node-html-parser - npm Package Compare versions

Comparing version 1.1.11 to 1.1.12

dist/index.d.ts

@@ -58,2 +58,5 @@ export declare enum NodeType {
 export declare class HTMLElement extends Node {
+    tagName: string;
+    private rawAttrs;
+    parentNode: Node;
     private _attrs;
@@ -63,7 +66,4 @@ private _rawAttrs;
     classNames: string[];
-    tagName: string;
-    rawAttrs: string;
     /**
      * Node Type declaration.
-     * @type {Number}
      */
@@ -73,10 +73,20 @@ nodeType: NodeType;
      * Creates an instance of HTMLElement.
-     * @param {string} name				tagName
-     * @param {KeyAttributes} keyAttrs	id and class attribute
-     * @param {string} [rawAttrs]	attributes in string
+     * @param keyAttrs	id and class attribute
+     * @param [rawAttrs]	attributes in string
      *
      * @memberof HTMLElement
      */
-    constructor(name: string, keyAttrs: KeyAttributes, rawAttrs?: string);
+    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs?: string, parentNode?: Node);
     /**
+     * Remove Child element from childNodes array
+     * @param {HTMLElement} node     node to remove
+     */
+    removeChild(node: Node): void;
+    /**
+     * Exchanges given child with new child
+     * @param {HTMLElement} oldNode     node to exchange
+     * @param {HTMLElement} newNode     new node
+     */
+    exchangeChild(oldNode: Node, newNode: Node): void;
+    /**
      * Get escpaed (as-it) text value of current node and its children.
@@ -122,3 +132,3 @@ * @return {string} text content
      */
-    querySelectorAll(selector: string | Matcher): Node[];
+    querySelectorAll(selector: string | Matcher): HTMLElement[];
     /**
@@ -130,3 +140,3 @@ * Query CSS Selector to find matching node.
      */
-    querySelector(selector: string | Matcher): Node;
+    querySelector(selector: string | Matcher): HTMLElement;
     /**
@@ -137,3 +147,3 @@ * Append a child node to childNodes
      */
-    appendChild(node: Node): Node;
+    appendChild<T extends Node = Node>(node: T): T;
     /**
@@ -207,3 +217,8 @@ * Get first child node
 export declare function parse(data: string, options?: {
-    lowerCaseTagName: boolean;
-}): HTMLElement;
+    lowerCaseTagName?: boolean;
+    noFix?: boolean;
+}): (TextNode & {
+    valid: boolean;
+}) | (HTMLElement & {
+    valid: boolean;
+});

