RDF Dereference

This library dereferences URLs to get its RDF contents.

This tool is useful in situations where you have a URL, and you just need the parsed triples/quads, without having to concern yourself with determining the correct content type and picking the correct parser.

RDF contents are returned as an RDF stream with RDFJS-compliant quads. This library takes care of all the necessary boilerplate automatically, such as content negotiation for getting appropriate RDF serialization, decompression, following redirects, setting base URLs, and so on. If the server did not emit any content type, then the content type will be guessed based on well-known extensions.

The following RDF serializations are supported:

Internally, this library makes use of RDF parsers from the Comunica framework, which enable streaming processing of RDF.

Internally, the following fully spec-compliant parsers are used:

• N3.js
• jsonld-streaming-parser.js
• rdfa-streaming-parser.js
• rdfxml-streaming-parser.js

If you need something more low-level with more control, have a look at rdf-parse.

Installation

$ npm install rdf-dereference

or

$ yarn add rdf-dereference

This package also works out-of-the-box in browsers via tools such as webpack and browserify.

Require

import rdfDereferencer from "rdf-dereference";

or

const rdfDereferencer = require("rdf-dereference").default;

Usage

Dereferencing an RDF document

The rdfDereferencer.dereference method accepts an URL, and outputs a promise resolving to an object containing a quad stream.

const { quads } = await rdfDereferencer.dereference('http://dbpedia.org/page/12_Monkeys');
quads.on('data', (quad) => console.log(quad))
     .on('error', (error) => console.error(error))
     .on('end', () => console.log('All done!'));

Such a stream is useful when the RDF document is huge, and you want to process it in a memory-efficient way.

Dereferencing works with any kind of RDF serialization, even HTML documents containing RDFa and JSON-LD:

const { quads1 } = await rdfDereferencer.dereference('https://www.rubensworks.net/');
const { quads2 } = await rdfDereferencer.dereference('https://www.netflix.com/title/80180182');

Dereferencing a local file

Similar as above, the rdfDereferencer.dereference method also accepts file paths.

const { quads } = await rdfDereferencer.dereference('path/to/file.ttl', { localFiles: true });
quads.on('data', (quad) => console.log(quad))
     .on('error', (error) => console.error(error))
     .on('end', () => console.log('All done!'));

Note that the localFiles flag MUST be enabled before local paths can be dereferenced for security reasons.

This feature is not available when this package is used within a browser environment.

Importing the resulting quads into a store

These resulting quads can easily be stored in a more convenient datastructure using tools such as rdf-store-stream:

import {storeStream} from "rdf-store-stream";

const store = await storeStream(quads);

const resultStream = store.match(namedNode('http://example.org/subject'));

Advanced features

Input: Passing custom headers

You can pass custom headers for the HTTP request via the options object:

const { quads } = await rdfDereferencer.dereference('https://www.netflix.com/title/80180182', {
  headers: {
    'Accept-Datetime': 'Thu, 31 May 2007 20:35:00 GMT',
  },
});

By default, the GET method will be used.

Input: Setting the HTTP method

You can define the HTTP method via the options object:

const { quads } = await rdfDereferencer.dereference('https://www.netflix.com/title/80180182', {
  method: 'POST',
});

By default, the GET method will be used.

Output: Determining the final URL

If dereferencing went through various redirects, it may be useful to determine the final URL. This can be done using the url field of the output object:

const { quads, url } = await rdfDereferencer.dereference('https://www.netflix.com/title/80180182');
console.log(url); // The final URL, e.g. https://www.netflix.com/at-en/title/80180182

Output: Response Headers

This library will return the HTTP response headers as a hash:

const { quads, headers } = await rdfDereferencer.dereference('https://ruben.verborgh.org/profile/');
console.log(headers); // Example: { 'content-length': '65701' }

Output: Triples or Quads

Some RDF serializations don't support named graphs, such as Turtle and N-Triples. In some cases, it may be valuable to know whether or not an RDF document was serialized with such a format. If this was the case, the triples flag will be set to true on the resulting object:

const { quads, triples } = await rdfDereferencer.dereference('https://ruben.verborgh.org/profile/');
console.log(triples); // If the document only supported triples, true in this case, since it returned Turtle.

Command line usage

A CLI version of this tool exists, which can be installed globally as follows:

$ npm install -g rdf-dereference

After that, you can dereference any URL to a compact JSON-based quad representation:

$ rdf-dereference https://www.rubensworks.net/
[
{"subject":"https://www.rubensworks.net/","predicate":"http://xmlns.com/foaf/0.1/primaryTopic","object":"https://www.rubensworks.net/#me","graph":""},
{"subject":"https://www.rubensworks.net/","predicate":"http://xmlns.com/foaf/0.1/maker","object":"https://www.rubensworks.net/#me","graph":""},
{"subject":"https://www.rubensworks.net/#me","predicate":"http://www.w3.org/1999/02/22-rdf-syntax-ns#type","object":"http://xmlns.com/foaf/0.1/Person","graph":""},
{"subject":"https://www.rubensworks.net/#me","predicate":"http://xmlns.com/foaf/0.1/name","object":"\"Ruben Taelman\"","graph":""},
...

After that, you can dereference local files, for which the content type will be identified by extension:

$ rdf-dereference path/to/file.ttl
...

License

This software is written by Ruben Taelman.

This code is released under the MIT license.