Add ActorExtractLinksHeaders for following auxiliary resources in Solid

engines/config-query-sparql-link-traversal/config/config-solid-default.json

@@ -7,6 +7,7 @@
     "ccqslt:config/config-solid-base.json",
     "ccqslt:config/extract-links/actors/predicates-common.json",
     "ccqslt:config/extract-links/actors/predicates-ldp.json",
+    "ccqslt:config/extract-links/actors/links-describedby.json",
     "ccqslt:config/extract-links/actors/predicates-solidstorage.json",
     "ccqslt:config/extract-links/actors/quad-pattern-query.json",
     "ccqslt:config/extract-links/actors/solid-type-index-noinference.json"

engines/config-query-sparql-link-traversal/config/extract-links/actors/links-describedby.json

@@ -0,0 +1,18 @@
+{
+  "@context": [
+    "https://linkedsoftwaredependencies.org/bundles/npm/@comunica/runner/^3.0.0/components/context.jsonld",
+
+    "https://linkedsoftwaredependencies.org/bundles/npm/@comunica/actor-extract-links-headers/^0.0.0/components/context.jsonld"
+  ],
+  "@id": "urn:comunica:default:Runner",
+  "@type": "Runner",
+  "actors": [
+    {
+      "@id": "urn:comunica:default:extract-links/actors#links-describedby",
+      "@type": "ActorExtractLinksHeaders",
+      "headersRegexes": [
+        "rel=\"describedby\""
+      ]
+    }
+  ]
+}

engines/query-sparql-link-traversal-solid/package.json

@@ -53,6 +53,7 @@
     "@comunica/actor-context-preprocess-set-defaults-link-traversal": "^0.3.0",
     "@comunica/actor-extract-links-all": "^0.3.0",
     "@comunica/actor-extract-links-content-policies": "^0.3.0",
+    "@comunica/actor-extract-links-headers": "^0.3.0",
     "@comunica/actor-extract-links-predicates": "^0.3.0",
     "@comunica/actor-extract-links-quad-pattern": "^0.3.0",
     "@comunica/actor-extract-links-quad-pattern-query": "^0.3.0",

engines/query-sparql-link-traversal/package.json

@@ -52,6 +52,7 @@
     "@comunica/actor-context-preprocess-set-defaults-link-traversal": "^0.3.0",
     "@comunica/actor-extract-links-all": "^0.3.0",
     "@comunica/actor-extract-links-content-policies": "^0.3.0",
+    "@comunica/actor-extract-links-headers": "^0.3.0",
     "@comunica/actor-extract-links-predicates": "^0.3.0",
     "@comunica/actor-extract-links-quad-pattern": "^0.3.0",
     "@comunica/actor-extract-links-quad-pattern-query": "^0.3.0",

packages/actor-extract-links-headers/README.md

@@ -0,0 +1,43 @@
+# Comunica Headers Extract Links Actor
+
+[![npm version](https://badge.fury.io/js/%40comunica%2Factor-extract-links-headers.svg)](https://www.npmjs.com/package/@comunica/actor-extract-links-headers)
+
+An [RDF Metadata Extract](https://github.com/comunica/comunica/tree/master/packages/bus-extract-links) actor that
+collects all links headers that match with any of the configured regular expressions.
+
+This module is part of the [Comunica framework](https://github.com/comunica/comunica),
+and should only be used by [developers that want to build their own query engine](https://comunica.dev/docs/modify/).
+
+[Click here if you just want to query with Comunica](https://comunica.dev/docs/query/).
+
+## Install
+
+```bash
+$ yarn add @comunica/actor-extract-links-headers
+```
+
+## Configure
+
+After installing, this package can be added to your engine's configuration as follows:
+```text
+{
+  "@context": [
+    ...
+    "https://linkedsoftwaredependencies.org/bundles/npm/@comunica/actor-extract-links-headers/^1.0.0/components/context.jsonld"
+  ],
+  "actors": [
+    ...
+    {
+      "@id": "urn:comunica:default:extract-links/actors#headers",
+      "@type": "ActorExtractLinksHeaders"
+      "headersRegexes": [
+        "rel=\"describedby\""
+      ]
+    }
+  ]
+}
+```
+
+### Config Parameters
+
+* `headersRegexes`: An array of regular expressions that will be matched with incoming link headers.