dist/index.js

@@ -78,19 +78,19 @@ "use strict";
      * Creates an instance of HTMLElement.
-     * @param {string} name				tagName
-     * @param {KeyAttributes} keyAttrs	id and class attribute
-     * @param {string} [rawAttrs]	attributes in string
+     * @param keyAttrs	id and class attribute
+     * @param [rawAttrs]	attributes in string
      *
      * @memberof HTMLElement
      */
-    constructor(name, keyAttrs, rawAttrs) {
+    constructor(tagName, keyAttrs, rawAttrs = '', parentNode = null) {
         super();
+        this.tagName = tagName;
+        this.rawAttrs = rawAttrs;
+        this.parentNode = parentNode;
         this.classNames = [];
         /**
          * Node Type declaration.
-         * @type {Number}
          */
         this.nodeType = NodeType.ELEMENT_NODE;
-        this.tagName = name;
         this.rawAttrs = rawAttrs || '';
-        // this.parentNode = null;
+        this.parentNode = parentNode || null;
         this.childNodes = [];
@@ -105,2 +105,26 @@ if (keyAttrs.id) {
     /**
+     * Remove Child element from childNodes array
+     * @param {HTMLElement} node     node to remove
+     */
+    removeChild(node) {
+        this.childNodes = this.childNodes.filter((child) => {
+            return (child !== node);
+        });
+    }
+    /**
+     * Exchanges given child with new child
+     * @param {HTMLElement} oldNode     node to exchange
+     * @param {HTMLElement} newNode     new node
+     */
+    exchangeChild(oldNode, newNode) {
+        let idx = -1;
+        for (let i = 0; i < this.childNodes.length; i++) {
+            if (this.childNodes[i] === oldNode) {
+                idx = i;
+                break;
+            }
+        }
+        this.childNodes[idx] = newNode;
+    }
+    /**
      * Get escpaed (as-it) text value of current node and its children.
@@ -211,19 +235,16 @@ * @return {string} text content
     trimRight(pattern) {
-        function dfs(node) {
-            for (let i = 0; i < node.childNodes.length; i++) {
-                const childNode = node.childNodes[i];
-                if (childNode.nodeType === NodeType.ELEMENT_NODE) {
-                    dfs(childNode);
+        for (let i = 0; i < this.childNodes.length; i++) {
+            const childNode = this.childNodes[i];
+            if (childNode.nodeType === NodeType.ELEMENT_NODE) {
+                childNode.trimRight(pattern);
+            }
+            else {
+                const index = childNode.rawText.search(pattern);
+                if (index > -1) {
+                    childNode.rawText = childNode.rawText.substr(0, index);
+                    // trim all following nodes.
+                    this.childNodes.length = i + 1;
                 }
-                else {
-                    const index = childNode.rawText.search(pattern);
-                    if (index > -1) {
-                        childNode.rawText = childNode.rawText.substr(0, index);
-                        // trim all following nodes.
-                        node.childNodes.length = i + 1;
-                    }
-                }
             }
         }
-        dfs(this);
         return this;
@@ -385,2 +406,5 @@ }
         this.childNodes.push(node);
+        if (node instanceof HTMLElement) {
+            node.parentNode = this;
+        }
         return node;
@@ -444,3 +468,3 @@ }
  */
-let functionCache = {
+const functionCache = {
     "f145": function (el, tagName, classes, attr_key, value) {
@@ -454,3 +478,3 @@ "use strict";
             return false;
-        for (var cls = classes, i = 0; i < cls.length; i++)
+        for (let cls = classes, i = 0; i < cls.length; i++)
             if (el.classNames.indexOf(cls[i]) === -1)
@@ -466,3 +490,3 @@ return false;
         value = value || "";
-        for (var cls = classes, i = 0; i < cls.length; i++)
+        for (let cls = classes, i = 0; i < cls.length; i++)
             if (el.classNames.indexOf(cls[i]) === -1)
@@ -506,4 +530,4 @@ return false;
         value = value || "";
-        var attrs = el.attributes;
-        for (var key in attrs) {
+        let attrs = el.attributes;
+        for (let key in attrs) {
             const val = attrs[key];
@@ -515,3 +539,3 @@ if (key == attr_key && val == value) {
         return false;
-        // for (var cls = classes, i = 0; i < cls.length; i++) {if (el.classNames.indexOf(cls[i]) === -1){ return false;}}
+        // for (let cls = classes, i = 0; i < cls.length; i++) {if (el.classNames.indexOf(cls[i]) === -1){ return false;}}
         // return true;
@@ -525,4 +549,4 @@ },
         value = value || "";
-        var attrs = el.attributes;
-        for (var key in attrs) {
+        let attrs = el.attributes;
+        for (let key in attrs) {
             const val = attrs[key];
@@ -542,4 +566,4 @@ if (key == attr_key && val == value) {
         value = value || "";
