node-html-parser 5.1.0 → 5.2.5

package/CHANGELOG.md

@@ -4,23 +4,9 @@ All notable changes to this project will be documented in this file. See [standa
 
 ## [5.1.0](https://github.com/taoqf/node-fast-html-parser/compare/v4.1.5...v5.1.0) (2021-10-28)
 
-
-### ⚠ BREAKING CHANGES
-
-* Add esm named export support (closes #160 closes #139)
-
 ### Features
 
-* Add esm named export support (closes [#160](https://github.com/taoqf/node-fast-html-parser/issues/160) closes [#139](https://github.com/taoqf/node-fast-html-parser/issues/139)) ([0d4b922](https://github.com/taoqf/node-fast-html-parser/commit/0d4b922eefd6210fe802991e464b21b0c69d5f63))
-* Added HTMLElement#getElementsByTagName ([d462e44](https://github.com/taoqf/node-fast-html-parser/commit/d462e449e7ebb00a5a43fb574133681ad5a62475))
-* Expose `HTMLElement#rawAttrs` (make public) ([34f1595](https://github.com/taoqf/node-fast-html-parser/commit/34f1595756c0974b6ae7ef5755a615f09e421f32))
-* Improved parsing performance + matching (closes [#164](https://github.com/taoqf/node-fast-html-parser/issues/164)) ([3c5b8e2](https://github.com/taoqf/node-fast-html-parser/commit/3c5b8e2a9104b01a8ca899a7970507463e42adaf))
-
-
-### Bug Fixes
-
-* Add null to return type for HTMLElement#querySelector (closes [#157](https://github.com/taoqf/node-fast-html-parser/issues/157)) ([2b65583](https://github.com/taoqf/node-fast-html-parser/commit/2b655839bd3868c41fb19cae5786ca097565bc7f))
-* blockTextElements incorrectly matching partial tag (detail) (fixes [#156](https://github.com/taoqf/node-fast-html-parser/issues/156) fixes [#124](https://github.com/taoqf/node-fast-html-parser/issues/124)) ([6823349](https://github.com/taoqf/node-fast-html-parser/commit/6823349fdf1809c7484c70d948aa24930ef4983f))
+* Exposed `HTMLElement#rawAttrs` (made public) ([34f1595](https://github.com/taoqf/node-fast-html-parser/commit/34f1595756c0974b6ae7ef5755a615f09e421f32))
 
 ## [5.0.0](https://github.com/taoqf/node-fast-html-parser/compare/v4.1.5...v5.0.0) (2021-10-10)
 

package/README.md

@@ -19,15 +19,18 @@ npm install --save node-html-parser
 
 ## Performance
 
-Faster than htmlparser2!
-
 ```shell
-htmlparser      :26.7111 ms/file ± 170.066
-cheerio         :24.2480 ms/file ± 17.1711
-parse5          :13.7239 ms/file ± 8.68561
-high5           :7.75466 ms/file ± 5.33549
-htmlparser2     :5.27376 ms/file ± 8.68456
-node-html-parser:2.85768 ms/file ± 2.87784
+cheerio         :12.0726 ms/file ± 7.31605
+parse5          :8.18615 ms/file ± 6.15337
+node-html-parser (last release):2.16533 ms/file ± 1.56924
+htmlparser      :17.0658 ms/file ± 120.901
+htmlparser2     :2.62695 ms/file ± 4.17579
+node-html-parser:2.14907 ms/file ± 1.66632
+html-parser     :24.6505 ms/file ± 18.9996
+htmljs-parser   :5.81797 ms/file ± 6.55537
+html-dom-parser :2.52265 ms/file ± 3.54858
+html5parser     :2.01144 ms/file ± 2.53570
+high5           :3.91342 ms/file ± 2.65563
 ```
 
 Tested with [htmlparser-benchmark](https://github.com/AndreasMadsen/htmlparser-benchmark).
@@ -70,15 +73,15 @@ var root = HTMLParser.parse('<ul id="list"><li>Hello World</li></ul>');
 
 ### parse(data[, options])
 
-Parse given data, and return root of the generated DOM.
+Parse the data provided, and return the root of the generated DOM.
 
