@shexjs/node

@shexjs/loader for a node environment: the same load() API, extended with

• file system access — a source may be a plain file path or a file: URL;
• stdin — the source - reads standard input;
• loadExtensions(globsOrPackageNames) — dynamic loading of semantic-action extensions, which is how @shexjs/cli's --extension works.

It reaches for node's fs, so it's not for browsers — use @shexjs/loader there.

Install

npm install @shexjs/node

load(schema, data, schemaOptions?, dataOptions?)

Load any mix of ShExC/ShExJ schema sources and Turtle/JSON-LD data sources into one schema and one graph, remembering where everything came from. A SOURCE may be:

• a file path or URL — where to load the item;
• {text: string, url: string} — an already-loaded resource;
• a ShExJ object (schema) or an RDF/JS store (data).

Given a local 1dotOr2dot.shex (say PREFIX : <http://a.example/> :S1 { :p1 . | :p2 .; :p3 . }) and a local p2p3.ttl (PREFIX : <http://a.example/> <x> :p2 "p2-0" ; :p3 "p3-0" .):

const ShExLoader = require("@shexjs/node")({
  rdfjs: require("n3"),          // an RDF/JS implementation
  fetch: globalThis.fetch,  // for http(s): URLs
});

ShExLoader.load(
  { shexc: [ "./1dotOr2dot.shex" ] },  // a file path
  { turtle: [                          // graphs merge
    "https://shex.io/webapps/packages/shex-cli/test/cli/p1.ttl",
    "./p2p3.ttl",
  ] }
).then(({schema, schemaMeta, data, dataMeta}) => {
  console.log("shapes:  " + schema.shapes.map(s => s.id).join(" "));
  console.log("sources: " + dataMeta.map(m => m.url).join("\n         "));
  console.log("triples: " + data.getQuads().map(
    q => ["subject", "predicate", "object"].map(t => q[t].value).join(" "))
    .join("\n         "));
});

shapes:  http://a.example/S1
sources: https://shex.io/webapps/packages/shex-cli/test/cli/p1.ttl
         file:///…/p2p3.ttl
triples: https://shex.io/webapps/packages/shex-cli/test/cli/x http://a.example/p1 p1-0
         file:///…/x http://a.example/p2 p2-0
         file:///…/x http://a.example/p3 p3-0

Relative IRIs in each document resolve against that document's own location — which is the point of loading through this module rather than Fs.readFile: the file: and https: triples above landed in one graph with their provenance intact, and schemaMeta/dataMeta carry each source's url, base and prefixes (ready for shape-map's parser).

See @shexjs/loader for the returned structure, JSON-LD support (pass jsonld: require("jsonld")) and the other options; everything there works here.

@shexjs/node is one of the shex.js packages; installing shex pulls in the whole suite, and its README maps them.