-        var attrs = el.attributes;
-        for (var key in attrs) {
+        let attrs = el.attributes;
+        for (let key in attrs) {
             const val = attrs[key];
@@ -560,3 +584,3 @@ if (key == attr_key && val == value) {
             return false;
-        for (var cls = classes, i = 0; i < cls.length; i++)
+        for (let cls = classes, i = 0; i < cls.length; i++)
             if (el.classNames.indexOf(cls[i]) === -1)
@@ -627,3 +651,3 @@ return false;
                     value = matcher[7] || matcher[8];
-                    source += `var attrs = el.attributes;for (var key in attrs){const val = attrs[key]; if (key == "${attr_key}" && val == "${value}"){return true;}} return false;`; //2
+                    source += `let attrs = el.attributes;for (let key in attrs){const val = attrs[key]; if (key == "${attr_key}" && val == "${value}"){return true;}} return false;`; //2
                     function_name += '2';
@@ -637,3 +661,3 @@ }
             if (classes.length > 0) {
-                source += 'for (var cls = ' + JSON.stringify(classes) + ', i = 0; i < cls.length; i++) if (el.classNames.indexOf(cls[i]) === -1) return false;'; //4
+                source += 'for (let cls = ' + JSON.stringify(classes) + ', i = 0; i < cls.length; i++) if (el.classNames.indexOf(cls[i]) === -1) return false;'; //4
                 function_name += '4';
@@ -699,9 +723,11 @@ }
 const kSelfClosingElements = {
-    meta: true,
+    area: true,
+    base: true,
+    br: true,
+    col: true,
+    hr: true,
     img: true,
+    input: true,
     link: true,
-    input: true,
-    area: true,
-    br: true,
-    hr: true
+    meta: true
 };
@@ -711,4 +737,11 @@ const kElementsClosedByOpening = {
     p: { p: true, div: true },
+    b: { div: true },
     td: { td: true, th: true },
-    th: { td: true, th: true }
+    th: { td: true, th: true },
+    h1: { h1: true },
+    h2: { h2: true },
+    h3: { h3: true },
+    h4: { h4: true },
+    h5: { h5: true },
+    h6: { h6: true }
 };
@@ -760,6 +793,6 @@ const kElementsClosedByClosing = {
             // not </ tags
-            var attrs = {};
-            for (var attMatch; attMatch = kAttributePattern.exec(match[3]);)
+            let attrs = {};
+            for (let attMatch; attMatch = kAttributePattern.exec(match[3]);) {
                 attrs[attMatch[2]] = attMatch[4] || attMatch[5] || attMatch[6];
-            // console.log(attrs);
+            }
             if (!match[4] && kElementsClosedByOpening[currentParent.tagName]) {
@@ -775,4 +808,4 @@ if (kElementsClosedByOpening[currentParent.tagName][match[2]]) {
                 // a little test to find next </script> or </style> ...
-                var closeMarkup = '</' + match[2] + '>';
-                var index = data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
+                let closeMarkup = '</' + match[2] + '>';
+                let index = data.indexOf(closeMarkup, kMarkupPattern.lastIndex);
                 if (options[match[2]]) {
@@ -787,4 +820,5 @@ let text;
                     }
-                    if (text.length > 0)
+                    if (text.length > 0) {
                         currentParent.appendChild(new TextNode(text));
+                    }
                 }
@@ -824,4 +858,44 @@ if (index == -1) {
     }
-    return root;
+    const valid = !!(stack.length === 1);
+    if (!options.noFix) {
+        const response = root;
+        response.valid = valid;
+        while (stack.length > 1) {
+            // Handle each error elements.
+            const last = stack.pop();
+            const oneBefore = arr_back(stack);
+            if (last.parentNode && last.parentNode.parentNode) {
+                if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
+                    // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
+                    oneBefore.removeChild(last);
+                    last.childNodes.forEach((child) => {
+                        oneBefore.parentNode.appendChild(child);
+                    });
+                    stack.pop();
+                }
+                else {
+                    // Single error  <div> <h3> </div> handle: Just removes <h3>
+                    oneBefore.removeChild(last);
+                    last.childNodes.forEach((child) => {
+                        oneBefore.appendChild(child);
+                    });
+                }
+            }
+            else {
+                // If it's final element just skip.
+            }
+        }
+        response.childNodes.forEach((node) => {
+            if (node instanceof HTMLElement) {
+                node.parentNode = null;
+            }
+        });
+        return response;
+    }
+    else {
+        const response = new TextNode(data);
+        response.valid = valid;
+        return response;
+    }
 }
 exports.parse = parse;