 - **data**, data to parse
 - **options**, parse options
 
   ```js
   {
-    lowerCaseTagName: false,  // convert tag name to lower case (hurt performance heavily)
-    comment: false,            // retrieve comments (hurt performance slightly)
+    lowerCaseTagName: false,  // convert tag name to lower case (hurts performance heavily)
+    comment: false,            // retrieve comments (hurts performance slightly)
     blockTextElements: {
       script: true,	// keep text content when parsing
       noscript: true,	// keep text content when parsing
@@ -90,7 +93,7 @@ Parse given data, and return root of the generated DOM.
 
 ### valid(data[, options])
 
-Parse given data, return true if the givent data is valid, and return false if not.
+Parse the data provided, return true if the given data is valid, and return false if not.
 
 ## HTMLElement Methods
 
@@ -106,7 +109,7 @@ Remove whitespaces in this sub tree.
 
 Query CSS selector to find matching nodes.
 
-Note: Full css3 selector supported since v3.0.0.
+Note: Full range of CSS3 selectors supported since v3.0.0.
 
 ### HTMLElement#querySelector(selector)
 
@@ -116,7 +119,7 @@ Query CSS Selector to find matching node.
 
 Get all elements with the specified tagName.
 
-Note: * for all elements.
+Note: Use * for all elements.
 
 ### HTMLElement#closest(selector)
 
@@ -128,7 +131,7 @@ Append a child node to childNodes
 
 ### HTMLElement#insertAdjacentHTML(where, html)
 
-parses the specified text as HTML and inserts the resulting nodes into the DOM tree at a specified position.
+Parses the specified text as HTML and inserts the resulting nodes into the DOM tree at a specified position.
 
 ### HTMLElement#setAttribute(key: string, value: string)
 
@@ -186,15 +189,19 @@ Remove class name.
 
 #### HTMLElement#classList.toggle(className: string):void
 
-Toggle class.
+Toggle class. Remove it if it is already included, otherwise add.
 
 #### HTMLElement#classList.contains(className: string): boolean
 
-Get if contains
+Returns true if the classname is already in the classList.
 
 #### HTMLElement#classList.values()
 
-get class names
+Get class names.
+
+#### Node#clone()
+
+Clone a node.
 
 ## HTMLElement Properties
 
@@ -205,28 +212,28 @@ Get unescaped text value of current node and its children. Like `innerText`.
 
 ### HTMLElement#rawText
 
-Get escaped (as-it) text value of current node and its children. May have
+Get escaped (as-is) text value of current node and its children. May have
 `&amp;` in it. (fast)
 
 ### HTMLElement#tagName
 
-Get tag name of HTMLElement. Notice: the returned value would be an uppercase string.
+Get or Set tag name of HTMLElement. Notice: the returned value would be an uppercase string.
 
 ### HTMLElement#structuredText
 
-Get structured Text
+Get structured Text.
 
 ### HTMLElement#structure
 
-Get DOM structure
+Get DOM structure.
 
 ### HTMLElement#firstChild
 
-Get first child node
+Get first child node.
 
 ### HTMLElement#lastChild
 
-Get last child node
+Get last child node.
 
 ### HTMLElement#innerHTML
 
@@ -244,6 +251,14 @@ Returns a reference to the next child node of the current element's parent.
 
 Returns a reference to the next child element of the current element's parent.
 
+### HTMLElement#previousSibling
+
+Returns a reference to the previous child node of the current element's parent.
+
+### HTMLElement#previousElementSibling
+
+Returns a reference to the previous child element of the current element's parent.
+
 ### HTMLElement#textContent
 
 Get or Set textContent of current element, more efficient than [set_content](#htmlelementset_contentcontent-string--node--node).

package/dist/main.js

@@ -1,3 +1,6 @@
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
 var __extends = (this && this.__extends) || (function () {
     var extendStatics = function (d, b) {
         extendStatics = Object.setPrototypeOf ||
@@ -13,9 +16,6 @@ var __extends = (this && this.__extends) || (function () {
         d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());
     };
 })();
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
 var __assign = (this && this.__assign) || function () {
     __assign = Object.assign || function(t) {
         for (var s, i = 1, n = arguments.length; i < n; i++) {
@@ -55,134 +55,51 @@ define("nodes/type", ["require", "exports"], function (require, exports) {
     })(NodeType || (NodeType = {}));
     exports.default = NodeType;
 });
-define("nodes/text", ["require", "exports", "he", "nodes/node", "nodes/type"], function (require, exports, he_1, node_1, type_1) {
+define("nodes/node", ["require", "exports", "he"], function (require, exports, he_1) {
     "use strict";
     Object.defineProperty(exports, "__esModule", { value: true });
-    node_1 = __importDefault(node_1);
-    type_1 = __importDefault(type_1);
     /**
-     * TextNode to contain a text element in DOM tree.
-     * @param {string} value [description]
+     * Node Class as base class for TextNode and HTMLElement.
      */
-    var TextNode = /** @class */ (function (_super) {
-        __extends(TextNode, _super);
-        function TextNode(rawText, parentNode, range) {
-            var _this = _super.call(this, parentNode, range) || this;
-            /**
-             * Node Type declaration.
-             * @type {Number}
-             */
-            _this.nodeType = type_1.default.TEXT_NODE;
-            _this._rawText = rawText;
-            return _this;
+    var Node = /** @class */ (function () {
+        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(TextNode.prototype, "rawText", {
-            get: function () {
-                return this._rawText;
-            },
-            /**
-             * Set rawText and invalidate trimmed caches
-             */
-            set: function (text) {
-                this._rawText = text;
-                this._trimmedRawText = void 0;
-                this._trimmedText = void 0;
-            },
-            enumerable: false,
-            configurable: true
-        });
-        Object.defineProperty(TextNode.prototype, "trimmedRawText", {
-            /**
-             * Returns raw text with all whitespace trimmed except single leading/trailing non-breaking space
-             */
-            get: function () {
-                if (this._trimmedRawText !== undefined)
-                    return this._trimmedRawText;
-                this._trimmedRawText = trimText(this.rawText);
-                return this._trimmedRawText;
-            },
-            enumerable: false,
-            configurable: true
-        });
-        Object.defineProperty(TextNode.prototype, "trimmedText", {
-            /**
-             * Returns text with all whitespace trimmed except single leading/trailing non-breaking space
