Linked Data Fetch: a HTTP client for RDF resources

For both the browser as the NodeJS framework it adds features specifically for getting RDF resources.

Supports small extra building blocks:

• Prefers to read TriG, but also reads N3, turtle, N-Triples, N-Quads, RDFa or JSON-LD (content-negotiation)
• Returns an N3 store containing the triples for the current document
• Emits events for: request, response, redirect, cache-hit, cache-miss and parsed

NodeJS framework in specific

• Automatically follows redirects
• Able to be configured with HTTP caching
• Able to limit the amount of concurrent requests and schedule these

npm install ldfetch

And using browserify you can also compile it for browser purposes:

browserify lib/ldfetch-browser.js > build.js

<script src="build.js"></script>
<script>
  var fetch = new window.ldfetch();
  //use it as described bellow
</script>

How to use it

A small example fetching the previous page of a hydra paged collection and returning its subjects

var ldfetch = require('../lib/ldfetch.js'),
    n3 = require('n3');
var options = {headers: {'Accept-Datetime': '2017-03-11T17:00:00.000Z'}}; // optional
var fetch = new ldfetch(options);
fetch.addPrefix("hydra","http://www.w3.org/ns/hydra/core#");
fetch.get(url).then(response => {
  console.error("Redirected to: " + response.url);
  response.store = new n3.Store(response.triples,{prefixes: response.prefixes});
  console.log("Requesting the previous page: " + response.store.getTriples(null,"hydra:previous")[0].object);
  fetch.get(response.store.getTriples(null,"hydra:previous")[0].object).then((response2) => {
    response2.store = new n3.Store(response2.triples,{prefixes: response2.prefixes});
    //return the subjects:
    console.log(response2.store.getSubjects());
  });
});

If HTTP requests with specific headers are needed, the options object may be used by defining an object inside of it, named headers containig HTTP header names and values.

The response object will look like this:

{
  "responseCode": 200,
  "triples": [], //Following the N3.js triple representation
  "prefixes": {"hydra": "http://www.w3.org/ns/hydra/core#"},
  "url": "//url after redirects"
}

What we want you to build

This library was built for the use within hypermedia agents that follow RDF links.

Adding browsing features (although I do not have a concrete use case for this today):

• Optionally keeps a history which allows you to jump back, forward, or study the list of visited URIs
• Able to stop loading a request
• Refresh a page based on HTTP caching headers (max-age and/or poll for eTag changes? Or just ping when the current page is out of date?)

Automatically exploiting Hydra links, such as hydra:next, hydra:previous, hydra:search, hydra:filter, etc. The Linked Data Fragments client for recognizing Triple Pattern Fragment and the Linked Connections client for recognizing an lc: departureTimeQuery

Architecture

Focuses on two parts:

1. 2 classes, one for in the browser, and one for in NodeJS (adding caching, redirects and concurrent requests), that take care of the HTTP abstraction
2. A class using 1. that picks the right parser and parses the data and stores it in a memory structure/store of choice