dist/umd/index.d.ts

@@ -58,2 +58,5 @@ export declare enum NodeType {
 export declare class HTMLElement extends Node {
+    tagName: string;
+    private rawAttrs;
+    parentNode: Node;
     private _attrs;
@@ -63,7 +66,4 @@ private _rawAttrs;
     classNames: string[];
-    tagName: string;
-    rawAttrs: string;
     /**
      * Node Type declaration.
-     * @type {Number}
      */
@@ -73,10 +73,20 @@ nodeType: NodeType;
      * Creates an instance of HTMLElement.
-     * @param {string} name				tagName
-     * @param {KeyAttributes} keyAttrs	id and class attribute
-     * @param {string} [rawAttrs]	attributes in string
+     * @param keyAttrs	id and class attribute
+     * @param [rawAttrs]	attributes in string
      *
      * @memberof HTMLElement
      */
-    constructor(name: string, keyAttrs: KeyAttributes, rawAttrs?: string);
+    constructor(tagName: string, keyAttrs: KeyAttributes, rawAttrs?: string, parentNode?: Node);
     /**
+     * Remove Child element from childNodes array
+     * @param {HTMLElement} node     node to remove
+     */
+    removeChild(node: Node): void;
+    /**
+     * Exchanges given child with new child
+     * @param {HTMLElement} oldNode     node to exchange
+     * @param {HTMLElement} newNode     new node
+     */
+    exchangeChild(oldNode: Node, newNode: Node): void;
+    /**
      * Get escpaed (as-it) text value of current node and its children.
@@ -122,3 +132,3 @@ * @return {string} text content
      */
-    querySelectorAll(selector: string | Matcher): Node[];
+    querySelectorAll(selector: string | Matcher): HTMLElement[];
     /**
@@ -130,3 +140,3 @@ * Query CSS Selector to find matching node.
      */
-    querySelector(selector: string | Matcher): Node;
+    querySelector(selector: string | Matcher): HTMLElement;
     /**
@@ -137,3 +147,3 @@ * Append a child node to childNodes
      */
-    appendChild(node: Node): Node;
+    appendChild<T extends Node = Node>(node: T): T;
     /**
@@ -207,3 +217,8 @@ * Get first child node
 export declare function parse(data: string, options?: {
-    lowerCaseTagName: boolean;
-}): HTMLElement;
+    lowerCaseTagName?: boolean;
+    noFix?: boolean;
+}): (TextNode & {
+    valid: boolean;
+}) | (HTMLElement & {
+    valid: boolean;
+});