-             */
+        Object.defineProperty(Node.prototype, "innerText", {
             get: function () {
-                if (this._trimmedText !== undefined)
-                    return this._trimmedText;
-                this._trimmedText = trimText(this.text);
-                return this._trimmedText;
+                return this.rawText;
             },
             enumerable: false,
             configurable: true
         });
-        Object.defineProperty(TextNode.prototype, "text", {
-            /**
-             * Get unescaped text value of current node and its children.
-             * @return {string} text content
-             */
+        Object.defineProperty(Node.prototype, "textContent", {
             get: function () {
                 return (0, he_1.decode)(this.rawText);
             },
-            enumerable: false,
-            configurable: true
-        });
-        Object.defineProperty(TextNode.prototype, "isWhitespace", {
-            /**
-             * Detect if the node contains only white space.
-             * @return {boolean}
-             */
-            get: function () {
-                return /^(\s|&nbsp;)*$/.test(this.rawText);
+            set: function (val) {
+                this.rawText = (0, he_1.encode)(val);
             },
             enumerable: false,
             configurable: true
         });
-        TextNode.prototype.toString = function () {
-            return this.rawText;
-        };
-        return TextNode;
-    }(node_1.default));
-    exports.default = TextNode;
-    /**
-     * Trim whitespace except single leading/trailing non-breaking space
-     */
-    function trimText(text) {
-        var i = 0;
-        var startPos;
-        var endPos;
-        while (i >= 0 && i < text.length) {
-            if (/\S/.test(text[i])) {
-                if (startPos === undefined) {
-                    startPos = i;
-                    i = text.length;
-                }
-                else {
-                    endPos = i;
-                    i = void 0;
-                }
-            }
-            if (startPos === undefined)
-                i++;
-            else
-                i--;
-        }
-        if (startPos === undefined)
-            startPos = 0;
-        if (endPos === undefined)
-            endPos = text.length - 1;
-        var hasLeadingSpace = startPos > 0 && /[^\S\r\n]/.test(text[startPos - 1]);
-        var hasTrailingSpace = endPos < (text.length - 1) && /[^\S\r\n]/.test(text[endPos + 1]);
-        return (hasLeadingSpace ? ' ' : '') + text.slice(startPos, endPos + 1) + (hasTrailingSpace ? ' ' : '');
-    }
+        return Node;
+    }());
+    exports.default = Node;
 });
-define("matcher", ["require", "exports", "nodes/type"], function (require, exports, type_2) {
+define("matcher", ["require", "exports", "nodes/type"], function (require, exports, type_1) {
     "use strict";
     Object.defineProperty(exports, "__esModule", { value: true });
-    type_2 = __importDefault(type_2);
+    type_1 = __importDefault(type_1);
     function isTag(node) {
-        return node && node.nodeType === type_2.default.ELEMENT_NODE;
+        return node && node.nodeType === type_1.default.ELEMENT_NODE;
     }
     function getAttributeValue(elem, name) {
         return isTag(elem) ? elem.getAttribute(name) : undefined;
@@ -282,21 +199,146 @@ define("matcher", ["require", "exports", "nodes/type"], function (require, expor
         findAll: findAll
     };
 });
-define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "nodes/type", "nodes/text", "matcher", "back", "nodes/comment"], function (require, exports, he_2, css_select_1, node_2, type_3, text_1, matcher_1, back_1, comment_1) {
+define("nodes/text", ["require", "exports", "he", "nodes/node", "nodes/type"], function (require, exports, he_2, node_1, type_2) {
+    "use strict";
+    Object.defineProperty(exports, "__esModule", { value: true });
+    node_1 = __importDefault(node_1);
+    type_2 = __importDefault(type_2);
+    /**
+     * TextNode to contain a text element in DOM tree.
+     * @param {string} value [description]
+     */
+    var TextNode = /** @class */ (function (_super) {
+        __extends(TextNode, _super);
+        function TextNode(rawText, parentNode, range) {
+            var _this = _super.call(this, parentNode, range) || this;
+            /**
+             * Node Type declaration.
+             * @type {Number}
+             */
+            _this.nodeType = type_2.default.TEXT_NODE;
+            _this._rawText = rawText;
+            return _this;
+        }
+        TextNode.prototype.clone = function () {
+            return new TextNode(this._rawText, null);
+        };
+        Object.defineProperty(TextNode.prototype, "rawText", {
+            get: function () {
+                return this._rawText;
+            },
+            /**
+             * Set rawText and invalidate trimmed caches
+             */
+            set: function (text) {
+                this._rawText = text;
+                this._trimmedRawText = void 0;
+                this._trimmedText = void 0;
+            },
+            enumerable: false,
+            configurable: true
+        });
+        Object.defineProperty(TextNode.prototype, "trimmedRawText", {
+            /**
+             * Returns raw text with all whitespace trimmed except single leading/trailing non-breaking space
+             */
+            get: function () {
+                if (this._trimmedRawText !== undefined)
+                    return this._trimmedRawText;
+                this._trimmedRawText = trimText(this.rawText);
+                return this._trimmedRawText;
+            },
+            enumerable: false,
+            configurable: true
+        });
+        Object.defineProperty(TextNode.prototype, "trimmedText", {
+            /**
+             * Returns text with all whitespace trimmed except single leading/trailing non-breaking space
+             */
+            get: function () {
+                if (this._trimmedText !== undefined)
+                    return this._trimmedText;
+                this._trimmedText = trimText(this.text);
+                return this._trimmedText;
+            },
+            enumerable: false,
+            configurable: true
+        });
+        Object.defineProperty(TextNode.prototype, "text", {
+            /**
+             * Get unescaped text value of current node and its children.
+             * @return {string} text content
+             */
+            get: function () {
+                return (0, he_2.decode)(this.rawText);
+            },
+            enumerable: false,
+            configurable: true
+        });
