node-html-parser 2.0.1 → 2.2.1

package/README.md

@@ -10,20 +10,24 @@ no closing `<li>`, `<td>` etc).
 
 ## Install
 
+
 ```shell
 npm install --save node-html-parser
 ```
 
+> Note: when using Fast HTML Parser in a Typescript project the minimum Typescript version supported is `^4.1.2`.
+
 ## Performance
 
 Faster than htmlparser2!
 
 ```shell
-node-html-parser:2.02346 ms/file ± 2.21481
-htmlparser      :26.0810 ms/file ± 171.313
-htmlparser2     :4.49111 ms/file ± 6.85512
-parse5          :14.8590 ms/file ± 10.9427
-high5           :7.71818 ms/file ± 4.88375
+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
 ```
 
 Tested with [htmlparser-benchmark](https://github.com/AndreasMadsen/htmlparser-benchmark).
@@ -162,6 +166,10 @@ Get unescaped text value of current node and its children. Like `innerText`.
 Get escpaed (as-it) 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.
+
 ### HTMLElement#structuredText
 
 Get structured Text
@@ -193,3 +201,7 @@ Returns a reference to the next child node of the current element's parent.
 ### HTMLElement#nextElementSibling
 
 Returns a reference to the next 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/esm/nodes/html.js

@@ -1,4 +1,4 @@
-import { decode } from 'he';
+import he from 'he';
 import Node from './node';
 import NodeType from './type';
 import TextNode from './text';
@@ -6,6 +6,7 @@ import Matcher from '../matcher';
 import arr_back from '../back';
 import CommentNode from './comment';
 import parse from '../parse';
+const { decode } = he;
 const kBlockElements = new Map();
 kBlockElements.set('DIV', true);
 kBlockElements.set('div', true);
@@ -117,6 +118,13 @@ export default class HTMLElement extends Node {
             return (pre += cur.rawText);
         }, '');
     }
+    get textContent() {
+        return this.rawText;
+    }
+    set textContent(val) {
+        const content = [new TextNode(val)];
+        this.childNodes = content;
+    }
     /**
      * Get unescaped text value of current node and its children.
      * @return {string} text content
@@ -233,7 +241,7 @@ export default class HTMLElement extends Node {
         function dfs(node) {
             const idStr = node.id ? (`#${node.id}`) : '';
             const classStr = node.classNames.length ? (`.${node.classNames.join('.')}`) : '';
-            write(node.rawTagName + idStr + classStr);
+            write(`${node.rawTagName}${idStr}${classStr}`);
             indention++;
             node.childNodes.forEach((childNode) => {
                 if (childNode.nodeType === NodeType.ELEMENT_NODE) {
@@ -409,9 +417,10 @@ export default class HTMLElement extends Node {
     }
     /**
      * Get attributes
+     * @access private
      * @return {Object} parsed and unescaped attributes
      */
-    get attributes() {
+    get attrs() {
         if (this._attrs) {
             return this._attrs;
         }
@@ -419,10 +428,19 @@ export default class HTMLElement extends Node {
         const attrs = this.rawAttributes;
         for (const key in attrs) {
             const val = attrs[key] || '';
-            this._attrs[key] = decode(val);
+            this._attrs[key.toLowerCase()] = decode(val);
         }
         return this._attrs;
     }
+    get attributes() {
+        const ret_attrs = {};
+        const attrs = this.rawAttributes;
+        for (const key in attrs) {
+            const val = attrs[key] || '';
+            ret_attrs[key] = decode(val);
+        }
+        return ret_attrs;
+    }
     /**
      * Get escaped (as-it) attributes
      * @return {Object} parsed attributes
@@ -459,14 +477,14 @@ export default class HTMLElement extends Node {
         }).join(' ');
     }
     hasAttribute(key) {
-        return key in this.attributes;
+        return key.toLowerCase() in this.attrs;
     }
     /**
      * Get an attribute
      * @return {string} value of the attribute
      */
     getAttribute(key) {
-        return this.attributes[key];
+        return this.attrs[key.toLowerCase()];
     }
     /**
      * Set an attribute value to the HTMLElement
@@ -477,10 +495,18 @@ export default class HTMLElement extends Node {
         if (arguments.length < 2) {
             throw new Error('Failed to execute \'setAttribute\' on \'Element\'');
         }
+        const k2 = key.toLowerCase();
         const attrs = this.rawAttributes;
+        for (const k in attrs) {
+            if (k.toLowerCase() === k2) {
+                key = k;
+                break;
+            }
+        }
         attrs[key] = String(value);
+        // update this.attrs
         if (this._attrs) {
-            this._attrs[key] = decode(attrs[key]);
+            this._attrs[k2] = decode(attrs[key]);
         }
         // Update rawString
         this.rawAttrs = Object.keys(attrs).map((name) => {

