node-html-parser 4.0.0 → 4.1.3

package/README.md

@@ -249,3 +249,7 @@ Get all attributes of current element. **Notice: do not try to change the return
 ### HTMLElement#classList
 
 Get all attributes of current element. **Notice: do not try to change the returned value.**
+
+### HTMLElement#range
+
+Corresponding source code start and end indexes (ie [ 0, 40 ])

package/dist/esm/nodes/comment.js

@@ -1,8 +1,8 @@
 import Node from './node';
 import NodeType from './type';
 export default class CommentNode extends Node {
-    constructor(rawText, parentNode) {
-        super(parentNode);
+    constructor(rawText, parentNode, range) {
+        super(parentNode, range);
         this.rawText = rawText;
         /**
          * Node Type declaration.

package/dist/esm/nodes/html.js

@@ -151,8 +151,8 @@ export default class HTMLElement extends Node {
      *
      * @memberof HTMLElement
      */
-    constructor(tagName, keyAttrs, rawAttrs = '', parentNode) {
-        super(parentNode);
+    constructor(tagName, keyAttrs, rawAttrs = '', parentNode, range) {
+        super(parentNode, range);
         this.rawAttrs = rawAttrs;
         /**
          * Node Type declaration.
@@ -946,39 +946,47 @@ export function base_parse(data, options = { lowerCaseTagName: false, comment: f
             return it.test(tag);
         });
     }
-    const root = new HTMLElement(null, {}, '', null);
+    const createRange = (startPos, endPos) => [startPos - frameFlagOffset, endPos - frameFlagOffset];
+    const root = new HTMLElement(null, {}, '', null, [0, data.length]);
     let currentParent = root;
     const stack = [root];
     let lastTextPos = -1;
     let match;
     // https://github.com/taoqf/node-html-parser/issues/38
     data = `<${frameflag}>${data}</${frameflag}>`;
+    const dataEndPos = data.length - (frameflag.length + 2);
+    const frameFlagOffset = frameflag.length + 2;
     while ((match = kMarkupPattern.exec(data))) {
+        const tagStartPos = kMarkupPattern.lastIndex - match[0].length;
+        const tagEndPos = kMarkupPattern.lastIndex;
+        // Add TextNode if content
         if (lastTextPos > -1) {
-            if (lastTextPos + match[0].length < kMarkupPattern.lastIndex) {
-                // if has content
-                const text = data.substring(lastTextPos, kMarkupPattern.lastIndex - match[0].length);
-                currentParent.appendChild(new TextNode(text, currentParent));
+            if (lastTextPos + match[0].length < tagEndPos) {
+                const text = data.substring(lastTextPos, tagStartPos);
+                currentParent.appendChild(new TextNode(text, currentParent, createRange(lastTextPos, tagStartPos)));
             }
         }
         lastTextPos = kMarkupPattern.lastIndex;
-        if (match[2] === frameflag) {
+        // https://github.com/taoqf/node-html-parser/issues/38
+        // Skip frameflag node
+        if (match[2] === frameflag)
             continue;
-        }
+        // Handle comments
         if (match[0][1] === '!') {
-            // this is a comment
             if (options.comment) {
                 // Only keep what is in between <!-- and -->
-                const text = data.substring(lastTextPos - 3, lastTextPos - match[0].length + 4);
-                currentParent.appendChild(new CommentNode(text, currentParent));
+                const text = data.substring(tagStartPos + 4, tagEndPos - 3);
+                currentParent.appendChild(new CommentNode(text, currentParent, createRange(tagStartPos, tagEndPos)));
             }
             continue;
         }
-        if (options.lowerCaseTagName) {
+        /* -- Handle tag matching -- */
+        // Fix tag casing if necessary
+        if (options.lowerCaseTagName)
             match[2] = match[2].toLowerCase();
-        }
+        // Handle opening tags (ie. <this> not </that>)
         if (!match[1]) {
-            // not </ tags
+            /* Populate attributes */
             const attrs = {};
             for (let attMatch; (attMatch = kAttributePattern.exec(match[3]));) {
                 attrs[attMatch[2].toLowerCase()] = attMatch[4] || attMatch[5] || attMatch[6];
@@ -990,45 +998,41 @@ export function base_parse(data, options = { lowerCaseTagName: false, comment: f
                     currentParent = arr_back(stack);
                 }
             }
-            // ignore container tag we add above
-            // https://github.com/taoqf/node-html-parser/issues/38
-            currentParent = currentParent.appendChild(new HTMLElement(match[2], attrs, match[3], null));
+            const tagEndPos = kMarkupPattern.lastIndex;
+            const tagStartPos = tagEndPos - match[0].length;
+            currentParent = currentParent.appendChild(
+            // Initialize range (end position updated later for closed tags)
+            new HTMLElement(match[2], attrs, match[3], null, createRange(tagStartPos, tagEndPos)));
             stack.push(currentParent);
             if (is_block_text_element(match[2])) {
-                // a little test to find next </script> or </style> ...
+                // Find closing tag
                 const closeMarkup = `</${match[2]}>`;
-                const index = (() => {
-                    if (options.lowerCaseTagName) {
-                        return data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex);
-                    }
-                    return data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
-                })();
+                const closeIndex = options.lowerCaseTagName
+                    ? data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex)
+                    : data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
+                const textEndPos = closeIndex === -1 ? dataEndPos : closeIndex;
                 if (element_should_be_ignore(match[2])) {
-                    let text;
-                    if (index === -1) {
-                        // there is no matching ending for the text element.
-                        text = data.substr(kMarkupPattern.lastIndex);
-                    }
-                    else {
-                        text = data.substring(kMarkupPattern.lastIndex, index);
-                    }
-                    if (text.length > 0) {
-                        currentParent.appendChild(new TextNode(text, currentParent));
+                    const text = data.substring(tagEndPos, textEndPos);
+                    if (text.length > 0 && /\S/.test(text)) {
+                        currentParent.appendChild(new TextNode(text, currentParent, createRange(tagEndPos, textEndPos)));
                     }
                 }
-                if (index === -1) {
+                if (closeIndex === -1) {
                     lastTextPos = kMarkupPattern.lastIndex = data.length + 1;
                 }
                 else {
-                    lastTextPos = kMarkupPattern.lastIndex = index + closeMarkup.length;
+                    lastTextPos = kMarkupPattern.lastIndex = closeIndex + closeMarkup.length;
+                    // Cause to be treated as self-closing, because no close found
                     match[1] = 'true';
                 }
             }
         }
+        // Handle closing tags or self-closed elements (ie </tag> or <br>)
         if (match[1] || match[4] || kSelfClosingElements[match[2]]) {
-            // </ or /> or <br> etc.
             while (true) {
                 if (currentParent.rawTagName === match[2]) {
+                    // Update range end for closed tag
+                    currentParent.range[1] = createRange(-1, Math.max(lastTextPos, tagEndPos))[1];
                     stack.pop();
                     currentParent = arr_back(stack);
                     break;

