Use validate-iri.js to validate IRIs

- By default the "pragmatic" mode is used (scheme validation + no invalid char according to RDF Turtle)
- The iriValidationStrategy option allows to customize the validation strategy

Benchmark using the GeoSpecies Knowledge Base (1.8M triples).
- no IRI validation: 8.338s
- pragmatic IRI validation: 9.116s
- full IRI validation: 12.053s

Author: Tpt

README.md

@@ -90,12 +90,13 @@ myParser.import(myTextStream)
 Optionally, the following parameters can be set in the `RdfXmlParser` constructor:
 
 * `dataFactory`: A custom [RDFJS DataFactory](http://rdf.js.org/#datafactory-interface) to construct terms and triples. _(Default: `require('@rdfjs/data-model')`)_
-* `baseIRI`: An initital default base IRI. _(Default: `''`)_
+* `baseIRI`: An initial default base IRI. _(Default: `''`)_
 * `defaultGraph`: The default graph for constructing [quads](http://rdf.js.org/#dom-datafactory-quad). _(Default: `defaultGraph()`)_
 * `strict`: If the internal SAX parser should parse XML in strict mode, and error if it is invalid. _(Default: `false`)_
 * `trackPosition`: If the internal position (line, column) should be tracked an emitted in error messages. _(Default: `false`)_
 * `allowDuplicateRdfIds`: By default [multiple occurrences of the same `rdf:ID` value are not allowed](https://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-ID-xml-base). By setting this option to `true`, this uniqueness check can be disabled. _(Default: `false`)_
 * `validateUri`: By default, the parser validates each URI. _(Default: `true`)_
+* `iriValidationStrategy`: Allows to customize the used IRI validation strategy using the `IriValidationStrategy` enumeration. _(Default: `IriValidationStrategy.Pragmatic`)_
 
 ```javascript
 new RdfXmlParser({

lib/RdfXmlParser.ts

@@ -2,15 +2,12 @@ import * as RDF from "@rdfjs/types";
 import {resolve} from "relative-to-absolute-iri";
 import {SaxesParser, SaxesTagNS} from "saxes";
 import {PassThrough, Transform} from "readable-stream";
-import EventEmitter = NodeJS.EventEmitter;
 import {ParseError} from "./ParseError";
 import {DataFactory} from "rdf-data-factory";
+import {IriValidationStrategy, validateIri} from "validate-iri";
+import EventEmitter = NodeJS.EventEmitter;
 
 export class RdfXmlParser extends Transform implements RDF.Sink<EventEmitter, RDF.Stream> {
-
-  // Regex for valid IRIs
-  public static readonly IRI_REGEX: RegExp = /^([A-Za-z][A-Za-z0-9+-.]*):[^ "<>{}|\\\[\]`]*$/;
-
   public static readonly MIME_TYPE = 'application/rdf+xml';
 
   public static readonly RDF = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#';
@@ -51,6 +48,7 @@ export class RdfXmlParser extends Transform implements RDF.Sink<EventEmitter, RD
   private readonly allowDuplicateRdfIds?: boolean;
   private readonly saxParser: SaxesParser;
   private readonly validateUri: boolean;
+  private readonly iriValidationStrategy: IriValidationStrategy;
 
   private readonly activeTagStack: IActiveTag[] = [];
   private readonly nodeIds: {[id: string]: boolean} = {};
@@ -74,25 +72,15 @@ export class RdfXmlParser extends Transform implements RDF.Sink<EventEmitter, RD
     if (this.validateUri !== false) {
       this.validateUri = true;
     }
+    if (!this.iriValidationStrategy) {
+      this.iriValidationStrategy = this.validateUri ? IriValidationStrategy.Pragmatic : IriValidationStrategy.None;
+    }
 
     this.saxParser = new SaxesParser({ xmlns: true, position: this.trackPosition });
 
     this.attachSaxListeners();
   }
 
-  /**