package/dist/esm/nodes/node.js

@@ -8,4 +8,10 @@ export default class Node {
     get innerText() {
         return this.rawText;
     }
+    get textContent() {
+        return this.rawText;
+    }
+    set textContent(val) {
+        this.rawText = val;
+    }
 }

package/dist/main.js

@@ -6,6 +6,8 @@ var __extends = (this && this.__extends) || (function () {
         return extendStatics(d, b);
     };
     return function (d, b) {
+        if (typeof b !== "function" && b !== null)
+            throw new TypeError("Class extends value " + String(b) + " is not a constructor or null");
         extendStatics(d, b);
         function __() { this.constructor = d; }
         d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());
@@ -14,12 +16,10 @@ var __extends = (this && this.__extends) || (function () {
 var __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
 };
-var __spreadArrays = (this && this.__spreadArrays) || function () {
-    for (var s = 0, i = 0, il = arguments.length; i < il; i++) s += arguments[i].length;
-    for (var r = Array(s), k = 0, i = 0; i < il; i++)
-        for (var a = arguments[i], j = 0, jl = a.length; j < jl; j++, k++)
-            r[k] = a[j];
-    return r;
+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;
 };
 define("back", ["require", "exports"], function (require, exports) {
     "use strict";
@@ -57,6 +57,16 @@ define("nodes/node", ["require", "exports"], function (require, exports) {
             enumerable: false,
             configurable: true
         });
+        Object.defineProperty(Node.prototype, "textContent", {
+            get: function () {
+                return this.rawText;
+            },
+            set: function (val) {
+                this.rawText = val;
+            },
+            enumerable: false,
+            configurable: true
+        });
         return Node;
     }());
     exports.default = Node;
@@ -450,6 +460,7 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
     "use strict";
     Object.defineProperty(exports, "__esModule", { value: true });
     exports.base_parse = void 0;
+    he_1 = __importDefault(he_1);
     node_3 = __importDefault(node_3);
     type_3 = __importDefault(type_3);
     text_1 = __importDefault(text_1);
@@ -457,6 +468,7 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
     back_2 = __importDefault(back_2);
     comment_1 = __importDefault(comment_1);
     parse_1 = __importDefault(parse_1);
+    var decode = he_1.default.decode;
     var kBlockElements = new Map();
     kBlockElements.set('DIV', true);
     kBlockElements.set('div', true);
@@ -581,13 +593,24 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
             enumerable: false,
             configurable: true
         });
+        Object.defineProperty(HTMLElement.prototype, "textContent", {
+            get: function () {
+                return this.rawText;
+            },
+            set: function (val) {
+                var content = [new text_1.default(val)];
+                this.childNodes = content;
+            },
+            enumerable: false,
+            configurable: true
+        });
         Object.defineProperty(HTMLElement.prototype, "text", {
             /**
              * Get unescaped text value of current node and its children.
              * @return {string} text content
              */
             get: function () {
-                return he_1.decode(this.rawText);
+                return decode(this.rawText);
             },
             enumerable: false,
             configurable: true
@@ -715,7 +738,7 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
                 function dfs(node) {
                     var idStr = node.id ? ("#" + node.id) : '';
                     var classStr = node.classNames.length ? ("." + node.classNames.join('.')) : '';
-                    write(node.rawTagName + idStr + classStr);
+                    write("" + node.rawTagName + idStr + classStr);
                     indention++;
                     node.childNodes.forEach(function (childNode) {
                         if (childNode.nodeType === type_3.default.ELEMENT_NODE) {
@@ -903,9 +926,10 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
             enumerable: false,
             configurable: true
         });
-        Object.defineProperty(HTMLElement.prototype, "attributes", {
+        Object.defineProperty(HTMLElement.prototype, "attrs", {
             /**
              * Get attributes
+             * @access private