package/dist/esm/nodes/node.js

@@ -3,9 +3,15 @@ import { decode, encode } from 'he';
  * Node Class as base class for TextNode and HTMLElement.
  */
 export default class Node {
-    constructor(parentNode = null) {
+    constructor(parentNode = null, range) {
         this.parentNode = parentNode;
         this.childNodes = [];
+        Object.defineProperty(this, 'range', {
+            enumerable: false,
+            writable: true,
+            configurable: true,
+            value: range !== null && range !== void 0 ? range : [-1, -1]
+        });
     }
     get innerText() {
         return this.rawText;

package/dist/esm/nodes/text.js

@@ -6,8 +6,8 @@ import NodeType from './type';
  * @param {string} value [description]
  */
 export default class TextNode extends Node {
-    constructor(rawText, parentNode) {
-        super(parentNode);
+    constructor(rawText, parentNode, range) {
+        super(parentNode, range);
         /**
          * Node Type declaration.
          * @type {Number}

package/dist/main.js

@@ -27,10 +27,14 @@ var __assign = (this && this.__assign) || function () {
     };
     return __assign.apply(this, arguments);
 };
-var __spreadArray = (this && this.__spreadArray) || function (to, from) {
-    for (var i = 0, il = from.length, j = to.length; i < il; i++, j++)
-        to[j] = from[i];
-    return to;
+var __spreadArray = (this && this.__spreadArray) || function (to, from, pack) {
+    if (pack || arguments.length === 2) for (var i = 0, l = from.length, ar; i < l; i++) {
+        if (ar || !(i in from)) {
+            if (!ar) ar = Array.prototype.slice.call(from, 0, i);
+            ar[i] = from[i];
+        }
+    }
+    return to.concat(ar || from);
 };
 define("back", ["require", "exports"], function (require, exports) {
     "use strict";
@@ -62,8 +66,8 @@ define("nodes/text", ["require", "exports", "he", "nodes/node", "nodes/type"], f
      */
     var TextNode = /** @class */ (function (_super) {
         __extends(TextNode, _super);
-        function TextNode(rawText, parentNode) {
-            var _this = _super.call(this, parentNode) || this;
+        function TextNode(rawText, parentNode, range) {
+            var _this = _super.call(this, parentNode, range) || this;
             /**
              * Node Type declaration.
              * @type {Number}
@@ -119,7 +123,7 @@ define("nodes/text", ["require", "exports", "he", "nodes/node", "nodes/type"], f
              * @return {string} text content
              */
             get: function () {
-                return he_1.decode(this.rawText);
+                return (0, he_1.decode)(this.rawText);
             },
             enumerable: false,
             configurable: true
@@ -446,9 +450,9 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
          *
          * @memberof HTMLElement
          */
-        function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode) {
+        function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range) {
             if (rawAttrs === void 0) { rawAttrs = ''; }
-            var _this = _super.call(this, parentNode) || this;
+            var _this = _super.call(this, parentNode, range) || this;
             _this.rawAttrs = rawAttrs;
             /**
              * Node Type declaration.
@@ -680,7 +684,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
             });
-            this.parentNode.childNodes = __spreadArray(__spreadArray(__spreadArray([], this.parentNode.childNodes.slice(0, idx)), content), this.parentNode.childNodes.slice(idx + 1));
+            this.parentNode.childNodes = __spreadArray(__spreadArray(__spreadArray([], this.parentNode.childNodes.slice(0, idx), true), content, true), this.parentNode.childNodes.slice(idx + 1), true);
         };
         Object.defineProperty(HTMLElement.prototype, "outerHTML", {
             get: function () {
@@ -773,7 +777,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
          * @return {HTMLElement[]}  matching elements
          */
         HTMLElement.prototype.querySelectorAll = function (selector) {
-            return css_select_1.selectAll(selector, this, {
+            return (0, css_select_1.selectAll)(selector, this, {
                 xmlMode: true,
                 adapter: matcher_1.default
             });
@@ -841,7 +845,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
          * @return {HTMLElement}    matching node
          */
         HTMLElement.prototype.querySelector = function (selector) {
-            return css_select_1.selectOne(selector, this, {
+            return (0, css_select_1.selectOne)(selector, this, {
                 xmlMode: true,
                 adapter: matcher_1.default
             });
@@ -914,17 +918,14 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             }
             el = this;
             while (el) {
-                var e = css_select_1.selectOne(selector, el, {
+                var e = (0, css_select_1.selectOne)(selector, el, {
                     xmlMode: true,
                     adapter: __assign(__assign({}, matcher_1.default), { getChildren: function (node) {
                             var child = mapChild.get(node);
                             return child && [child];
-                        },
-                        getSiblings: function (node) {
+                        }, getSiblings: function (node) {
                             return [node];
-                        },
-                        findOne: findOne,
-                        findAll: function () {
+                        }, findOne: findOne, findAll: function () {
                             return [];
                         } })
                 });
@@ -963,7 +964,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
              * @return {Node} last child node
              */
             get: function () {
-                return back_1.default(this.childNodes);
+                return (0, back_1.default)(this.childNodes);
             },
             enumerable: false,
             configurable: true
@@ -1125,7 +1126,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 var idx = this.parentNode.childNodes.findIndex(function (child) {
                     return child === _this;
                 });
