Perf/better quad ids

README.md

@@ -12,14 +12,14 @@ It offers:
   [TriG](https://www.w3.org/TR/trig/),
   [N-Triples](https://www.w3.org/TR/n-triples/),
   [N-Quads](https://www.w3.org/TR/n-quads/),
-  [RDF*](https://blog.liu.se/olafhartig/2019/01/10/position-statement-rdf-star-and-sparql-star/)
+  [RDF-star](https://www.w3.org/2021/12/rdf-star.html)
   and [Notation3 (N3)](https://www.w3.org/TeamSubmission/n3/)
 - [**Writing**](#writing) triples/quads to
   [Turtle](https://www.w3.org/TR/turtle/),
   [TriG](https://www.w3.org/TR/trig/),
   [N-Triples](https://www.w3.org/TR/n-triples/),
   [N-Quads](https://www.w3.org/TR/n-quads/)
-  and [RDF*](https://blog.liu.se/olafhartig/2019/01/10/position-statement-rdf-star-and-sparql-star/)
+  and [RDF-star](https://www.w3.org/2021/12/rdf-star.html)
 - [**Storage**](#storing) of triples/quads in memory
 
 Parsing and writing is:
@@ -358,16 +358,16 @@ The N3.js parser and writer is fully compatible with the following W3C specifica
 
 In addition, the N3.js parser also supports [Notation3 (N3)](https://www.w3.org/TeamSubmission/n3/) (no official specification yet).
 
-The N3.js parser and writer are also fully compatible with the RDF* variants
+The N3.js parser and writer are also fully compatible with the RDF-star variants
 of the W3C specifications.
 
 The default mode is permissive
-and allows a mixture of different syntaxes, including RDF*.
+and allows a mixture of different syntaxes, including RDF-star.
 Pass a `format` option to the constructor with the name or MIME type of a format
 for strict, fault-intolerant behavior.
 If a format string contains `star` or `*`
 (e.g., `turtlestar` or `TriG*`),
-RDF* support for that format will be enabled.
+RDF-star support for that format will be enabled.
 
 ### Interface specifications
 The N3.js submodules are compatible with the following [RDF.js](http://rdf.js.org) interfaces:

package.json

@@ -42,7 +42,7 @@
     "mocha": "^8.0.0",
     "nyc": "^14.1.1",
     "pre-commit": "^1.2.2",
-    "rdf-test-suite": "^1.19.2",
+    "rdf-test-suite": "^1.20.0",
     "streamify-string": "^1.0.1",
     "uglify-js": "^3.14.3"
   },
@@ -54,7 +54,7 @@
     "mocha": "mocha",
     "lint": "eslint src perf test spec",
     "prepare": "npm run build",
-    "spec": "npm run spec-turtle && npm run spec-ntriples && npm run spec-nquads && npm run spec-trig",
+    "spec": "npm run spec-turtle && npm run spec-ntriples && npm run spec-nquads && npm run spec-trig && npm run spec-rdf-star",
     "spec-earl": "npm run spec-earl-turtle && npm run spec-earl-ntriples && npm run spec-earl-nquads && npm run spec-earl-trig",
     "spec-ntriples": "rdf-test-suite spec/parser.js http://w3c.github.io/rdf-tests/ntriples/manifest.ttl -i '{ \"format\": \"n-triples\" }' -c .rdf-test-suite-cache/",
     "spec-nquads": "rdf-test-suite spec/parser.js http://w3c.github.io/rdf-tests/nquads/manifest.ttl -i '{ \"format\": \"n-quads\" }' -c .rdf-test-suite-cache/",
@@ -64,6 +64,12 @@
     "spec-earl-nquads": "rdf-test-suite spec/parser.js http://w3c.github.io/rdf-tests/nquads/manifest.ttl -i '{ \"format\": \"n-quads\" }' -c .rdf-test-suite-cache/ -o earl -p spec/earl-meta.json > spec/earl-nquads.ttl",
     "spec-earl-turtle": "rdf-test-suite spec/parser.js http://w3c.github.io/rdf-tests/turtle/manifest.ttl -i '{ \"format\": \"turtle\" }' -c .rdf-test-suite-cache/ -o earl -p spec/earl-meta.json > spec/earl-turtle.ttl",
     "spec-earl-trig": "rdf-test-suite spec/parser.js http://w3c.github.io/rdf-tests/trig/manifest.ttl -i '{ \"format\": \"trig\" }' -c .rdf-test-suite-cache/ -o earl -p spec/earl-meta.json > spec/earl-trig.ttl",
+    "spec-rdf-star": "npm run spec-trig-rdf-star && npm run spec-trig-eval-rdf-star && npm run spec-turtle-rdf-star && npm run spec-turtle-eval-rdf-star",
+    "spec-trig-rdf-star": "node ../rdf-test-suite.js/bin/Runner.js spec/parser.js https://w3c.github.io/rdf-star/tests/trig/syntax/manifest.jsonld -i '{ \"format\": \"trig-star\" }' -c .rdf-test-suite-cache/",
+    "spec-trig-eval-rdf-star": "node ../rdf-test-suite.js/bin/Runner.js spec/parser.js https://w3c.github.io/rdf-star/tests/trig/eval/manifest.jsonld -i '{ \"format\": \"trig-star\" }' -c .rdf-test-suite-cache/",
+    "spec-turtle-rdf-star": "node ../rdf-test-suite.js/bin/Runner.js spec/parser.js https://w3c.github.io/rdf-star/tests/turtle/syntax/manifest.jsonld -i '{ \"format\": \"turtle-star\" }' -c .rdf-test-suite-cache/",
+    "spec-turtle-eval-rdf-star": "node ../rdf-test-suite.js/bin/Runner.js spec/parser.js https://w3c.github.io/rdf-star/tests/turtle/eval/manifest.jsonld -i '{ \"format\": \"turtle-star\" }' -c .rdf-test-suite-cache/",
+    "spec-ntriples-rdf-star": "node ../rdf-test-suite.js/bin/Runner.js spec/parser.js https://w3c.github.io/rdf-star/tests/nt/syntax/manifest.jsonld -i '{ \"format\": \"n-quads-star\" }' -c .rdf-test-suite-cache/",
     "spec-clean": "rm -r .rdf-test-suite-cache/",
     "docs": "cd src && docco *.js -o ../docs && cd .."
   },