              * @return {Object} parsed and unescaped attributes
              */
             get: function () {
@@ -916,13 +940,26 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
                 var attrs = this.rawAttributes;
                 for (var key in attrs) {
                     var val = attrs[key] || '';
-                    this._attrs[key] = he_1.decode(val);
+                    this._attrs[key.toLowerCase()] = decode(val);
                 }
                 return this._attrs;
             },
             enumerable: false,
             configurable: true
         });
+        Object.defineProperty(HTMLElement.prototype, "attributes", {
+            get: function () {
+                var ret_attrs = {};
+                var attrs = this.rawAttributes;
+                for (var key in attrs) {
+                    var val = attrs[key] || '';
+                    ret_attrs[key] = decode(val);
+                }
+                return ret_attrs;
+            },
+            enumerable: false,
+            configurable: true
+        });
         Object.defineProperty(HTMLElement.prototype, "rawAttributes", {
             /**
              * Get escaped (as-it) attributes
@@ -963,14 +1000,14 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
             }).join(' ');
         };
         HTMLElement.prototype.hasAttribute = function (key) {
-            return key in this.attributes;
+            return key.toLowerCase() in this.attrs;
         };
         /**
          * Get an attribute
          * @return {string} value of the attribute
          */
         HTMLElement.prototype.getAttribute = function (key) {
-            return this.attributes[key];
+            return this.attrs[key.toLowerCase()];
         };
         /**
          * Set an attribute value to the HTMLElement
@@ -981,10 +1018,18 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
             if (arguments.length < 2) {
                 throw new Error('Failed to execute \'setAttribute\' on \'Element\'');
             }
+            var k2 = key.toLowerCase();
             var attrs = this.rawAttributes;
+            for (var k in attrs) {
+                if (k.toLowerCase() === k2) {
+                    key = k;
+                    break;
+                }
+            }
             attrs[key] = String(value);
+            // update this.attrs
             if (this._attrs) {
-                this._attrs[key] = he_1.decode(attrs[key]);
+                this._attrs[k2] = decode(attrs[key]);
             }
             // Update rawString
             this.rawAttrs = Object.keys(attrs).map(function (name) {
@@ -1028,7 +1073,7 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
                 var idx = this.parentNode.childNodes.findIndex(function (child) {
                     return child === _this;
                 });
-                (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArrays([idx + 1, 0], p.childNodes));
+                (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes));
                 p.childNodes.forEach(function (n) {
                     if (n instanceof HTMLElement) {
                         n.parentNode = _this.parentNode;
@@ -1047,7 +1092,7 @@ define("nodes/html", ["require", "exports", "he", "nodes/node", "nodes/type", "n
                 var idx = this.parentNode.childNodes.findIndex(function (child) {
                     return child === _this;
                 });
-                (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArrays([idx, 0], p.childNodes));
+                (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes));
                 p.childNodes.forEach(function (n) {
                     if (n instanceof HTMLElement) {
                         n.parentNode = _this.parentNode;

package/dist/nodes/comment.d.ts

@@ -13,5 +13,5 @@ export default class CommentNode extends Node {
      * @return {string} text content
      */
     get text(): string;
-    toString(): string;
+    toString(): `<!--${string}-->`;
 }

package/dist/nodes/comment.js

@@ -7,6 +7,8 @@ var __extends = (this && this.__extends) || (function () {
         return extendStatics(d, b);
     };
     return function (d, b) {
+        if (typeof b !== "function" && b !== null)
+            throw new TypeError("Class extends value " + String(b) + " is not a constructor or null");
         extendStatics(d, b);
         function __() { this.constructor = d; }
         d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());