-                (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes));
+                (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes, false));
                 p.childNodes.forEach(function (n) {
                     if (n instanceof HTMLElement) {
                         n.parentNode = _this.parentNode;
@@ -1144,7 +1145,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 var idx = this.parentNode.childNodes.findIndex(function (child) {
                     return child === _this;
                 });
-                (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes));
+                (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes, false));
                 p.childNodes.forEach(function (n) {
                     if (n instanceof HTMLElement) {
                         n.parentNode = _this.parentNode;
@@ -1319,39 +1320,49 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 return it.test(tag);
             });
         }
-        var root = new HTMLElement(null, {}, '', null);
+        var createRange = function (startPos, endPos) {
+            return [startPos - frameFlagOffset, endPos - frameFlagOffset];
+        };
+        var root = new HTMLElement(null, {}, '', null, [0, data.length]);
         var currentParent = root;
         var stack = [root];
         var lastTextPos = -1;
         var match;
         // https://github.com/taoqf/node-html-parser/issues/38
         data = "<" + frameflag + ">" + data + "</" + frameflag + ">";
-        var _loop_1 = function () {
+        var dataEndPos = data.length - (frameflag.length + 2);
+        var frameFlagOffset = frameflag.length + 2;
+        while ((match = kMarkupPattern.exec(data))) {
+            var tagStartPos = kMarkupPattern.lastIndex - match[0].length;
+            var tagEndPos = kMarkupPattern.lastIndex;
+            // Add TextNode if content
             if (lastTextPos > -1) {
-                if (lastTextPos + match[0].length < kMarkupPattern.lastIndex) {
-                    // if has content
-                    var text = data.substring(lastTextPos, kMarkupPattern.lastIndex - match[0].length);
-                    currentParent.appendChild(new text_1.default(text, currentParent));
+                if (lastTextPos + match[0].length < tagEndPos) {
+                    var text = data.substring(lastTextPos, tagStartPos);
+                    currentParent.appendChild(new text_1.default(text, currentParent, createRange(lastTextPos, tagStartPos)));
                 }
             }
             lastTextPos = kMarkupPattern.lastIndex;
-            if (match[2] === frameflag) {
-                return "continue";
-            }
+            // https://github.com/taoqf/node-html-parser/issues/38
+            // Skip frameflag node
+            if (match[2] === frameflag)
+                continue;
+            // Handle comments
             if (match[0][1] === '!') {
-                // this is a comment
                 if (options.comment) {
                     // Only keep what is in between <!-- and -->
-                    var text = data.substring(lastTextPos - 3, lastTextPos - match[0].length + 4);
-                    currentParent.appendChild(new comment_1.default(text, currentParent));
+                    var text = data.substring(tagStartPos + 4, tagEndPos - 3);
+                    currentParent.appendChild(new comment_1.default(text, currentParent, createRange(tagStartPos, tagEndPos)));
                 }
-                return "continue";
+                continue;
             }
-            if (options.lowerCaseTagName) {
+            /* -- Handle tag matching -- */
+            // Fix tag casing if necessary
+            if (options.lowerCaseTagName)
                 match[2] = match[2].toLowerCase();
-            }
+            // Handle opening tags (ie. <this> not </that>)
             if (!match[1]) {
-                // not </ tags
+                /* Populate attributes */
                 var attrs = {};
                 for (var attMatch = void 0; (attMatch = kAttributePattern.exec(match[3]));) {
                     attrs[attMatch[2].toLowerCase()] = attMatch[4] || attMatch[5] || attMatch[6];
@@ -1360,50 +1371,46 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 if (!match[4] && kElementsClosedByOpening[tagName]) {
                     if (kElementsClosedByOpening[tagName][match[2]]) {
                         stack.pop();
-                        currentParent = back_1.default(stack);
+                        currentParent = (0, back_1.default)(stack);
                     }
                 }
-                // ignore container tag we add above
-                // https://github.com/taoqf/node-html-parser/issues/38
-                currentParent = currentParent.appendChild(new HTMLElement(match[2], attrs, match[3], null));
+                var tagEndPos_1 = kMarkupPattern.lastIndex;
+                var tagStartPos_1 = tagEndPos_1 - match[0].length;
+                currentParent = currentParent.appendChild(
+                // Initialize range (end position updated later for closed tags)
+                new HTMLElement(match[2], attrs, match[3], null, createRange(tagStartPos_1, tagEndPos_1)));
                 stack.push(currentParent);
                 if (is_block_text_element(match[2])) {
-                    // a little test to find next </script> or </style> ...
-                    var closeMarkup_1 = "</" + match[2] + ">";
-                    var index = (function () {
-                        if (options.lowerCaseTagName) {
-                            return data.toLocaleLowerCase().indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
-                        }
-                        return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
-                    })();
+                    // Find closing tag
+                    var closeMarkup = "</" + match[2] + ">";
+                    var closeIndex = options.lowerCaseTagName
+                        ? data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex)
+                        : data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
+                    var textEndPos = closeIndex === -1 ? dataEndPos : closeIndex;
                     if (element_should_be_ignore(match[2])) {