+        Object.defineProperty(TextNode.prototype, "isWhitespace", {
+            /**
+             * Detect if the node contains only white space.
+             * @return {boolean}
+             */
+            get: function () {
+                return /^(\s|&nbsp;)*$/.test(this.rawText);
+            },
+            enumerable: false,
+            configurable: true
+        });
+        TextNode.prototype.toString = function () {
+            return this.rawText;
+        };
+        return TextNode;
+    }(node_1.default));
+    exports.default = TextNode;
+    /**
+     * Trim whitespace except single leading/trailing non-breaking space
+     */
+    function trimText(text) {
+        var i = 0;
+        var startPos;
+        var endPos;
+        while (i >= 0 && i < text.length) {
+            if (/\S/.test(text[i])) {
+                if (startPos === undefined) {
+                    startPos = i;
+                    i = text.length;
+                }
+                else {
+                    endPos = i;
+                    i = void 0;
+                }
+            }
+            if (startPos === undefined)
+                i++;
+            else
+                i--;
+        }
+        if (startPos === undefined)
+            startPos = 0;
+        if (endPos === undefined)
+            endPos = text.length - 1;
+        var hasLeadingSpace = startPos > 0 && /[^\S\r\n]/.test(text[startPos - 1]);
+        var hasTrailingSpace = endPos < (text.length - 1) && /[^\S\r\n]/.test(text[endPos + 1]);
+        return (hasLeadingSpace ? ' ' : '') + text.slice(startPos, endPos + 1) + (hasTrailingSpace ? ' ' : '');
+    }
+});
+define("nodes/html", ["require", "exports", "css-select", "he", "back", "matcher", "nodes/comment", "nodes/node", "nodes/text", "nodes/type"], function (require, exports, css_select_1, he_3, back_1, matcher_1, comment_1, node_2, text_1, type_3) {
     "use strict";
     Object.defineProperty(exports, "__esModule", { value: true });
     exports.parse = exports.base_parse = void 0;
-    he_2 = __importDefault(he_2);
-    node_2 = __importDefault(node_2);
-    type_3 = __importDefault(type_3);
-    text_1 = __importDefault(text_1);
-    matcher_1 = __importDefault(matcher_1);
+    he_3 = __importDefault(he_3);
     back_1 = __importDefault(back_1);
+    matcher_1 = __importDefault(matcher_1);
     comment_1 = __importDefault(comment_1);
+    node_2 = __importDefault(node_2);
+    text_1 = __importDefault(text_1);
+    type_3 = __importDefault(type_3);
     var voidTags = new Set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']);
     function decode(val) {
         // clone string
-        return JSON.parse(JSON.stringify(he_2.default.decode(val)));
+        return JSON.parse(JSON.stringify(he_3.default.decode(val)));
     }
     // https://developer.mozilla.org/en-US/docs/Web/HTML/Block-level_elements
     var Htags = ['h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'header', 'hgroup'];
@@ -332,7 +374,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
         }
         DOMTokenList.prototype._validate = function (c) {
             if (/\s/.test(c)) {
-                throw new Error("DOMException in DOMTokenList.add: The token '" + c + "' contains HTML space characters, which are not valid in tokens.");
+                throw new Error("DOMException in DOMTokenList.add: The token '".concat(c, "' contains HTML space characters, which are not valid in tokens."));
             }
         };
         DOMTokenList.prototype.add = function (c) {
@@ -416,14 +458,14 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             );
             if (keyAttrs.id) {
                 if (!rawAttrs) {
-                    _this.rawAttrs = "id=\"" + keyAttrs.id + "\"";
+                    _this.rawAttrs = "id=\"".concat(keyAttrs.id, "\"");
                 }
             }
             if (keyAttrs.class) {
                 if (!rawAttrs) {
-                    var cls = "class=\"" + _this.classList.toString() + "\"";
+                    var cls = "class=\"".concat(_this.classList.toString(), "\"");
                     if (_this.rawAttrs) {
-                        _this.rawAttrs += " " + cls;
+                        _this.rawAttrs += " ".concat(cls);
                     }
                     else {
                         _this.rawAttrs = cls;
@@ -482,6 +524,9 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             get: function () {
                 return this.rawTagName ? this.rawTagName.toUpperCase() : this.rawTagName;
             },
+            set: function (newname) {
+                this.rawTagName = newname.toLowerCase();
+            },
             enumerable: false,
             configurable: true
         });
@@ -565,7 +610,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                         else {
                             var text = node.trimmedText;
                             if (currentBlock.prependWhitespace) {
-                                text = " " + text;
+                                text = " ".concat(text);
                                 currentBlock.prependWhitespace = false;
                             }
                             currentBlock.push(text);
@@ -586,8 +631,8 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
         HTMLElement.prototype.toString = function () {
             var tag = this.rawTagName;
             if (tag) {
-                var attrs = this.rawAttrs ? " " + this.rawAttrs : '';
-                return this.isVoidElement ? "<" + tag + attrs + ">" : "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
+                var attrs = this.rawAttrs ? " ".concat(this.rawAttrs) : '';
+                return this.isVoidElement ? "<".concat(tag).concat(attrs, ">") : "<".concat(tag).concat(attrs, ">").concat(this.innerHTML, "</").concat(tag, ">");
             }
             return this.innerHTML;
         };
@@ -683,9 +728,9 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                     res.push('  '.repeat(indention) + str);
                 }
                 function dfs(node) {
-                    var idStr = node.id ? "#" + node.id : '';
-                    var classStr = node.classList.length ? "." + node.classList.value.join('.') : ''; // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-member-access, @typescript-eslint/restrict-template-expressions, @typescript-eslint/no-unsafe-call
-                    write("" + node.rawTagName + idStr + classStr);
+                    var idStr = node.id ? "#".concat(node.id) : '';
+                    var classStr = node.classList.length ? ".".concat(node.classList.value.join('.')) : ''; // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-member-access, @typescript-eslint/restrict-template-expressions, @typescript-eslint/no-unsafe-call