src/N3DataFactory.js

@@ -10,7 +10,6 @@ let DEFAULTGRAPH;
 let _blankNodeCounter = 0;
 
 const escapedLiteral = /^"(.*".*)(?="[^"]*$)/;
-const quadId = /^<<("(?:""|[^"])*"[^ ]*|[^ ]+) ("(?:""|[^"])*"[^ ]*|[^ ]+) ("(?:""|[^"])*"[^ ]*|[^ ]+) ?("(?:""|[^"])*"[^ ]*|[^ ]+)?>>$/;
 
 // ## DataFactory singleton
 const DataFactory = {
@@ -188,9 +187,12 @@ export class DefaultGraph extends Term {
 // ## DefaultGraph singleton
 DEFAULTGRAPH = new DefaultGraph();
 
-
 // ### Constructs a term from the given internal string ID
-export function termFromId(id, factory) {
+// The third 'nested' parameter of this function is to aid
+// with recursion over nested terms. It should not be used
+// by consumers of this library.
+// See https://github.com/rdfjs/N3.js/pull/311#discussion_r1061042725
+export function termFromId(id, factory, nested) {
   factory = factory || DataFactory;
 
   // Falsy value or empty string indicate the default graph
@@ -215,21 +217,28 @@ export function termFromId(id, factory) {
     return factory.literal(id.substr(1, endPos - 1),
             id[endPos + 1] === '@' ? id.substr(endPos + 2)
                                    : factory.namedNode(id.substr(endPos + 3)));
-  case '<':
-    const components = quadId.exec(id);
-    return factory.quad(
-      termFromId(unescapeQuotes(components[1]), factory),
-      termFromId(unescapeQuotes(components[2]), factory),
-      termFromId(unescapeQuotes(components[3]), factory),
-      components[4] && termFromId(unescapeQuotes(components[4]), factory)
-    );
+  case '[':
+    id = JSON.parse(id);
+    break;
   default:
-    return factory.namedNode(id);
+    if (!nested || !Array.isArray(id)) {
+      return factory.namedNode(id);
+    }
   }
+  return factory.quad(
+    termFromId(id[0], factory, true),
+    termFromId(id[1], factory, true),
+    termFromId(id[2], factory, true),
+    id[3] && termFromId(id[3], factory, true)
+  );
 }
 