-                        var text = void 0;
-                        if (index === -1) {
-                            // there is no matching ending for the text element.
-                            text = data.substr(kMarkupPattern.lastIndex);
-                        }
-                        else {
-                            text = data.substring(kMarkupPattern.lastIndex, index);
-                        }
-                        if (text.length > 0) {
-                            currentParent.appendChild(new text_1.default(text, currentParent));
+                        var text = data.substring(tagEndPos_1, textEndPos);
+                        if (text.length > 0 && /\S/.test(text)) {
+                            currentParent.appendChild(new text_1.default(text, currentParent, createRange(tagEndPos_1, textEndPos)));
                         }
                     }
-                    if (index === -1) {
+                    if (closeIndex === -1) {
                         lastTextPos = kMarkupPattern.lastIndex = data.length + 1;
                     }
                     else {
-                        lastTextPos = kMarkupPattern.lastIndex = index + closeMarkup_1.length;
+                        lastTextPos = kMarkupPattern.lastIndex = closeIndex + closeMarkup.length;
+                        // Cause to be treated as self-closing, because no close found
                         match[1] = 'true';
                     }
                 }
             }
+            // Handle closing tags or self-closed elements (ie </tag> or <br>)
             if (match[1] || match[4] || kSelfClosingElements[match[2]]) {
-                // </ or /> or <br> etc.
                 while (true) {
                     if (currentParent.rawTagName === match[2]) {
+                        // Update range end for closed tag
+                        currentParent.range[1] = createRange(-1, Math.max(lastTextPos, tagEndPos))[1];
                         stack.pop();
-                        currentParent = back_1.default(stack);
+                        currentParent = (0, back_1.default)(stack);
                         break;
                     }
                     else {
@@ -1412,7 +1419,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                         if (kElementsClosedByClosing[tagName]) {
                             if (kElementsClosedByClosing[tagName][match[2]]) {
                                 stack.pop();
-                                currentParent = back_1.default(stack);
+                                currentParent = (0, back_1.default)(stack);
                                 continue;
                             }
                         }
@@ -1421,9 +1428,6 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                     }
                 }
             }
-        };
-        while ((match = kMarkupPattern.exec(data))) {
-            _loop_1();
         }
         return stack;
     }
@@ -1436,10 +1440,10 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
         if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
         var stack = base_parse(data, options);
         var root = stack[0];
-        var _loop_2 = function () {
+        var _loop_1 = function () {
             // Handle each error elements.
             var last = stack.pop();
-            var oneBefore = back_1.default(stack);
+            var oneBefore = (0, back_1.default)(stack);
             if (last.parentNode && last.parentNode.parentNode) {
                 if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
                     // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
@@ -1462,7 +1466,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             }
         };
         while (stack.length > 1) {
-            _loop_2();
+            _loop_1();
         }
         // response.childNodes.forEach((node) => {
         // 	if (node instanceof HTMLElement) {
@@ -1480,10 +1484,16 @@ define("nodes/node", ["require", "exports", "he"], function (require, exports, h
      * Node Class as base class for TextNode and HTMLElement.
      */
     var Node = /** @class */ (function () {
-        function Node(parentNode) {
+        function Node(parentNode, range) {
             if (parentNode === void 0) { parentNode = null; }
             this.parentNode = parentNode;
             this.childNodes = [];
+            Object.defineProperty(this, 'range', {
+                enumerable: false,
+                writable: true,
+                configurable: true,
+                value: range !== null && range !== void 0 ? range : [-1, -1]
+            });
         }
         Object.defineProperty(Node.prototype, "innerText", {
             get: function () {
@@ -1494,10 +1504,10 @@ define("nodes/node", ["require", "exports", "he"], function (require, exports, h
         });
         Object.defineProperty(Node.prototype, "textContent", {
             get: function () {
-                return he_3.decode(this.rawText);
+                return (0, he_3.decode)(this.rawText);
             },
             set: function (val) {
-                this.rawText = he_3.encode(val);
+                this.rawText = (0, he_3.encode)(val);
             },
             enumerable: false,
             configurable: true
@@ -1513,8 +1523,8 @@ define("nodes/comment", ["require", "exports", "nodes/node", "nodes/type"], func
     type_4 = __importDefault(type_4);
     var CommentNode = /** @class */ (function (_super) {
         __extends(CommentNode, _super);
-        function CommentNode(rawText, parentNode) {
-            var _this = _super.call(this, parentNode) || this;
+        function CommentNode(rawText, parentNode, range) {
+            var _this = _super.call(this, parentNode, range) || this;
             _this.rawText = rawText;
             /**
              * Node Type declaration.
@@ -1556,7 +1566,7 @@ define("valid", ["require", "exports", "nodes/html"], function (require, exports
      */
     function valid(data, options) {
         if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
-        var stack = html_2.base_parse(data, options);
+        var stack = (0, html_2.base_parse)(data, options);
         return Boolean(stack.length === 1);
     }
     exports.default = valid;

