update language definitions

Author: redmer

README.md

@@ -12,7 +12,7 @@ If you're not using a build system and just want to embed this in your webpage:
 
 ```html
 <script type="text/javascript" src="/path/to/highlight.pack.js"></script>
-<script type="text/javascript" src="/path/to/highlightjs-turtle/turtle.js"></script>
+<script type="text/javascript" src="/path/to/highlightjs-turtle/src/languages/turtle.js"></script>
 <script type="text/javascript">
     hljs.registerLanguage('turtle', window.hljsDefineTurtle);
     hljs.initHighlightingOnLoad();
@@ -35,4 +35,8 @@ This is a pretty simple package, the only thing you might want to do differently
 
 ### About the author
 
-The code for Turtle was taken from the non-merged pull request by Mark Ellis, at https://github.com/ellismarkf/highlight.js/blob/c65ea836b3785a5b2ee2078461edaf38c005247d/src/languages/sparql.js
+Redmer KRONEMEIJER is an information modeller. Read more about him at https://rdmr.eu.
+
+The code for Turtle was taken from two unmaintained repositories:
+- Mark Ellis, at https://github.com/ellismarkf/highlight.js
+- Vladimir Alexiev, at https://github.com/VladimirAlexiev/highlight.js

package-lock.json

@@ -1,32 +1,32 @@
 {
-  "name": "highlightjs-solidity",
-  "version": "1.0.19",
-  "description": "highlight.js syntax definition for Ethereum's Solidity language",
-  "main": "solidity.js",
+  "name": "highlightjs-turtle",
+  "version": "1.0.1",
+  "description": "highlight.js syntax definition for the semantic web language Turtle",
+  "main": "src/languages/turtle.js",
   "files": [],
   "scripts": {
-    "test": "mocha"
+    "test": "mocha test/markup/turtle/default.js"
   },
   "repository": {
     "type": "git",
-    "url": "git+ssh://[email protected]/pospi/highlightjs-solidity.git"
+    "url": "git+ssh://[email protected]/redmer/highlightjs-turtle.git"
   },
   "keywords": [
-    "Solidity",
-    "Ethereum",
+    "turtle",
+    "semantic web",
     "highlight.js",
     "highlightjs",
     "syntax"
   ],
-  "author": "pospi",
+  "author": "redmer",
   "license": "MIT",
   "bugs": {
-    "url": "https://github.com/highlightjs/highlightjs-solidity/issues"
+    "url": "https://github.com/redmer/highlightjs-turtle/issues"
   },
-  "homepage": "https://github.com/highlightjs/highlightjs-solidity#readme",
+  "homepage": "https://github.com/redmer/highlightjs-turtle#readme",
   "devDependencies": {
     "highlightjs": "^9.12.0",
     "mocha": "^6.2.1",
     "parse5": "^5.1.0"
   }
-}
+}

package.json

@@ -0,0 +1,121 @@
+/*
+Language: Turtle
+Author: Redmer KRONEMEIJER <[email protected]>
+Contributors: Mark ELLIS <[email protected]>, Vladimir ALEXIEV <[email protected]>
+*/
+
+var module = module ? module : {};     // shim for browser use
+
+function hljsDefineTurtle(hljs) {
+  var KEYWORDS = {
+    keyword: 'BASE|10 PREFIX|10 @base|10 @prefix|10',
+    literal: 'true|0 false|0',
+    built_in: 'a|0'
+  };
+
+  var IRI_LITERAL = {// https://www.w3.org/TR/turtle/#grammar-production-IRIREF
+    className: 'literal',
+    relevance: 1, // XML tags look also like relative IRIs
+    begin: /</,
+    end: />/,