 // ### Constructs an internal string ID from the given term or ID string
-export function termToId(term) {
+// The third 'nested' parameter of this function is to aid
+// with recursion over nested terms. It should not be used
+// by consumers of this library.
+// See https://github.com/rdfjs/N3.js/pull/311#discussion_r1061042725
+export function termToId(term, nested) {
   if (typeof term === 'string')
     return term;
   if (term instanceof Term && term.termType !== 'Quad')
@@ -247,17 +256,15 @@ export function termToId(term) {
     term.language ? `@${term.language}` :
       (term.datatype && term.datatype.value !== xsd.string ? `^^${term.datatype.value}` : '')}`;
   case 'Quad':
-    // To identify RDF* quad components, we escape quotes by doubling them.
-    // This avoids the overhead of backslash parsing of Turtle-like syntaxes.
-    return `<<${
-        escapeQuotes(termToId(term.subject))
-      } ${
-        escapeQuotes(termToId(term.predicate))
-      } ${
-        escapeQuotes(termToId(term.object))
-      }${
-        (isDefaultGraph(term.graph)) ? '' : ` ${termToId(term.graph)}`
-      }>>`;
+    const res = [
+      termToId(term.subject, true),
+      termToId(term.predicate, true),
+      termToId(term.object, true),
+    ];
+    if (!isDefaultGraph(term.graph)) {
+      res.push(termToId(term.graph, true));
+    }
+    return nested ? res : JSON.stringify(res);
   default: throw new Error(`Unexpected termType: ${term.termType}`);
   }
 }

src/N3Lexer.js

@@ -294,20 +294,33 @@ export default class N3Lexer {
       case '!':
         if (!this._n3Mode)
           break;
+      case '{':
+        // Note the input[0] === '{' is required as this could be a fall-through from the above case
+        if (input.length > 1 && input[0] === '{' && input[1] === '|') {
+          type = '{|', matchLength = 2;
+          break;
+        }
       case ',':
       case ';':
       case '[':
       case ']':
       case '(':
       case ')':
-      case '{':
       case '}':
-        if (!this._lineMode) {
+        if (
+          !this._lineMode &&
+          // The token might actually be {| and we just have not encountered the pipe yet
+          (input !== '{' || input.length > 1)
+          ) {
           matchLength = 1;
           type = firstChar;
         }
         break;
-
+      case '|':
+        if (input.length > 1 && input[1] === '}') {
+          type = '|}', matchLength = 2;
+          break;
+        }
       default:
         inconclusive = true;
       }

src/N3Parser.js

@@ -197,6 +197,10 @@ export default class N3Parser {
                         this._subject = this._blankNode(), null, null);
       return this._readBlankNodeHead;
     case '(':
+      // Lists are not allowed inside quoted triples
+      if (this._contextStack.length > 0 && this._contextStack[this._contextStack.length - 1].type === '<<') {
+        return this._error('Unexpected list inside quoted triple', token);
+      }
       // Start a new list
       this._saveContext('list', this._graph, this.RDF_NIL, null, null);
       this._subject = null;
@@ -239,7 +243,7 @@ export default class N3Parser {
       break;
     case '<<':
       if (!this._supportsRDFStar)
-        return this._error('Unexpected RDF* syntax', token);
+        return this._error('Unexpected RDF-star syntax', token);
       this._saveContext('<<', this._graph, null, null, null);
       this._graph = null;
       return this._readSubject;
@@ -315,6 +319,10 @@ export default class N3Parser {
                         this._subject = this._blankNode());
       return this._readBlankNodeHead;
     case '(':
+      // Lists are not allowed inside quoted triples
+      if (this._contextStack.length > 0 && this._contextStack[this._contextStack.length - 1].type === '<<') {
+        return this._error('Unexpected list inside quoted triple', token);
+      }
       // Start a new list
       this._saveContext('list', this._graph, this._subject, this._predicate, this.RDF_NIL);
       this._subject = null;
@@ -328,7 +336,7 @@ export default class N3Parser {
       return this._readSubject;
     case '<<':
       if (!this._supportsRDFStar)
-        return this._error('Unexpected RDF* syntax', token);
+        return this._error('Unexpected RDF-star syntax', token);
       this._saveContext('<<', this._graph, this._subject, this._predicate, null);
       this._graph = null;
       return this._readSubject;
@@ -363,6 +371,9 @@ export default class N3Parser {
       this._subject = null;
       return this._readBlankNodeTail(token);
     }
+    else if (this._contextStack.length > 1 && this._contextStack[this._contextStack.length - 2].type === '<<') {
+      return this._error('Compound blank node expressions not permitted within quoted triple', token);
+    }
     else {
       this._predicate = null;
       return this._readPredicate(token);
@@ -473,6 +484,16 @@ export default class N3Parser {
       this._saveContext('formula', this._graph, this._subject, this._predicate,
                         this._graph = this._blankNode());
       return this._readSubject;
+    case '<<':
+      if (!this._supportsRDFStar)
+        return this._error('Unexpected RDF-star syntax', token);
+
+      this._saveContext('<<', this._graph, this._subject, null, null);
+      return this._readSubject;
+    case '>>':
+      item = this._graph;
+      this._graph = null;
+      break;
     default:
       if ((item = this._readEntity(token)) === undefined)
         return;
@@ -614,6 +635,18 @@ export default class N3Parser {
     case ',':
       next = this._readObject;
       break;
+    case '{|':
+      if (!this._supportsRDFStar)
+        return this._error('Unexpected RDF-star syntax', token);
+
+      this._saveContext('{|', this._graph, this._subject, this._predicate, this._object);
+
+      // As a convention we use set the graph term as the Default Graph in quads representing quoted triples
+      // see https://github.com/rdfjs/N3.js/pull/311#discussion_r1061039556 for details
+      this._subject = this._quad(this._subject, this._predicate, this._object, this.DEFAULTGRAPH);
+      this._predicate = null;
+      this._object = null;
+      return this._readPredicate;
     default:
       // An entity means this is a quad (only allowed if not already inside a graph)
       if (this._supportsQuads && this._graph === null && (graph = this._readEntity(token)) !== undefined) {
@@ -835,25 +868,21 @@ export default class N3Parser {
     return this._readPath;
   }
 