package/dist/nodes/comment.d.ts

@@ -3,7 +3,7 @@ import NodeType from './type';
 import HTMLElement from './html';
 export default class CommentNode extends Node {
     rawText: string;
-    constructor(rawText: string, parentNode: HTMLElement);
+    constructor(rawText: string, parentNode: HTMLElement, range?: [number, number]);
     /**
      * Node Type declaration.
      * @type {Number}
@@ -14,5 +14,5 @@ export default class CommentNode extends Node {
      * @return {string} text content
      */
     get text(): string;
-    toString(): `<!--${string}-->`;
+    toString(): string;
 }

package/dist/nodes/comment.js

@@ -22,8 +22,8 @@ var node_1 = __importDefault(require("./node"));
 var type_1 = __importDefault(require("./type"));
 var CommentNode = /** @class */ (function (_super) {
     __extends(CommentNode, _super);
-    function CommentNode(rawText, parentNode) {
-        var _this = _super.call(this, parentNode) || this;
+    function CommentNode(rawText, parentNode, range) {
+        var _this = _super.call(this, parentNode, range) || this;
         _this.rawText = rawText;
         /**
          * Node Type declaration.

package/dist/nodes/html.d.ts

@@ -59,7 +59,7 @@ export default class HTMLElement extends Node {
      *
      * @memberof HTMLElement
      */
-    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs: string, parentNode: HTMLElement | null);
+    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs: string, parentNode: HTMLElement | null, range?: [number, number]);
     /**
      * Remove current element
      */