-   * Check if the given IRI is valid.
-   * @param {string} iri A potential IRI.
-   * @return {boolean} If the given IRI is valid.
-   */
-  public static isValidIri(iri: string): boolean {
-    return RdfXmlParser.IRI_REGEX.test(iri);
-  }
-
-  get uriValidationEnabled() {
-    return this.validateUri;
-  }
-
   /**
    * Parses the given text stream into a quad stream.
    * @param {NodeJS.EventEmitter} stream A text stream.
@@ -148,8 +136,9 @@ export class RdfXmlParser extends Transform implements RDF.Sink<EventEmitter, RD
    */
   public uriToNamedNode(uri: string): RDF.NamedNode {
     // Validate URI
-    if (this.uriValidationEnabled && !RdfXmlParser.isValidIri(uri)) {
-      throw this.newParseError(`Invalid URI: ${uri}`);
+    const uriValidationResult = validateIri(uri, this.iriValidationStrategy);
+    if (uriValidationResult instanceof Error) {
+      throw this.newParseError(uriValidationResult.message);
     }
     return this.dataFactory.namedNode(uri);
   }
@@ -701,6 +690,11 @@ export interface IRdfXmlParserArgs {
    * By default, it is equal to true.
    */
   validateUri?: boolean;
+  /**
+   * Allows to customize the used IRI validation strategy using the `IriValidationStrategy` enumeration.
+   * By default, the "pragmatic" strategy is used.
+   */
+  iriValidationStrategy?: IriValidationStrategy;
 }
 
 export interface IActiveTag {

package.json

@@ -95,7 +95,8 @@
     "rdf-data-factory": "^1.1.0",
     "relative-to-absolute-iri": "^1.0.0",
     "readable-stream": "^4.0.0",
-    "saxes": "^6.0.0"
+    "saxes": "^6.0.0",
+    "validate-iri": "^1.0.0"
   },
   "sideEffects": false
 }

test/RdfXmlParser-test.ts

@@ -7,6 +7,7 @@ import {DataFactory} from "rdf-data-factory";
 const streamifyString = require('streamify-string');
 const streamifyArray = require('streamify-array');
 import arrayifyStream from 'arrayify-stream';
+import {IriValidationStrategy} from "validate-iri";
 const quad = require('rdf-quad');
 const DF = new DataFactory();
 
@@ -26,7 +27,7 @@ describe('RdfXmlParser', () => {
     expect((<any> instance).defaultGraph).toBe(DF.defaultGraph());
     expect((<any> instance).saxParser).toBeInstanceOf(SaxesParser);
     expect((<any> instance).validateUri).toBeTruthy();
-    expect(instance.uriValidationEnabled).toBeTruthy();
+    expect((<any> instance).iriValidationStrategy).toEqual(IriValidationStrategy.Pragmatic);
   });
 
   it('should be constructable with empty args', () => {
@@ -37,7 +38,7 @@ describe('RdfXmlParser', () => {
     expect((<any> instance).defaultGraph).toBe(DF.defaultGraph());
     expect((<any> instance).saxParser).toBeInstanceOf(SaxesParser);
     expect((<any> instance).validateUri).toBeTruthy();
-    expect(instance.uriValidationEnabled).toBeTruthy();
+    expect((<any> instance).iriValidationStrategy).toEqual(IriValidationStrategy.Pragmatic);
   });
 
   it('should be constructable with args with a custom data factory', () => {
@@ -84,41 +85,7 @@ describe('RdfXmlParser', () => {
     const instance = new RdfXmlParser({ validateUri: false });
     expect(instance).toBeInstanceOf(RdfXmlParser);
     expect((<any> instance).validateUri).toBeFalsy();
-    expect(instance.uriValidationEnabled).toBeFalsy();
-  });
-
-  describe('#isValidIri', () => {
-    it('should be false for null', async () => {
-      expect(RdfXmlParser.isValidIri(null)).toBeFalsy();
-    });
-
-    it('should be false for an empty string', async () => {
-      expect(RdfXmlParser.isValidIri('')).toBeFalsy();
-    });
-