-  // ### `_readRDFStarTailOrGraph` reads the graph of a nested RDF* quad or the end of a nested RDF* triple
-  _readRDFStarTailOrGraph(token) {
-    if (token.type !== '>>') {
-      // An entity means this is a quad (only allowed if not already inside a graph)
-      if (this._supportsQuads && this._graph === null && (this._graph = this._readEntity(token)) !== undefined)
-        return this._readRDFStarTail;
-      return this._error(`Expected >> to follow "${this._object.id}"`, token);
-    }
-    return this._readRDFStarTail(token);
-  }
-
-  // ### `_readRDFStarTail` reads the end of a nested RDF* triple
+  // ### `_readRDFStarTail` reads the end of a nested RDF-star triple
   _readRDFStarTail(token) {
     if (token.type !== '>>')
-      return this._error(`Expected >> but got ${token.type}`, token);
+      return this._error(`Expected >> to follow "${this._object.id}" but got ${token.type}`, token);
     // Read the quad and restore the previous context
     const quad = this._quad(this._subject, this._predicate, this._object,
       this._graph || this.DEFAULTGRAPH);
     this._restoreContext('<<', token);
+
+    // If the triple is in a list then return to reading the remaining elements
+    if (this._contextStack.length > 0 && this._contextStack[this._contextStack.length - 1].type === 'list') {
+      this._graph = quad;
+      return this._readListItem(token);
+    }
+
     // If the triple was the subject, continue by reading the predicate.
     if (this._subject === null) {
       this._subject = quad;
@@ -866,6 +895,29 @@ export default class N3Parser {
     }
   }
 
+    // ### `_readRDFStarTail` reads the end of a nested RDF-star triple
+  _readAnnotatedTail(token) {
+    if (token.type === '{|') {
+      this._saveContext('{|', this._graph, this._subject, this._predicate, this._object);
+
+      // As a convention we use set the graph term as the Default Graph in quads representing quoted triples
+      // see https://github.com/rdfjs/N3.js/pull/311#discussion_r1061039556 for details
+      this._subject = this._quad(this._subject, this._predicate, this._object, this.DEFAULTGRAPH);
+      this._predicate = null;
+      this._object = null;
+      return this._readPredicate;
+    }
+    else {
+      this._emit(this._subject, this._predicate, this._object, this._graph);
+    }
+
+    // If the quoted triple is not finished, the next token must be a predicate
+    if (token.type !== '|}')
+      return this._readPredicate;
+    this._restoreContext('{|', token);
+    return this._getContextEndReader();
+  }
+
   // ### `_getContextEndReader` gets the next reader function at the end of a context
   _getContextEndReader() {
     const contextStack = this._contextStack;
@@ -880,7 +932,9 @@ export default class N3Parser {
     case 'formula':
       return this._readFormulaTail;
     case '<<':
-      return this._readRDFStarTailOrGraph;
+      return this._readRDFStarTail;
+    case '{|':
+      return this._readAnnotatedTail;
     }
   }