package/dist/nodes/html.js

@@ -25,10 +25,14 @@ var __assign = (this && this.__assign) || function () {
     };
     return __assign.apply(this, arguments);
 };
-var __spreadArray = (this && this.__spreadArray) || function (to, from) {
-    for (var i = 0, il = from.length, j = to.length; i < il; i++, j++)
-        to[j] = from[i];
-    return to;
+var __spreadArray = (this && this.__spreadArray) || function (to, from, pack) {
+    if (pack || arguments.length === 2) for (var i = 0, l = from.length, ar; i < l; i++) {
+        if (ar || !(i in from)) {
+            if (!ar) ar = Array.prototype.slice.call(from, 0, i);
+            ar[i] = from[i];
+        }
+    }
+    return to.concat(ar || from);
 };
 var __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
@@ -200,9 +204,9 @@ var HTMLElement = /** @class */ (function (_super) {
      *
      * @memberof HTMLElement
      */
-    function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode) {
+    function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode, range) {
         if (rawAttrs === void 0) { rawAttrs = ''; }
-        var _this = _super.call(this, parentNode) || this;
+        var _this = _super.call(this, parentNode, range) || this;
         _this.rawAttrs = rawAttrs;
         /**
          * Node Type declaration.
@@ -434,7 +438,7 @@ var HTMLElement = /** @class */ (function (_super) {
         var idx = this.parentNode.childNodes.findIndex(function (child) {
             return child === _this;
         });
-        this.parentNode.childNodes = __spreadArray(__spreadArray(__spreadArray([], this.parentNode.childNodes.slice(0, idx)), content), this.parentNode.childNodes.slice(idx + 1));
+        this.parentNode.childNodes = __spreadArray(__spreadArray(__spreadArray([], this.parentNode.childNodes.slice(0, idx), true), content, true), this.parentNode.childNodes.slice(idx + 1), true);
     };
     Object.defineProperty(HTMLElement.prototype, "outerHTML", {
         get: function () {
@@ -527,7 +531,7 @@ var HTMLElement = /** @class */ (function (_super) {
      * @return {HTMLElement[]}  matching elements
      */
     HTMLElement.prototype.querySelectorAll = function (selector) {
-        return css_select_1.selectAll(selector, this, {
+        return (0, css_select_1.selectAll)(selector, this, {
             xmlMode: true,
             adapter: matcher_1.default
         });
@@ -595,7 +599,7 @@ var HTMLElement = /** @class */ (function (_super) {
      * @return {HTMLElement}    matching node
      */
     HTMLElement.prototype.querySelector = function (selector) {
-        return css_select_1.selectOne(selector, this, {
+        return (0, css_select_1.selectOne)(selector, this, {
             xmlMode: true,
             adapter: matcher_1.default
         });
@@ -668,17 +672,14 @@ var HTMLElement = /** @class */ (function (_super) {
         }
         el = this;
         while (el) {
-            var e = css_select_1.selectOne(selector, el, {
+            var e = (0, css_select_1.selectOne)(selector, el, {
                 xmlMode: true,
                 adapter: __assign(__assign({}, matcher_1.default), { getChildren: function (node) {
                         var child = mapChild.get(node);
                         return child && [child];
-                    },
-                    getSiblings: function (node) {
+                    }, getSiblings: function (node) {
                         return [node];
-                    },
-                    findOne: findOne,
-                    findAll: function () {
+                    }, findOne: findOne, findAll: function () {
                         return [];
                     } })
             });
@@ -717,7 +718,7 @@ var HTMLElement = /** @class */ (function (_super) {
          * @return {Node} last child node
          */
         get: function () {
-            return back_1.default(this.childNodes);
+            return (0, back_1.default)(this.childNodes);
         },
         enumerable: false,
         configurable: true
@@ -879,7 +880,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
             });
-            (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes));
+            (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes, false));
             p.childNodes.forEach(function (n) {
                 if (n instanceof HTMLElement) {
                     n.parentNode = _this.parentNode;
@@ -898,7 +899,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
             });
-            (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes));
+            (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes, false));
             p.childNodes.forEach(function (n) {
                 if (n instanceof HTMLElement) {
                     n.parentNode = _this.parentNode;
@@ -1073,39 +1074,49 @@ function base_parse(data, options) {
             return it.test(tag);
         });
     }
-    var root = new HTMLElement(null, {}, '', null);
+    var createRange = function (startPos, endPos) {
+        return [startPos - frameFlagOffset, endPos - frameFlagOffset];
+    };
+    var root = new HTMLElement(null, {}, '', null, [0, data.length]);
     var currentParent = root;
     var stack = [root];
     var lastTextPos = -1;
     var match;
     // https://github.com/taoqf/node-html-parser/issues/38
     data = "<" + frameflag + ">" + data + "</" + frameflag + ">";
-    var _loop_1 = function () {
+    var dataEndPos = data.length - (frameflag.length + 2);
+    var frameFlagOffset = frameflag.length + 2;
+    while ((match = kMarkupPattern.exec(data))) {
+        var tagStartPos = kMarkupPattern.lastIndex - match[0].length;
+        var tagEndPos = kMarkupPattern.lastIndex;
+        // Add TextNode if content
         if (lastTextPos > -1) {
-            if (lastTextPos + match[0].length < kMarkupPattern.lastIndex) {
-                // if has content
-                var text = data.substring(lastTextPos, kMarkupPattern.lastIndex - match[0].length);
-                currentParent.appendChild(new text_1.default(text, currentParent));
+            if (lastTextPos + match[0].length < tagEndPos) {
+                var text = data.substring(lastTextPos, tagStartPos);
+                currentParent.appendChild(new text_1.default(text, currentParent, createRange(lastTextPos, tagStartPos)));
             }
         }
         lastTextPos = kMarkupPattern.lastIndex;
-        if (match[2] === frameflag) {
-            return "continue";
-        }
+        // https://github.com/taoqf/node-html-parser/issues/38
+        // Skip frameflag node
+        if (match[2] === frameflag)
+            continue;
+        // Handle comments
         if (match[0][1] === '!') {
-            // this is a comment