packages/actor-extract-links-headers/lib/ActorExtractLinksHeaders.ts

@@ -0,0 +1,48 @@
+import type { IActionExtractLinks, IActorExtractLinksOutput } from '@comunica/bus-extract-links';
+import { ActorExtractLinks } from '@comunica/bus-extract-links';
+import type { ILink } from '@comunica/bus-rdf-resolve-hypermedia-links';
+import type { IActorArgs, IActorTest } from '@comunica/core';
+
+/**
+ * A comunica Traverse Predicates RDF Link Header Actor.
+ */
+export class ActorExtractLinksHeaders extends ActorExtractLinks {
+  private readonly headers: RegExp[];
+  private readonly linkRegEx = /<(.*)>/u;
+
+  public constructor(args: IActorExtractLinksTraverseHeadersArgs) {
+    super(args);
+    this.headers = args.headersRegexes.map(stringRegex => new RegExp(stringRegex, 'u'));
+  }
+
+  public async test(_action: IActionExtractLinks): Promise<IActorTest> {
+    return true;
+  }
+
+  public async run(action: IActionExtractLinks): Promise<IActorExtractLinksOutput> {
+    return new Promise((resolve, _reject) => {
+      const headers = action.headers;
+      const links: ILink[] = [];
+
+      for (const regex of this.headers) {
+        const linkHeaders = headers?.get('link')?.split(',');
+        if (linkHeaders) {
+          for (const header of linkHeaders) {
+            if (regex.test(header)) {
+              const match = this.linkRegEx.exec(header);
+              if (match) {
+                links.push({ url: new URL(match[1], action.url).href });
+              }
+            }
+          }
+        }
+      }
+      resolve({ links });
+    });
+  }
+}
+
+export interface IActorExtractLinksTraverseHeadersArgs
+  extends IActorArgs<IActionExtractLinks, IActorTest, IActorExtractLinksOutput> {
+  headersRegexes: string[];
+}

packages/actor-extract-links-headers/lib/index.ts

@@ -0,0 +1 @@
+export * from './ActorExtractLinksHeaders';

packages/actor-extract-links-headers/package.json

@@ -0,0 +1,44 @@
+{
+  "name": "@comunica/actor-extract-links-headers",
+  "version": "0.3.0",
+  "description": "A headers extract-links actor",
+  "lsd:module": true,
+  "license": "MIT",
+  "homepage": "https://comunica.dev/",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/comunica/comunica.git",
+    "directory": "packages/actor-extract-links-headers"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "bugs": {
+    "url": "https://github.com/comunica/comunica/issues"
+  },
+  "keywords": [
+    "comunica",
+    "actor",
+    "extract-links",
+    "headers"
+  ],
+  "sideEffects": false,
+  "main": "lib/index.js",
+  "typings": "lib/index",
+  "files": [
+    "components",
+    "lib/**/*.d.ts",
+    "lib/**/*.js",
+    "lib/**/*.js.map"
+  ],
+  "scripts": {
+    "build": "npm run build:ts && npm run build:components",
+    "build:ts": "node \"../../node_modules/typescript/bin/tsc\"",
+    "build:components": "componentsjs-generator"
+  },
+  "dependencies": {
+    "@comunica/bus-extract-links": "^0.3.0",
+    "@comunica/bus-rdf-resolve-hypermedia-links": "^3.0.1",
+    "@comunica/core": "^3.0.1"
+  }
+}

packages/actor-extract-links-headers/test/ActorExtractLinksHeaders-test.ts

@@ -0,0 +1,50 @@
+import type { Readable } from 'node:stream';
+import { ActionContext, Bus } from '@comunica/core';
+import { ActorExtractLinksHeaders } from '../lib/ActorExtractLinksHeaders';
+
+const quad = require('rdf-quad');
+const stream = require('streamify-array');
+
+describe('ActorExtractLinksHeaders', () => {
+  let bus: any;
+
+  beforeEach(() => {
+    bus = new Bus({ name: 'bus' });
+  });
+
+  describe('An ActorExtractLinksHeaders instance with check subject', () => {
+    let actor: ActorExtractLinksHeaders;
+    let inputMetadata: Readable;
+    let input: Headers;
+
+    beforeEach(() => {
+      actor = new ActorExtractLinksHeaders({
+        name: 'actor',
+        bus,
+        headersRegexes: [
+          'rel="describedby"',
+        ],
+      });
+      inputMetadata = stream([]);
+      input = new Headers();
+      input.append('Content-Type', 'text-turtle');
+      input.append('Location', '/storage/resource');
+      input.append('Content-Length', '1024');
+      input.append('Link', '</storage/resource.meta>;rel="describedby"');
+    });
+
+    it('should test ', async() => {
+      await expect(actor.test({ url: 'http://pod.example.com/storage/resource', metadata: inputMetadata, headers: input, requestTime: 0, context: new ActionContext() }))
+        .resolves.toBe(true);
+    });
+
+    it('should run on headers and return all describedby values', async() => {
+      await expect(actor.run({ url: 'http://pod.example.com/storage/resource', metadata: inputMetadata, headers: input, requestTime: 0, context: new ActionContext() })).resolves
+        .toEqual({
+          links: [
+            { url: 'http://pod.example.com/storage/resource.meta' },
+          ],
+        });
+    });
+  });
+});