+                    write("".concat(node.rawTagName).concat(idStr).concat(classStr));
                     indention++;
                     node.childNodes.forEach(function (childNode) {
                         if (childNode.nodeType === type_3.default.ELEMENT_NODE) {
@@ -944,7 +989,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 if (val === undefined || val === 'null') {
                     return name;
                 }
-                return name + "=" + val;
+                return "".concat(name, "=").concat(val);
             })
                 .join(' ');
             // Update this.id
@@ -991,7 +1036,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 var val = _this.quoteAttribute(attrs[name]);
                 if (val === 'null' || val === '""')
                     return name;
-                return name + "=" + val;
+                return "".concat(name, "=").concat(val);
             })
                 .join(' ');
             // Update this.id
@@ -1019,7 +1064,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 var val = attributes[name];
                 if (val === 'null' || val === '""')
                     return name;
-                return name + "=" + _this.quoteAttribute(String(val));
+                return "".concat(name, "=").concat(_this.quoteAttribute(String(val)));
             })
                 .join(' ');
         };
@@ -1061,7 +1106,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 });
             }
             else {
-                throw new Error("The value provided ('" + where + "') is not one of 'beforebegin', 'afterbegin', 'beforeend', or 'afterend'");
+                throw new Error("The value provided ('".concat(where, "') is not one of 'beforebegin', 'afterbegin', 'beforeend', or 'afterend'"));
             }
             // if (!where || html === undefined || html === null) {
             // 	return;
@@ -1106,6 +1151,45 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             enumerable: false,
             configurable: true
         });
+        Object.defineProperty(HTMLElement.prototype, "previousSibling", {
+            get: function () {
+                if (this.parentNode) {
+                    var children = this.parentNode.childNodes;
+                    var i = children.length;
+                    while (i > 0) {
+                        var child = children[--i];
+                        if (this === child)
+                            return children[i - 1] || null;
+                    }
+                    return null;
+                }
+            },
+            enumerable: false,
+            configurable: true
+        });
+        Object.defineProperty(HTMLElement.prototype, "previousElementSibling", {
+            get: function () {
+                if (this.parentNode) {
+                    var children = this.parentNode.childNodes;
+                    var i = children.length;
+                    var find = false;
+                    while (i > 0) {
+                        var child = children[--i];
+                        if (find) {
+                            if (child instanceof HTMLElement) {
+                                return child || null;
+                            }
+                        }
+                        else if (this === child) {
+                            find = true;
+                        }
+                    }
+                    return null;
+                }
+            },
+            enumerable: false,
+            configurable: true
+        });
         Object.defineProperty(HTMLElement.prototype, "classNames", {
             get: function () {
                 return this.classList.toString();
@@ -1113,6 +1197,12 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             enumerable: false,
             configurable: true
         });
+        /**
+         * Clone this Node
+         */
+        HTMLElement.prototype.clone = function () {
+            return parse(this.toString()).firstChild;
+        };
         return HTMLElement;
     }(node_2.default));
     exports.default = HTMLElement;
@@ -1205,8 +1295,8 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             pre: true,
         };
         var element_names = Object.keys(elements);
-        var kBlockTextElements = element_names.map(function (it) { return new RegExp("^" + it + "$", 'i'); });
-        var kIgnoreElements = element_names.filter(function (it) { return elements[it]; }).map(function (it) { return new RegExp("^" + it + "$", 'i'); });
+        var kBlockTextElements = element_names.map(function (it) { return new RegExp("^".concat(it, "$"), 'i'); });
+        var kIgnoreElements = element_names.filter(function (it) { return elements[it]; }).map(function (it) { return new RegExp("^".concat(it, "$"), 'i'); });
         function element_should_be_ignore(tag) {
             return kIgnoreElements.some(function (it) { return it.test(tag); });
         }
@@ -1221,7 +1311,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
         var noNestedTagIndex = undefined;
         var match;
         // https://github.com/taoqf/node-html-parser/issues/38
-        data = "<" + frameflag + ">" + data + "</" + frameflag + ">";
+        data = "<".concat(frameflag, ">").concat(data, "</").concat(frameflag, ">");
         var lowerCaseTagName = options.lowerCaseTagName;
         var dataEndPos = data.length - (frameflag.length + 2);
         var frameFlagOffset = frameflag.length + 2;
@@ -1289,7 +1379,7 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
                 stack.push(currentParent);
                 if (is_block_text_element(tagName)) {
                     // Find closing tag
-                    var closeMarkup = "</" + tagName + ">";
+                    var closeMarkup = "</".concat(tagName, ">");
                     var closeIndex = lowerCaseTagName
                         ? data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex)
                         : data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
@@ -1356,18 +1446,25 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
             if (last.parentNode && last.parentNode.parentNode) {
                 if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
                     // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
-                    oneBefore.removeChild(last);
-                    last.childNodes.forEach(function (child) {
-                        oneBefore.parentNode.appendChild(child);
-                    });
-                    stack.pop();