             if (options.comment) {
                 // Only keep what is in between <!-- and -->
-                var text = data.substring(lastTextPos - 3, lastTextPos - match[0].length + 4);
-                currentParent.appendChild(new comment_1.default(text, currentParent));
+                var text = data.substring(tagStartPos + 4, tagEndPos - 3);
+                currentParent.appendChild(new comment_1.default(text, currentParent, createRange(tagStartPos, tagEndPos)));
             }
-            return "continue";
+            continue;
         }
-        if (options.lowerCaseTagName) {
+        /* -- Handle tag matching -- */
+        // Fix tag casing if necessary
+        if (options.lowerCaseTagName)
             match[2] = match[2].toLowerCase();
-        }
+        // Handle opening tags (ie. <this> not </that>)
         if (!match[1]) {
-            // not </ tags
+            /* Populate attributes */
             var attrs = {};
             for (var attMatch = void 0; (attMatch = kAttributePattern.exec(match[3]));) {
                 attrs[attMatch[2].toLowerCase()] = attMatch[4] || attMatch[5] || attMatch[6];
@@ -1114,50 +1125,46 @@ function base_parse(data, options) {
             if (!match[4] && kElementsClosedByOpening[tagName]) {
                 if (kElementsClosedByOpening[tagName][match[2]]) {
                     stack.pop();
-                    currentParent = back_1.default(stack);
+                    currentParent = (0, back_1.default)(stack);
                 }
             }
-            // ignore container tag we add above
-            // https://github.com/taoqf/node-html-parser/issues/38
-            currentParent = currentParent.appendChild(new HTMLElement(match[2], attrs, match[3], null));
+            var tagEndPos_1 = kMarkupPattern.lastIndex;
+            var tagStartPos_1 = tagEndPos_1 - match[0].length;
+            currentParent = currentParent.appendChild(
+            // Initialize range (end position updated later for closed tags)
+            new HTMLElement(match[2], attrs, match[3], null, createRange(tagStartPos_1, tagEndPos_1)));
             stack.push(currentParent);
             if (is_block_text_element(match[2])) {
-                // a little test to find next </script> or </style> ...
-                var closeMarkup_1 = "</" + match[2] + ">";
-                var index = (function () {
-                    if (options.lowerCaseTagName) {
-                        return data.toLocaleLowerCase().indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
-                    }
-                    return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
-                })();
+                // Find closing tag
+                var closeMarkup = "</" + match[2] + ">";
+                var closeIndex = options.lowerCaseTagName
+                    ? data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex)
+                    : data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
+                var textEndPos = closeIndex === -1 ? dataEndPos : closeIndex;
                 if (element_should_be_ignore(match[2])) {
-                    var text = void 0;
-                    if (index === -1) {
-                        // there is no matching ending for the text element.
-                        text = data.substr(kMarkupPattern.lastIndex);
-                    }
-                    else {
-                        text = data.substring(kMarkupPattern.lastIndex, index);
-                    }
-                    if (text.length > 0) {
-                        currentParent.appendChild(new text_1.default(text, currentParent));
+                    var text = data.substring(tagEndPos_1, textEndPos);
+                    if (text.length > 0 && /\S/.test(text)) {
+                        currentParent.appendChild(new text_1.default(text, currentParent, createRange(tagEndPos_1, textEndPos)));
                     }
                 }
-                if (index === -1) {
+                if (closeIndex === -1) {
                     lastTextPos = kMarkupPattern.lastIndex = data.length + 1;
                 }
                 else {
-                    lastTextPos = kMarkupPattern.lastIndex = index + closeMarkup_1.length;
+                    lastTextPos = kMarkupPattern.lastIndex = closeIndex + closeMarkup.length;
+                    // Cause to be treated as self-closing, because no close found
                     match[1] = 'true';
                 }
             }
         }
+        // Handle closing tags or self-closed elements (ie </tag> or <br>)
         if (match[1] || match[4] || kSelfClosingElements[match[2]]) {
-            // </ or /> or <br> etc.
             while (true) {
                 if (currentParent.rawTagName === match[2]) {
+                    // Update range end for closed tag
+                    currentParent.range[1] = createRange(-1, Math.max(lastTextPos, tagEndPos))[1];
                     stack.pop();
-                    currentParent = back_1.default(stack);
+                    currentParent = (0, back_1.default)(stack);
                     break;
                 }
                 else {
@@ -1166,7 +1173,7 @@ function base_parse(data, options) {
                     if (kElementsClosedByClosing[tagName]) {
                         if (kElementsClosedByClosing[tagName][match[2]]) {
                             stack.pop();
-                            currentParent = back_1.default(stack);
+                            currentParent = (0, back_1.default)(stack);
                             continue;
                         }
                     }
@@ -1175,9 +1182,6 @@ function base_parse(data, options) {
                 }
             }
         }
-    };
-    while ((match = kMarkupPattern.exec(data))) {
-        _loop_1();
     }
     return stack;
 }
@@ -1190,10 +1194,10 @@ function parse(data, options) {
     if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
     var stack = base_parse(data, options);
     var root = stack[0];
-    var _loop_2 = function () {
+    var _loop_1 = function () {
         // Handle each error elements.
         var last = stack.pop();
-        var oneBefore = back_1.default(stack);
+        var oneBefore = (0, back_1.default)(stack);
         if (last.parentNode && last.parentNode.parentNode) {
             if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
                 // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
@@ -1216,7 +1220,7 @@ function parse(data, options) {
         }
     };
     while (stack.length > 1) {
-        _loop_2();
+        _loop_1();
     }
     // response.childNodes.forEach((node) => {
     // 	if (node instanceof HTMLElement) {