package/dist/nodes/html.d.ts

@@ -62,6 +62,8 @@ export default class HTMLElement extends Node {
      * @return {string} text content
      */
     get rawText(): string;
+    get textContent(): string;
+    set textContent(val: string);
     /**
      * Get unescaped text value of current node and its children.
      * @return {string} text content
@@ -124,9 +126,11 @@ export default class HTMLElement extends Node {
     get lastChild(): Node;
     /**
      * Get attributes
+     * @access private
      * @return {Object} parsed and unescaped attributes
      */
-    get attributes(): Attributes;
+    get attrs(): Attributes;
+    get attributes(): Record<string, string>;
     /**
      * Get escaped (as-it) attributes
      * @return {Object} parsed attributes

package/dist/nodes/html.js

@@ -7,24 +7,24 @@ var __extends = (this && this.__extends) || (function () {
         return extendStatics(d, b);
     };
     return function (d, b) {
+        if (typeof b !== "function" && b !== null)
+            throw new TypeError("Class extends value " + String(b) + " is not a constructor or null");
         extendStatics(d, b);
         function __() { this.constructor = d; }
         d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());
     };
 })();
-var __spreadArrays = (this && this.__spreadArrays) || function () {
-    for (var s = 0, i = 0, il = arguments.length; i < il; i++) s += arguments[i].length;
-    for (var r = Array(s), k = 0, i = 0; i < il; i++)
-        for (var a = arguments[i], j = 0, jl = a.length; j < jl; j++, k++)
-            r[k] = a[j];
-    return r;
+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 __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
 };
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.base_parse = void 0;
-var he_1 = require("he");
+var he_1 = __importDefault(require("he"));
 var node_1 = __importDefault(require("./node"));
 var type_1 = __importDefault(require("./type"));
 var text_1 = __importDefault(require("./text"));
@@ -32,6 +32,7 @@ var matcher_1 = __importDefault(require("../matcher"));
 var back_1 = __importDefault(require("../back"));
 var comment_1 = __importDefault(require("./comment"));
 var parse_1 = __importDefault(require("../parse"));
+var decode = he_1.default.decode;
 var kBlockElements = new Map();
 kBlockElements.set('DIV', true);
 kBlockElements.set('div', true);
@@ -156,13 +157,24 @@ var HTMLElement = /** @class */ (function (_super) {
         enumerable: false,
         configurable: true
     });
+    Object.defineProperty(HTMLElement.prototype, "textContent", {
+        get: function () {
+            return this.rawText;
+        },
+        set: function (val) {
+            var content = [new text_1.default(val)];
+            this.childNodes = content;
+        },
+        enumerable: false,
+        configurable: true
+    });
     Object.defineProperty(HTMLElement.prototype, "text", {
         /**
          * Get unescaped text value of current node and its children.
          * @return {string} text content
          */
         get: function () {
-            return he_1.decode(this.rawText);
+            return decode(this.rawText);
         },
         enumerable: false,
         configurable: true
@@ -290,7 +302,7 @@ var HTMLElement = /** @class */ (function (_super) {
             function dfs(node) {
                 var idStr = node.id ? ("#" + node.id) : '';
                 var classStr = node.classNames.length ? ("." + node.classNames.join('.')) : '';
-                write(node.rawTagName + idStr + classStr);
+                write("" + node.rawTagName + idStr + classStr);
                 indention++;
                 node.childNodes.forEach(function (childNode) {
                     if (childNode.nodeType === type_1.default.ELEMENT_NODE) {
@@ -478,9 +490,10 @@ var HTMLElement = /** @class */ (function (_super) {
         enumerable: false,
         configurable: true
     });
-    Object.defineProperty(HTMLElement.prototype, "attributes", {
+    Object.defineProperty(HTMLElement.prototype, "attrs", {
         /**
          * Get attributes
+         * @access private
          * @return {Object} parsed and unescaped attributes
          */
         get: function () {
@@ -491,13 +504,26 @@ var HTMLElement = /** @class */ (function (_super) {
             var attrs = this.rawAttributes;
             for (var key in attrs) {
                 var val = attrs[key] || '';
-                this._attrs[key] = he_1.decode(val);
+                this._attrs[key.toLowerCase()] = decode(val);
             }
             return this._attrs;
         },
         enumerable: false,
         configurable: true
     });
+    Object.defineProperty(HTMLElement.prototype, "attributes", {
+        get: function () {
+            var ret_attrs = {};
+            var attrs = this.rawAttributes;
+            for (var key in attrs) {
+                var val = attrs[key] || '';
+                ret_attrs[key] = decode(val);
+            }
+            return ret_attrs;
+        },
+        enumerable: false,
+        configurable: true
+    });
     Object.defineProperty(HTMLElement.prototype, "rawAttributes", {
         /**
          * Get escaped (as-it) attributes
@@ -538,14 +564,14 @@ var HTMLElement = /** @class */ (function (_super) {
         }).join(' ');
     };
     HTMLElement.prototype.hasAttribute = function (key) {
-        return key in this.attributes;
+        return key.toLowerCase() in this.attrs;
     };
     /**
      * Get an attribute
      * @return {string} value of the attribute
      */
     HTMLElement.prototype.getAttribute = function (key) {
-        return this.attributes[key];
+        return this.attrs[key.toLowerCase()];
     };
     /**
      * Set an attribute value to the HTMLElement
@@ -556,10 +582,18 @@ var HTMLElement = /** @class */ (function (_super) {
         if (arguments.length < 2) {
             throw new Error('Failed to execute \'setAttribute\' on \'Element\'');
         }
+        var k2 = key.toLowerCase();
         var attrs = this.rawAttributes;
+        for (var k in attrs) {
+            if (k.toLowerCase() === k2) {
+                key = k;
+                break;
+            }
+        }
         attrs[key] = String(value);
+        // update this.attrs
         if (this._attrs) {
-            this._attrs[key] = he_1.decode(attrs[key]);
+            this._attrs[k2] = decode(attrs[key]);
         }
         // Update rawString
         this.rawAttrs = Object.keys(attrs).map(function (name) {
@@ -603,7 +637,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
             });
-            (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArrays([idx + 1, 0], p.childNodes));
+            (_a = this.parentNode.childNodes).splice.apply(_a, __spreadArray([idx + 1, 0], p.childNodes));
             p.childNodes.forEach(function (n) {
                 if (n instanceof HTMLElement) {
                     n.parentNode = _this.parentNode;
@@ -622,7 +656,7 @@ var HTMLElement = /** @class */ (function (_super) {
             var idx = this.parentNode.childNodes.findIndex(function (child) {
                 return child === _this;
             });
-            (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArrays([idx, 0], p.childNodes));
+            (_c = this.parentNode.childNodes).splice.apply(_c, __spreadArray([idx, 0], p.childNodes));
             p.childNodes.forEach(function (n) {
                 if (n instanceof HTMLElement) {
                     n.parentNode = _this.parentNode;