+                    // this is wrong, becouse this will put the H3 outside the current right position which should be inside the current Html Element, see issue 152 for more info
+                    if (options.parseNoneClosedTags !== true) {
+                        oneBefore.removeChild(last);
+                        last.childNodes.forEach(function (child) {
+                            oneBefore.parentNode.appendChild(child);
+                        });
+                        stack.pop();
+                    }
                 }
                 else {
                     // Single error  <div> <h3> </div> handle: Just removes <h3>
-                    oneBefore.removeChild(last);
-                    last.childNodes.forEach(function (child) {
-                        oneBefore.appendChild(child);
-                    });
+                    // Why remove? this is already a HtmlElement and the missing <H3> is already added in this case. see issue 152 for more info
+                    // eslint-disable-next-line no-lonely-if
+                    if (options.parseNoneClosedTags !== true) {
+                        oneBefore.removeChild(last);
+                        last.childNodes.forEach(function (child) {
+                            oneBefore.appendChild(child);
+                        });
+                    }
                 }
             }
             else {
@@ -1386,45 +1483,6 @@ define("nodes/html", ["require", "exports", "he", "css-select", "nodes/node", "n
     }
     exports.parse = parse;
 });
-define("nodes/node", ["require", "exports", "he"], function (require, exports, he_3) {
-    "use strict";
-    Object.defineProperty(exports, "__esModule", { value: true });
-    /**
-     * Node Class as base class for TextNode and HTMLElement.
-     */
-    var Node = /** @class */ (function () {
-        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 () {
-                return this.rawText;
-            },
-            enumerable: false,
-            configurable: true
-        });
-        Object.defineProperty(Node.prototype, "textContent", {
-            get: function () {
-                return (0, he_3.decode)(this.rawText);
-            },
-            set: function (val) {
-                this.rawText = (0, he_3.encode)(val);
-            },
-            enumerable: false,
-            configurable: true
-        });
-        return Node;
-    }());
-    exports.default = Node;
-});
 define("nodes/comment", ["require", "exports", "nodes/node", "nodes/type"], function (require, exports, node_3, type_4) {
     "use strict";
     Object.defineProperty(exports, "__esModule", { value: true });
@@ -1442,6 +1500,9 @@ define("nodes/comment", ["require", "exports", "nodes/node", "nodes/type"], func
             _this.nodeType = type_4.default.COMMENT_NODE;
             return _this;
         }
+        CommentNode.prototype.clone = function () {
+            return new CommentNode(this.rawText, null);
+        };
         Object.defineProperty(CommentNode.prototype, "text", {
             /**
              * Get unescaped text value of current node and its children.
@@ -1454,7 +1515,7 @@ define("nodes/comment", ["require", "exports", "nodes/node", "nodes/type"], func
             configurable: true
         });
         CommentNode.prototype.toString = function () {
-            return "<!--" + this.rawText + "-->";
+            return "<!--".concat(this.rawText, "-->");
         };
         return CommentNode;
     }(node_3.default));

package/dist/nodes/comment.d.ts

@@ -1,8 +1,9 @@
+import HTMLElement from './html';
 import Node from './node';
 import NodeType from './type';
-import HTMLElement from './html';
 export default class CommentNode extends Node {
     rawText: string;
+    clone(): CommentNode;
     constructor(rawText: string, parentNode: HTMLElement, range?: [number, number]);
     /**
      * Node Type declaration.

package/dist/nodes/comment.js

@@ -32,6 +32,9 @@ var CommentNode = /** @class */ (function (_super) {
         _this.nodeType = type_1.default.COMMENT_NODE;
         return _this;
     }
+    CommentNode.prototype.clone = function () {
+        return new CommentNode(this.rawText, null);
+    };
     Object.defineProperty(CommentNode.prototype, "text", {
         /**
          * Get unescaped text value of current node and its children.
@@ -44,7 +47,7 @@ var CommentNode = /** @class */ (function (_super) {
         configurable: true
     });
     CommentNode.prototype.toString = function () {
-        return "<!--" + this.rawText + "-->";
+        return "<!--".concat(this.rawText, "-->");
     };
     return CommentNode;
 }(node_1.default));

package/dist/nodes/html.d.ts

@@ -76,6 +76,7 @@ export default class HTMLElement extends Node {
      */
     exchangeChild(oldNode: Node, newNode: Node): void;
     get tagName(): string;
+    set tagName(newname: string);
     get localName(): string;
     get isVoidElement(): boolean;
     /**
@@ -188,11 +189,18 @@ export default class HTMLElement extends Node {
     insertAdjacentHTML(where: InsertPosition, html: string): void;
     get nextSibling(): Node;
     get nextElementSibling(): HTMLElement;
+    get previousSibling(): Node;
+    get previousElementSibling(): HTMLElement;
     get classNames(): string;
+    /**
+     * Clone this Node
+     */
+    clone(): Node;
 }
 export interface Options {
     lowerCaseTagName: boolean;
     comment: boolean;
+    parseNoneClosedTags?: boolean;
     blockTextElements: {
         [tag: string]: boolean;
     };

package/dist/nodes/html.js

@@ -39,14 +39,14 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
 };
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parse = exports.base_parse = void 0;
-var he_1 = __importDefault(require("he"));
 var css_select_1 = require("css-select");
-var node_1 = __importDefault(require("./node"));
-var type_1 = __importDefault(require("./type"));
-var text_1 = __importDefault(require("./text"));
-var matcher_1 = __importDefault(require("../matcher"));
+var he_1 = __importDefault(require("he"));
 var back_1 = __importDefault(require("../back"));
+var matcher_1 = __importDefault(require("../matcher"));
 var comment_1 = __importDefault(require("./comment"));
+var node_1 = __importDefault(require("./node"));
+var text_1 = __importDefault(require("./text"));
+var type_1 = __importDefault(require("./type"));
 var voidTags = new Set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']);
 function decode(val) {
     // clone string
@@ -86,7 +86,7 @@ var DOMTokenList = /** @class */ (function () {
     }
     DOMTokenList.prototype._validate = function (c) {
         if (/\s/.test(c)) {
-            throw new Error("DOMException in DOMTokenList.add: The token '" + c + "' contains HTML space characters, which are not valid in tokens.");