+    illegal: /[^\x00-\x20<>"{}|^`]/, // TODO: https://www.w3.org/TR/turtle/#grammar-production-UCHAR
+  };
+
+  // https://www.w3.org/TR/turtle/#terminals
+  var PN_CHARS_BASE = 'A-Za-z\u00C0-\u00D6\u00D8-\u00F6\u00F8-\u02FF\u0370-\u037D\u037F-\u1FFF\u200C-\u200D\u2070-\u218F\u2C00-\u2FEF\u3001-\uD7FF\uF900-\uFDCF\uFDF0-\uFFFD\u10000-\uEFFFF';
+  var PN_CHARS_U = PN_CHARS_BASE + '_';
+  var PN_CHARS = '-' + PN_CHARS_U + '0-9\u00B7\u0300-\u036F\u203F-\u2040';
+  var BLANK_NODE_LABEL = '_:[' + PN_CHARS_U + '0-9]([' + PN_CHARS + '.]*[' + PN_CHARS + '])?';
+  var PN_PREFIX = '[' + PN_CHARS_BASE + ']([' + PN_CHARS + '.]*[' + PN_CHARS + '])?';
+  var PERCENT = '%[0-9A-Fa-f][0-9A-Fa-f]';
+  var PN_LOCAL_ESC = '\\\\[_~.!$&\'()*+,;=/?#@%-]';
+  var PLX = PERCENT + '|' + PN_LOCAL_ESC;
+  var PNAME_NS = '(' + PN_PREFIX + ')?:';
+  var PN_LOCAL = '([' + PN_CHARS_U + ':0-9]|' + PLX + ')([' + PN_CHARS + '.:]|' + PLX + ')*([' + PN_CHARS + ':]|' + PLX + ')?';
+  var PNAME_LN = PNAME_NS + PN_LOCAL;
+  var PNAME_NS_or_LN = PNAME_NS + '(' + PN_LOCAL + ')?';
+
+  var PNAME = {
+    begin: PNAME_NS_or_LN,
+    relevance: 0,
+    className: 'symbol',
+  };
+
+  var BLANK_NODE = {
+    begin: BLANK_NODE_LABEL,
+    relevance: 10,
+    className: 'template-variable',
+  };
+
+  var LANGTAG = {
+    begin: /@[a-zA-Z]+([a-zA-Z0-9-]+)*/,
+    className: 'type',
+    relevance: 5, // also catches objectivec keywords like: @protocol, @optional
+  };
+
+  var DATATYPE = {
+    begin: '\\^\\^' + PNAME_LN,
+    className: 'type',
+    relevance: 10,
+  };
+
+  var TRIPLE_APOS_STRING = {
+    begin: /'''/,
+    end: /'''/,
+    className: 'string',
+    relevance: 0,
+  };
+
+  var TRIPLE_QUOTE_STRING = {
+    begin: /"""/,
+    end: /"""/,
+    className: 'string',
+    relevance: 0,
+  };
+
+  var APOS_STRING_LITERAL = JSON.parse(JSON.stringify(hljs.APOS_STRING_MODE));
+  APOS_STRING_LITERAL.relevance = 0;
+
+  var QUOTE_STRING_LITERAL = JSON.parse(JSON.stringify(hljs.QUOTE_STRING_MODE));
+  QUOTE_STRING_LITERAL.relevance = 0;
+
+  var NUMBER = JSON.parse(JSON.stringify(hljs.C_NUMBER_MODE));
+  NUMBER.relevance = 0;
+
+  return {
+    name: "Turtle",
+    case_insensitive: false, // was: true, but PREFIX and @prefix are oblig. cased thus
+    keywords: KEYWORDS,
+    aliases: ['turtle', 'ttl', 'n3'],
+    contains: [
+      LANGTAG,
+      DATATYPE,
+      IRI_LITERAL,
+      BLANK_NODE,
+      PNAME,
+      TRIPLE_APOS_STRING, TRIPLE_QUOTE_STRING, // order matters
+      APOS_STRING_LITERAL, QUOTE_STRING_LITERAL,
+      NUMBER,
+      hljs.HASH_COMMENT_MODE,
+    ],
+    exports: {
+      LANGTAG: LANGTAG,
+      DATATYPE: DATATYPE,
+      IRI_LITERAL: IRI_LITERAL,
+      BLANK_NODE: BLANK_NODE,
+      PNAME: PNAME,
+      TRIPLE_APOS_STRING: TRIPLE_APOS_STRING,
+      TRIPLE_QUOTE_STRING: TRIPLE_QUOTE_STRING,
+      APOS_STRING_LITERAL: APOS_STRING_LITERAL,
+      QUOTE_STRING_LITERAL: QUOTE_STRING_LITERAL,
+      NUMBER: NUMBER,
+      KEYWORDS: KEYWORDS,
+    }
+  };
+}
+
+module.exports = function (hljs) {
+  hljs.registerLanguage('turtle', hljsDefineTurtle);
+};
+
+module.exports.definer = hljsDefineTurtle;

src/languages/turtle.js

@@ -0,0 +1,26 @@
+@base <http://example.org/> .