-    it('should be false for an abc', async () => {
-      expect(RdfXmlParser.isValidIri('abc')).toBeFalsy();
-    });
-
-    it('should be true for an abc:def', async () => {
-      expect(RdfXmlParser.isValidIri('abc:def')).toBeTruthy();
-    });
-
-    it('should be true for an http://google.com', async () => {
-      expect(RdfXmlParser.isValidIri('http://google.com')).toBeTruthy();
-    });
-
-    it('should be false for an http://google.com<', async () => {
-      expect(RdfXmlParser.isValidIri('http://google.com<')).toBeFalsy();
-    });
-
-    it('should be false for an http://google .com', async () => {
-      expect(RdfXmlParser.isValidIri('http://google .com')).toBeFalsy();
-    });
-
-    it('should be false for an invalid scheme', async () => {
-      expect(RdfXmlParser.isValidIri('%http://google.com')).toBeFalsy();
-    });
+    expect((<any> instance).iriValidationStrategy).toEqual(IriValidationStrategy.None);
   });
 
   describe('a default instance', () => {
@@ -162,12 +129,12 @@ abc`)).rejects.toBeTruthy();
 
       it('create error on a URI with an invalid scheme', () => {
         expect(() => parser.valueToUri('%https://example.com/', {}))
-          .toThrow(new Error('Invalid URI: %https://example.com/'));
+          .toThrow(new Error('Invalid IRI according to RDF Turtle: \'%https://example.com/\''));
       });
 
       it('create error on a URI with an invalid character', () => {
         expect(() => parser.valueToUri('https://example.com/<', {}))
-          .toThrow(new Error('Invalid URI: https://example.com/<'));
+          .toThrow(new Error('Invalid IRI according to RDF Turtle: \'https://example.com/<\''));
       });
 
       it('create a named node from a relative URI when a baseIRI is given', () => {
@@ -397,7 +364,7 @@ abc`)).rejects.toBeTruthy();
     <ex:prop>1</ex:prop>
   <rdf:Description>
 </rdf:RDF>`)).rejects.toEqual(
-          new Error('Invalid URI: #abc'));
+          new Error('Invalid IRI according to RDF Turtle: \'#abc\''));
       });
 
       // 2.10
@@ -782,7 +749,7 @@ abc`)).rejects.toBeTruthy();
     <ns1:b rdf:resource="_:bnode"/>
   </rdf:Description>
 
-</rdf:RDF>`)).rejects.toEqual(new Error('Invalid URI: _:bnode'));
+</rdf:RDF>`)).rejects.toEqual(new Error('Invalid IRI according to RDF Turtle: \'_:bnode\''));
       });
 
       // Illegal XML name production
@@ -2373,6 +2340,17 @@ abc`)).rejects.toBeTruthy();
         expect(streamParser.read(1)).toBeFalsy();
         expect(streamParser.writable).toBeFalsy();
       });
+
+
+      it('should properly support XML encoded URIs', async () => {
+        expect(parse(parser, `<?xml version="1.0" encoding="utf-8" ?>
+<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
+         xmlns:ns0="b:">
+  <rdf:Description rdf:about="a:&#xA;">
+    <ns0:b rdf:resource="c:c"/>
+  </rdf:Description>
+</rdf:RDF>`)).rejects.toBeTruthy();
+      });
     });
   });
 

yarn.lock

@@ -3981,6 +3981,11 @@ v8-to-istanbul@^9.0.1:
     "@types/istanbul-lib-coverage" "^2.0.1"
     convert-source-map "^1.6.0"
 
+validate-iri@^1.0.0:
+  version "1.0.0"
+  resolved "https://registry.yarnpkg.com/validate-iri/-/validate-iri-1.0.0.tgz#a109600246e8a7515ecd238cdcddb7ca54b95d2d"
+  integrity sha512-4htbVgPOAS8RihVeyp/Pq5bnpKKhA2FcpsYTTION9rejFSZuIUX80SzO/+WMtSR3OVV1NFJx3J6DjrAZw83eCA==
+
 validate-npm-package-license@^3.0.1:
   version "3.0.4"
   resolved "https://registry.yarnpkg.com/validate-npm-package-license/-/validate-npm-package-license-3.0.4.tgz#fc91f6b9c7ba15c857f4cb2c5defeec39d4f410a"