+            throw new Error("DOMException in DOMTokenList.add: The token '".concat(c, "' contains HTML space characters, which are not valid in tokens."));
         }
     };
     DOMTokenList.prototype.add = function (c) {
@@ -170,14 +170,14 @@ var HTMLElement = /** @class */ (function (_super) {
         );
         if (keyAttrs.id) {
             if (!rawAttrs) {
-                _this.rawAttrs = "id=\"" + keyAttrs.id + "\"";
+                _this.rawAttrs = "id=\"".concat(keyAttrs.id, "\"");
             }
         }
         if (keyAttrs.class) {
             if (!rawAttrs) {
-                var cls = "class=\"" + _this.classList.toString() + "\"";
+                var cls = "class=\"".concat(_this.classList.toString(), "\"");
                 if (_this.rawAttrs) {
-                    _this.rawAttrs += " " + cls;
+                    _this.rawAttrs += " ".concat(cls);
                 }
                 else {
                     _this.rawAttrs = cls;
@@ -236,6 +236,9 @@ var HTMLElement = /** @class */ (function (_super) {
         get: function () {
             return this.rawTagName ? this.rawTagName.toUpperCase() : this.rawTagName;
         },
+        set: function (newname) {
+            this.rawTagName = newname.toLowerCase();
+        },
         enumerable: false,
         configurable: true
     });
@@ -319,7 +322,7 @@ var HTMLElement = /** @class */ (function (_super) {
                     else {
                         var text = node.trimmedText;
                         if (currentBlock.prependWhitespace) {
-                            text = " " + text;
+                            text = " ".concat(text);
                             currentBlock.prependWhitespace = false;
                         }
                         currentBlock.push(text);
@@ -340,8 +343,8 @@ var HTMLElement = /** @class */ (function (_super) {
     HTMLElement.prototype.toString = function () {
         var tag = this.rawTagName;
         if (tag) {
-            var attrs = this.rawAttrs ? " " + this.rawAttrs : '';
-            return this.isVoidElement ? "<" + tag + attrs + ">" : "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
+            var attrs = this.rawAttrs ? " ".concat(this.rawAttrs) : '';
+            return this.isVoidElement ? "<".concat(tag).concat(attrs, ">") : "<".concat(tag).concat(attrs, ">").concat(this.innerHTML, "</").concat(tag, ">");
         }
         return this.innerHTML;
     };
@@ -437,9 +440,9 @@ var HTMLElement = /** @class */ (function (_super) {
                 res.push('  '.repeat(indention) + str);
             }
             function dfs(node) {
-                var idStr = node.id ? "#" + node.id : '';
-                var classStr = node.classList.length ? "." + node.classList.value.join('.') : ''; // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-member-access, @typescript-eslint/restrict-template-expressions, @typescript-eslint/no-unsafe-call
-                write("" + node.rawTagName + idStr + classStr);
+                var idStr = node.id ? "#".concat(node.id) : '';
+                var classStr = node.classList.length ? ".".concat(node.classList.value.join('.')) : ''; // eslint-disable-line @typescript-eslint/no-unsafe-member-access, @typescript-eslint/no-unsafe-member-access, @typescript-eslint/restrict-template-expressions, @typescript-eslint/no-unsafe-call
+                write("".concat(node.rawTagName).concat(idStr).concat(classStr));
                 indention++;
                 node.childNodes.forEach(function (childNode) {
                     if (childNode.nodeType === type_1.default.ELEMENT_NODE) {
@@ -698,7 +701,7 @@ var HTMLElement = /** @class */ (function (_super) {
             if (val === undefined || val === 'null') {
                 return name;
             }
-            return name + "=" + val;
+            return "".concat(name, "=").concat(val);
         })
             .join(' ');
         // Update this.id
@@ -745,7 +748,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var val = _this.quoteAttribute(attrs[name]);
             if (val === 'null' || val === '""')
                 return name;
-            return name + "=" + val;
+            return "".concat(name, "=").concat(val);
         })
             .join(' ');
         // Update this.id
@@ -773,7 +776,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var val = attributes[name];
             if (val === 'null' || val === '""')
                 return name;
-            return name + "=" + _this.quoteAttribute(String(val));
+            return "".concat(name, "=").concat(_this.quoteAttribute(String(val)));
         })
             .join(' ');
     };
@@ -815,7 +818,7 @@ var HTMLElement = /** @class */ (function (_super) {
             });
         }
         else {
-            throw new Error("The value provided ('" + where + "') is not one of 'beforebegin', 'afterbegin', 'beforeend', or 'afterend'");
+            throw new Error("The value provided ('".concat(where, "') is not one of 'beforebegin', 'afterbegin', 'beforeend', or 'afterend'"));
         }
         // if (!where || html === undefined || html === null) {
         // 	return;
@@ -860,6 +863,45 @@ var HTMLElement = /** @class */ (function (_super) {
         enumerable: false,
         configurable: true
     });
+    Object.defineProperty(HTMLElement.prototype, "previousSibling", {
+        get: function () {
+            if (this.parentNode) {
+                var children = this.parentNode.childNodes;
+                var i = children.length;
+                while (i > 0) {
+                    var child = children[--i];
+                    if (this === child)
+                        return children[i - 1] || null;
+                }
+                return null;
+            }
+        },
+        enumerable: false,
+        configurable: true
+    });
+    Object.defineProperty(HTMLElement.prototype, "previousElementSibling", {
+        get: function () {
+            if (this.parentNode) {
+                var children = this.parentNode.childNodes;
+                var i = children.length;
+                var find = false;
+                while (i > 0) {
+                    var child = children[--i];
+                    if (find) {
+                        if (child instanceof HTMLElement) {
+                            return child || null;
+                        }
+                    }
+                    else if (this === child) {
+                        find = true;
+                    }
+                }
+                return null;
+            }
+        },
+        enumerable: false,
+        configurable: true
+    });
