMLSchema#"
  },
  "@graph": [
    {
      "@id": "http://example.org/work/1",
      "@type": "ex:Work",
      "dc:title": {"@value": "Cat's cradle", "@language": "en"},
      "ex:creator: "http://example.org/person/1"
      "ex:themes": ["free will", "science", "nucelar war"]
    },
    {
      "@id": "http://example.org/person/1",
      "ex:name": "Kurt Vonnegut",
      "ex:born": {
        "@value": "1922",
        "@type": "xsd:gYear"
      },
      "ex:dead": {
        "@value": "2007",
        "@type": "xsd:gYear"
      }
    },
    {
      "@id": "http://example.org/publication/1",
      "@type": "ex:Publication",
      "ex:publicationOf: "http://example.org/work/1"
      "ex:date": "1963"
      "ex:hasExemplars": [
        {"@id": "http://example.org/item/1"},
        {"@id": "http://example.org/item/2"},
      ]
    },
    {
      "@id": "http://example.org/item/1",
      "@type": "ex:Item",
      "ex:exemplarOf: "http://example.org/publication/1"
      "ex:status": "available"
    },
    {
      "@id": "http://example.org/item/2",
      "@type": "ex:Item",
      "ex:exemplarOf: "http://example.org/publication/1"
      "ex:status": "on-loan"
    }
  ]
}

Import the library:

var graph = require("ld-graph");

Parse a RDF graph in JSON-LD format into a graph object:

var g = graph.parse(JSONLD);

You can now acces a node by Id(URI):

var w = g.byId("http://example.org/work/1"),
    p = g.byId("http://example.org/publication/1"),
    i1 = g.byId("http://example.org/item/1"),
    i2 = g.byId("http://example.org/item/2");

w.id
\\ => "http://example.org/work/1"

You can also ask for nodes by type, which returns an array of nodes.

var items = g.byType("Item");

Given a node, you can ask for it's literal values by the get method, which returns a Literal object with it's value, datatype and language-tag (if present):

w.get("title")
// => {value: "Cat's cradle", type: "http://www.w3.org/2000/01/rdf-schema#langString", "lang": "en"}

If the node doesn't have a property defined for the given key, an error is raised.

The get method only returns the first literal, even if there are more defined. To get them all, use getAll, which returns an array of Literal objects (or an empty array, if the property is not defined for the node):

w.getAll("themes").map(function (l) { return l.value; })
\\ => ["free will", "science", "nucelar war"]

w.getAll("xyz")
\\ => []

To navigate from one node to another - use the in method for incoming relations and the out method for outgoing relations.

w.in("publicationOf").id
\\ => "http://example.org/publication/1"

p.out("hasExemplar").id
\\ => "http://example.org/item/1"

To get an array of all matching nodes, use inAll and outAll:

p.inAll("exemplarOf").map(function(n) { return n.id })
\\ => ["http://example.org/item/1", "http://example.org/item/2"]

p.outAll("hasExemplar").map(function(n) { return n.id })
\\ => ["http://example.org/item/1", "http://example.org/item/2"]

Limitations

• JSON-LD is a very flexible format - allowing you to represent the same information in multiple ways. Not all of these representations are supported yet, but may eventually be so when I encounter them in my work. Currently unsupported keywords: @container, @list, @reverse, @index, @vocab, @base, @set, :.
• No support for blank nodes whatsoever.