package/dist/nodes/node.d.ts

@@ -9,4 +9,6 @@ export default abstract class Node {
     abstract rawText: string;
     abstract toString(): string;
     get innerText(): string;
+    get textContent(): string;
+    set textContent(val: string);
 }

package/dist/nodes/node.js

@@ -14,6 +14,16 @@ var Node = /** @class */ (function () {
         enumerable: false,
         configurable: true
     });
+    Object.defineProperty(Node.prototype, "textContent", {
+        get: function () {
+            return this.rawText;
+        },
+        set: function (val) {
+            this.rawText = val;
+        },
+        enumerable: false,
+        configurable: true
+    });
     return Node;
 }());
 exports.default = Node;

package/dist/nodes/text.js

@@ -7,6 +7,8 @@ var __extends = (this && this.__extends) || (function () {
         return extendStatics(d, b);
     };
     return function (d, b) {
+        if (typeof b !== "function" && b !== null)
+            throw new TypeError("Class extends value " + String(b) + " is not a constructor or null");
         extendStatics(d, b);
         function __() { this.constructor = d; }
         d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "node-html-parser",
-	"version": "2.0.1",
+	"version": "2.2.1",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
 	"main": "dist/index.js",
 	"module": "dist/esm/index.js",
@@ -38,6 +38,7 @@
 		"@typescript-eslint/eslint-plugin-tslint": "latest",
 		"@typescript-eslint/parser": "latest",
 		"blanket": "latest",
+		"cheerio": "^1.0.0-rc.5",
 		"del-cli": "latest",
 		"eslint": "latest",
 		"eslint-config-prettier": "latest",