     Object.defineProperty(HTMLElement.prototype, "classNames", {
         get: function () {
             return this.classList.toString();
@@ -867,6 +909,12 @@ var HTMLElement = /** @class */ (function (_super) {
         enumerable: false,
         configurable: true
     });
+    /**
+     * Clone this Node
+     */
+    HTMLElement.prototype.clone = function () {
+        return parse(this.toString()).firstChild;
+    };
     return HTMLElement;
 }(node_1.default));
 exports.default = HTMLElement;
@@ -959,8 +1007,8 @@ function base_parse(data, options) {
         pre: true,
     };
     var element_names = Object.keys(elements);
-    var kBlockTextElements = element_names.map(function (it) { return new RegExp("^" + it + "$", 'i'); });
-    var kIgnoreElements = element_names.filter(function (it) { return elements[it]; }).map(function (it) { return new RegExp("^" + it + "$", 'i'); });
+    var kBlockTextElements = element_names.map(function (it) { return new RegExp("^".concat(it, "$"), 'i'); });
+    var kIgnoreElements = element_names.filter(function (it) { return elements[it]; }).map(function (it) { return new RegExp("^".concat(it, "$"), 'i'); });
     function element_should_be_ignore(tag) {
         return kIgnoreElements.some(function (it) { return it.test(tag); });
     }
@@ -975,7 +1023,7 @@ function base_parse(data, options) {
     var noNestedTagIndex = undefined;
     var match;
     // https://github.com/taoqf/node-html-parser/issues/38
-    data = "<" + frameflag + ">" + data + "</" + frameflag + ">";
+    data = "<".concat(frameflag, ">").concat(data, "</").concat(frameflag, ">");
     var lowerCaseTagName = options.lowerCaseTagName;
     var dataEndPos = data.length - (frameflag.length + 2);
     var frameFlagOffset = frameflag.length + 2;
@@ -1043,7 +1091,7 @@ function base_parse(data, options) {
             stack.push(currentParent);
             if (is_block_text_element(tagName)) {
                 // Find closing tag
-                var closeMarkup = "</" + tagName + ">";
+                var closeMarkup = "</".concat(tagName, ">");
                 var closeIndex = lowerCaseTagName
                     ? data.toLocaleLowerCase().indexOf(closeMarkup, kMarkupPattern.lastIndex)
                     : data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
@@ -1110,18 +1158,25 @@ function parse(data, options) {
         if (last.parentNode && last.parentNode.parentNode) {
             if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
                 // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
-                oneBefore.removeChild(last);
-                last.childNodes.forEach(function (child) {
-                    oneBefore.parentNode.appendChild(child);
-                });
-                stack.pop();
+                // this is wrong, becouse this will put the H3 outside the current right position which should be inside the current Html Element, see issue 152 for more info
+                if (options.parseNoneClosedTags !== true) {
+                    oneBefore.removeChild(last);
+                    last.childNodes.forEach(function (child) {
+                        oneBefore.parentNode.appendChild(child);
+                    });
+                    stack.pop();
+                }
             }
             else {
                 // Single error  <div> <h3> </div> handle: Just removes <h3>
-                oneBefore.removeChild(last);
-                last.childNodes.forEach(function (child) {
-                    oneBefore.appendChild(child);
-                });
+                // Why remove? this is already a HtmlElement and the missing <H3> is already added in this case. see issue 152 for more info
+                // eslint-disable-next-line no-lonely-if
+                if (options.parseNoneClosedTags !== true) {
+                    oneBefore.removeChild(last);
+                    last.childNodes.forEach(function (child) {
+                        oneBefore.appendChild(child);
+                    });
+                }
             }
         }
         else {

package/dist/nodes/node.d.ts

@@ -11,6 +11,7 @@ export default abstract class Node {
     abstract text: string;
     abstract rawText: string;
     abstract toString(): string;
+    abstract clone(): Node;
     constructor(parentNode?: HTMLElement, range?: [number, number]);
     get innerText(): string;
     get textContent(): string;

package/dist/nodes/text.d.ts

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

package/dist/nodes/text.js

@@ -37,6 +37,9 @@ var TextNode = /** @class */ (function (_super) {
         _this._rawText = rawText;
         return _this;
     }
+    TextNode.prototype.clone = function () {
+        return new TextNode(this._rawText, null);
+    };
     Object.defineProperty(TextNode.prototype, "rawText", {
         get: function () {
             return this._rawText;

package/package.json

@@ -1,9 +1,14 @@
 {
 	"name": "node-html-parser",
-	"version": "5.1.0",
+	"version": "5.2.5",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
 	"main": "dist/index.js",
 	"types": "dist/index.d.ts",
+	"exports": {
+		"require": "./dist/index.js",
+		"import": "./esm/index.js",
+		"types": "./dist/index.d.ts"
+	},
 	"scripts": {
 		"compile": "tsc",
 		"build": "npm run lint && npm run clean && npm run compile:cjs && npm run compile:amd",
@@ -48,7 +53,7 @@
 		"registry": "https://registry.npmjs.org"
 	},
 	"dependencies": {
-		"css-select": "^4.1.3",
+		"css-select": "^4.2.1",
 		"he": "1.2.0"
 	},
 	"devDependencies": {
@@ -60,25 +65,31 @@
 		"@typescript-eslint/parser": "latest",
 		"blanket": "latest",
 		"cheerio": "^1.0.0-rc.5",
-		"rimraf": "^3.0.2",
+		"cross-env": "^7.0.3",
 		"eslint": "^7.32.0",
 		"eslint-config-prettier": "latest",
 		"eslint-plugin-import": "latest",
 		"high5": "^1.0.0",
+		"html-dom-parser": "^1.0.4",
+		"html-parser": "^0.11.0",
+		"html5parser": "^2.0.2",
+		"htmljs-parser": "^2.11.1",
 		"htmlparser": "^1.7.7",
 		"htmlparser-benchmark": "^1.1.3",
 		"htmlparser2": "^6.0.0",
 		"mocha": "latest",
 		"mocha-each": "^2.0.1",
+		"neutron-html5parser": "^0.2.0",
 		"np": "latest",
 		"parse5": "^6.0.1",
+		"rimraf": "^3.0.2",
+		"saxes": "^6.0.0",
 		"should": "latest",
 		"spec": "latest",
 		"standard-version": "^9.3.1",
 		"travis-cov": "latest",
 		"ts-node": "^10.2.1",
-		"typescript": "latest",
-		"cross-env": "^7.0.3"
+		"typescript": "latest"
 	},
 	"config": {
 		"blanket": {
@@ -102,9 +113,5 @@
 		"url": "https://github.com/taoqf/node-fast-html-parser/issues"
 	},
 	"homepage": "https://github.com/taoqf/node-fast-html-parser",
-	"sideEffects": false,
-	"exports": {
-		"require": "./dist/index.js",
-		"import": "./esm/index.js"
-	}
+	"sideEffects": false
 }