+@prefix : <http://example.org/>.
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+@prefix foaf: <http://xmlns.com/foaf/0.1/> .
+@prefix rel: <http://www.perceive.net/schemas/relationship/> .
+@prefix ff-map:  <http://factforge.net/ff2016-mapping/>.
+
+<#green-goblin>
+  rel:enemyOf <#spiderman> ;
+  a foaf:Person ; # in the context of the Marvel universe
+  foaf:description """Long TRIPLE_QUOTE_STRING with ''' embedded
+apostrophes '''""",
+    '''The librarian said, "Perhaps you would enjoy 'War and Peace'."'''.
+
+<#spiderman> rel:enemyOf <#green-goblin> ;
+  foaf:name "Spiderman"@en, "Человек-паук"@ru ;
+  foaf:age "23"^^xsd:integer, 23;
+  foaf:height 1.68, 0.168e1, "1.68"^^xsd:decimal, true, false;
+  foaf:mbox  <mailto:[email protected]>;
+  rdfs:label 'chat'@fr, 
+
+_:blank1 fb:object.type fb:object.type.car. # trailing dot not part of symbol
+: rel:weird :bar, rel: ; rel:weird rel:.
+
+</doc#char=10,20> ff-map:mentionsEntity </entity>.

test.js

@@ -0,0 +1,61 @@
+const assert = require('assert');
+const parse5 = require('parse5');
+
+const hljs = require('highlightjs');
+const defineTurtle = require('../../../src/languages/turtle');
+
+defineTurtle(hljs);
+
+// Receives a Turtle snippet and returns an array of [type, text] tuples.
+// Type is the detected token type, and text the corresponding source text.
+function getTokens(source) {
+  const { value } = hljs.highlight('turtle', source);
+  const frag = parse5.parseFragment(value);
+
+  return frag.childNodes.map(function (node) {
+    if (node.nodeName === '#text') {
+      return ['none', node.value];
+    } else {
+      const type = node.attrs.find(a => a.name === 'class').value.replace(/^hljs-/, '');
+      assert(
+        node.childNodes.length === 1 && node.childNodes[0].nodeName === '#text',
+        'Unexpected nested tags',
+      );
+      return [type, node.childNodes[0].value];
+    }
+  });
+}
+
+// Taken from the Solidity repo.
+it('numbers', function () {
+  const ok = [
+    '-5',
+    '-5.0',
+    '4.2E9',
+  ];
+
+  const fail = [
+    '1-5',
+    '5,0',
+  ];
+
+  for (const n of ok) {
+    assert.deepStrictEqual(getTokens(n), [['number', n]]);
+  }
+
+  for (const n of fail) {
+    assert.notDeepStrictEqual(getTokens(n), [['number', n]]);
+  }
+});
+
+// it('namespaces', function () {
+//   assert.deepStrictEqual(getTokens(':local'), [['symbol', ':local']]);
+//   assert.deepStrictEqual(getTokens('_:local'), [['template-variable', '_:local']]);
+//   assert.deepStrictEqual(getTokens('ns1:hello'), [['symbol', 'ns1:hello']]);
+//   assert.notDeepStrictEqual(getTokens('ns1:ns2:wrong'), [['symbol', 'ns1:ns2:wrong']]);
+// });
+
+it('strings', function () {
+  assert.deepStrictEqual(getTokens('"hello"@en'), [['string', '"hello"'], ['type', '@en']]);
+  assert.deepStrictEqual(getTokens('"daar"@nl-NL'), [['string', '"daar"'], ['type', '@nl-NL']]);
+});