package/dist/nodes/node.d.ts

@@ -7,10 +7,11 @@ export default abstract class Node {
     parentNode: HTMLElement;
     abstract nodeType: NodeType;
     childNodes: Node[];
+    range: readonly [number, number];
     abstract text: string;
     abstract rawText: string;
     abstract toString(): string;
-    constructor(parentNode?: HTMLElement);
+    constructor(parentNode?: HTMLElement, range?: [number, number]);
     get innerText(): string;
     get textContent(): string;
     set textContent(val: string);

package/dist/nodes/node.js

@@ -5,10 +5,16 @@ var he_1 = require("he");
  * Node Class as base class for TextNode and HTMLElement.
  */
 var Node = /** @class */ (function () {
-    function Node(parentNode) {
+    function Node(parentNode, range) {
         if (parentNode === void 0) { parentNode = null; }
         this.parentNode = parentNode;
         this.childNodes = [];
+        Object.defineProperty(this, 'range', {
+            enumerable: false,
+            writable: true,
+            configurable: true,
+            value: range !== null && range !== void 0 ? range : [-1, -1]
+        });
     }
     Object.defineProperty(Node.prototype, "innerText", {
         get: function () {
@@ -19,10 +25,10 @@ var Node = /** @class */ (function () {
     });
     Object.defineProperty(Node.prototype, "textContent", {
         get: function () {
-            return he_1.decode(this.rawText);
+            return (0, he_1.decode)(this.rawText);
         },
         set: function (val) {
-            this.rawText = he_1.encode(val);
+            this.rawText = (0, he_1.encode)(val);
         },
         enumerable: false,
         configurable: true

package/dist/nodes/text.d.ts

@@ -6,7 +6,7 @@ import NodeType from './type';
  * @param {string} value [description]
  */
 export default class TextNode extends Node {
-    constructor(rawText: string, parentNode: HTMLElement);
+    constructor(rawText: string, parentNode: HTMLElement, range?: [number, number]);
     /**
      * Node Type declaration.
      * @type {Number}

package/dist/nodes/text.js

@@ -27,8 +27,8 @@ var type_1 = __importDefault(require("./type"));
  */
 var TextNode = /** @class */ (function (_super) {
     __extends(TextNode, _super);
-    function TextNode(rawText, parentNode) {
-        var _this = _super.call(this, parentNode) || this;
+    function TextNode(rawText, parentNode, range) {
+        var _this = _super.call(this, parentNode, range) || this;
         /**
          * Node Type declaration.
          * @type {Number}
@@ -84,7 +84,7 @@ var TextNode = /** @class */ (function (_super) {
          * @return {string} text content
          */
         get: function () {
-            return he_1.decode(this.rawText);
+            return (0, he_1.decode)(this.rawText);
         },
         enumerable: false,
         configurable: true

package/dist/valid.js

@@ -7,7 +7,7 @@ var html_1 = require("./nodes/html");
  */
 function valid(data, options) {
     if (options === void 0) { options = { lowerCaseTagName: false, comment: false }; }
-    var stack = html_1.base_parse(data, options);
+    var stack = (0, html_1.base_parse)(data, options);
     return Boolean(stack.length === 1);
 }
 exports.default = valid;

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "node-html-parser",
-	"version": "4.0.0",
+	"version": "4.1.3",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
 	"main": "dist/index.js",
 	"module": "dist/esm/index.js",
@@ -11,7 +11,7 @@
 		"clean": "del-cli ./dist/",
 		"ts:cjs": "tsc -m commonjs",
 		"ts:amd": "tsc -t es5 -m amd -d false --outFile ./dist/main.js",
-		"ts:esm": "tsc -t esnext -m esnext -d false --outDir ./dist/esm/",
+		"ts:esm": "tsc -t es2019 -m esnext -d false --outDir ./dist/esm/",
 		"build": "npm run lint && npm run clean && npm run ts:cjs && npm run ts:amd && npm run ts:esm",
 		"dev": "tsc -w & mocha -w ./test/*.js",
 		"pretest": "tsc -m commonjs",
@@ -54,6 +54,7 @@
 		"htmlparser-benchmark": "^1.1.3",
 		"htmlparser2": "^6.0.0",
 		"mocha": "latest",
+		"mocha-each": "^2.0.1",
 		"np": "latest",
 		"parse5": "^6.0.1",
 		"should": "latest",