node-html-parser - npm Package Compare versions

Comparing version 5.2.0 to 5.2.2

dist/main.js

@@ -0,1 +1,4 @@
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
 var __extends = (this && this.__extends) || (function () {
@@ -16,5 +19,2 @@ var extendStatics = function (d, b) {
 })();
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
 var __assign = (this && this.__assign) || function () {
@@ -59,82 +59,33 @@ __assign = Object.assign || function(t) {
 });
-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", {
+        Object.defineProperty(Node.prototype, "innerText", {
             get: function () {
-                return this._rawText;
+                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
-             */
+        Object.defineProperty(Node.prototype, "textContent", {
             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_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| )*$/.test(this.rawText);
+            set: function (val) {
+                this.rawText = (0, he_1.encode)(val);
             },
@@ -144,46 +95,12 @@ enumerable: false,
         });
-        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;
     }
@@ -288,17 +205,142 @@ function getAttributeValue(elem, name) {
 });
-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);
+    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);
-    text_1 = __importDefault(text_1);
-    matcher_1 = __importDefault(matcher_1);
-    back_1 = __importDefault(back_1);
-    comment_1 = __importDefault(comment_1);
     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)));
     }
@@ -1116,2 +1158,41 @@ // https://developer.mozilla.org/en-US/docs/Web/HTML/Block-level_elements
         });
+        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", {
@@ -1124,2 +1205,8 @@ get: function () {
         });
+        /**
+         * Clone this Node
+         */
+        HTMLElement.prototype.clone = function () {
+            return parse(this.toString()).firstChild;
+        };
         return HTMLElement;
@@ -1368,14 +1455,21 @@ }(node_2.default));
                     // 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);
+                        });
+                    }
                 }
@@ -1399,41 +1493,2 @@ }
 });
-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) {
@@ -1456,2 +1511,5 @@ "use strict";
         }
+        CommentNode.prototype.clone = function () {
+            return new CommentNode(this.rawText, null);
+        };
         Object.defineProperty(CommentNode.prototype, "text", {
@@ -1458,0 +1516,0 @@ /**

dist/nodes/comment.d.ts

@@ -0,6 +1,7 @@
+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]);
@@ -7,0 +8,0 @@ /**

dist/nodes/comment.js

@@ -35,2 +35,5 @@ "use strict";
     }
+    CommentNode.prototype.clone = function () {
+        return new CommentNode(this.rawText, null);
+    };
     Object.defineProperty(CommentNode.prototype, "text", {
@@ -37,0 +40,0 @@ /**

dist/nodes/html.d.ts

@@ -192,3 +192,9 @@ import Node from './node';
     get nextElementSibling(): HTMLElement;
+    get previousSibling(): Node;
+    get previousElementSibling(): HTMLElement;
     get classNames(): string;
+    /**
+     * Clone this Node
+     */
+    clone(): Node;
 }
@@ -198,2 +204,3 @@ export interface Options {
     comment: boolean;
+    parseNoneClosedTags?: boolean;
     blockTextElements: {
@@ -200,0 +207,0 @@ [tag: string]: boolean;

dist/nodes/html.js

@@ -42,10 +42,10 @@ "use strict";
 exports.parse = exports.base_parse = void 0;
+var css_select_1 = require("css-select");
 var he_1 = __importDefault(require("he"));
-var css_select_1 = require("css-select");
+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 text_1 = __importDefault(require("./text"));
-var matcher_1 = __importDefault(require("../matcher"));
-var back_1 = __importDefault(require("../back"));
-var comment_1 = __importDefault(require("./comment"));
 var voidTags = new Set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr']);
@@ -867,2 +867,41 @@ function decode(val) {
     });
+    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", {
@@ -875,2 +914,8 @@ get: function () {
     });
+    /**
+     * Clone this Node
+     */
+    HTMLElement.prototype.clone = function () {
+        return parse(this.toString()).firstChild;
+    };
     return HTMLElement;
@@ -1119,14 +1164,21 @@ }(node_1.default));
                 // 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);
+                    });
+                }
             }
@@ -1133,0 +1185,0 @@ }

dist/nodes/node.d.ts

@@ -14,2 +14,3 @@ import NodeType from './type';
     abstract toString(): string;
+    abstract clone(): Node;
     constructor(parentNode?: HTMLElement, range?: [number, number]);
@@ -16,0 +17,0 @@ get innerText(): string;

dist/nodes/text.d.ts

@@ -9,2 +9,3 @@ import HTMLElement from './html';
 export default class TextNode extends Node {
+    clone(): TextNode;
     constructor(rawText: string, parentNode: HTMLElement, range?: [number, number]);
@@ -11,0 +12,0 @@ /**

dist/nodes/text.js

@@ -40,2 +40,5 @@ "use strict";
     }
+    TextNode.prototype.clone = function () {
+        return new TextNode(this._rawText, null);
+    };
     Object.defineProperty(TextNode.prototype, "rawText", {
@@ -42,0 +45,0 @@ get: function () {

package.json

 {
 	"name": "node-html-parser",
-	"version": "5.2.0",
+	"version": "5.2.2",
 	"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": {
@@ -51,3 +56,3 @@ "compile": "tsc",
 	"dependencies": {
-		"css-select": "^4.1.3",
+		"css-select": "^4.2.1",
 		"he": "1.2.0"
@@ -113,7 +118,3 @@ },
 	"homepage": "https://github.com/taoqf/node-fast-html-parser",
-	"sideEffects": false,
-	"exports": {
-		"require": "./dist/index.js",
-		"import": "./esm/index.js"
-	}
+	"sideEffects": false
 }

README.md

@@ -202,2 +202,6 @@ # Fast HTML Parser [![NPM version](https://badge.fury.io/js/node-html-parser.png)](http://badge.fury.io/js/node-html-parser) [![Build Status](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Factions-badge.atrox.dev%2Ftaoqf%2Fnode-html-parser%2Fbadge%3Fref%3Dmain&style=flat)](https://actions-badge.atrox.dev/taoqf/node-html-parser/goto?ref=main)
 
+#### Node#clone()
+
+Clone a node.
+
 ## HTMLElement Properties
@@ -251,2 +255,10 @@
 
+### 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
@@ -253,0 +265,0 @@

No alert changes

Improved metrics

Total package byte prevSize

146722

increased by3.72%

Lines of code

3484

increased by3.75%

Number of lines in readme file

277

increased by4.53%

Dependency changes

• Updatedcss-select@^4.2.1