dist/umd/index.js

@@ -113,19 +113,21 @@ var __extends = (this && this.__extends) || (function () {
          * Creates an instance of HTMLElement.
-         * @param {string} name				tagName
-         * @param {KeyAttributes} keyAttrs	id and class attribute
-         * @param {string} [rawAttrs]	attributes in string
+         * @param keyAttrs	id and class attribute
+         * @param [rawAttrs]	attributes in string
          *
          * @memberof HTMLElement
          */
-        function HTMLElement(name, keyAttrs, rawAttrs) {
+        function HTMLElement(tagName, keyAttrs, rawAttrs, parentNode) {
+            if (rawAttrs === void 0) { rawAttrs = ''; }
+            if (parentNode === void 0) { parentNode = null; }
             var _this = _super.call(this) || this;
+            _this.tagName = tagName;
+            _this.rawAttrs = rawAttrs;
+            _this.parentNode = parentNode;
             _this.classNames = [];
             /**
              * Node Type declaration.
-             * @type {Number}
              */
             _this.nodeType = NodeType.ELEMENT_NODE;
-            _this.tagName = name;
             _this.rawAttrs = rawAttrs || '';
-            // this.parentNode = null;
+            _this.parentNode = parentNode || null;
             _this.childNodes = [];
@@ -140,2 +142,26 @@ if (keyAttrs.id) {
         }
+        /**
+         * Remove Child element from childNodes array
+         * @param {HTMLElement} node     node to remove
+         */
+        HTMLElement.prototype.removeChild = function (node) {
+            this.childNodes = this.childNodes.filter(function (child) {
+                return (child !== node);
+            });
+        };
+        /**
+         * Exchanges given child with new child
+         * @param {HTMLElement} oldNode     node to exchange
+         * @param {HTMLElement} newNode     new node
+         */
+        HTMLElement.prototype.exchangeChild = function (oldNode, newNode) {
+            var idx = -1;
+            for (var i = 0; i < this.childNodes.length; i++) {
+                if (this.childNodes[i] === oldNode) {
+                    idx = i;
+                    break;
+                }
+            }
+            this.childNodes[idx] = newNode;
+        };
         Object.defineProperty(HTMLElement.prototype, "rawText", {
@@ -267,19 +293,16 @@ /**
         HTMLElement.prototype.trimRight = function (pattern) {
-            function dfs(node) {
-                for (var i = 0; i < node.childNodes.length; i++) {
-                    var childNode = node.childNodes[i];
-                    if (childNode.nodeType === NodeType.ELEMENT_NODE) {
-                        dfs(childNode);
+            for (var i = 0; i < this.childNodes.length; i++) {
+                var childNode = this.childNodes[i];
+                if (childNode.nodeType === NodeType.ELEMENT_NODE) {
+                    childNode.trimRight(pattern);
+                }
+                else {
+                    var index = childNode.rawText.search(pattern);
+                    if (index > -1) {
+                        childNode.rawText = childNode.rawText.substr(0, index);
+                        // trim all following nodes.
+                        this.childNodes.length = i + 1;
                     }
-                    else {
-                        var index = childNode.rawText.search(pattern);
-                        if (index > -1) {
-                            childNode.rawText = childNode.rawText.substr(0, index);
-                            // trim all following nodes.
-                            node.childNodes.length = i + 1;
-                        }
-                    }
                 }
             }
-            dfs(this);
             return this;
@@ -445,2 +468,5 @@ };
             this.childNodes.push(node);
+            if (node instanceof HTMLElement) {
+                node.parentNode = this;
+            }
             return node;
@@ -588,3 +614,3 @@ };
             return false;
-            // for (var cls = classes, i = 0; i < cls.length; i++) {if (el.classNames.indexOf(cls[i]) === -1){ return false;}}
+            // for (let cls = classes, i = 0; i < cls.length; i++) {if (el.classNames.indexOf(cls[i]) === -1){ return false;}}
             // return true;
@@ -697,3 +723,3 @@ },
                         value = matcher_1[7] || matcher_1[8];
-                        source += "var attrs = el.attributes;for (var key in attrs){const val = attrs[key]; if (key == \"" + attr_key + "\" && val == \"" + value + "\"){return true;}} return false;"; //2
+                        source += "let attrs = el.attributes;for (let key in attrs){const val = attrs[key]; if (key == \"" + attr_key + "\" && val == \"" + value + "\"){return true;}} return false;"; //2
                         function_name += '2';
@@ -707,3 +733,3 @@ }
                 if (classes.length > 0) {
-                    source += 'for (var cls = ' + JSON.stringify(classes) + ', i = 0; i < cls.length; i++) if (el.classNames.indexOf(cls[i]) === -1) return false;'; //4
+                    source += 'for (let cls = ' + JSON.stringify(classes) + ', i = 0; i < cls.length; i++) if (el.classNames.indexOf(cls[i]) === -1) return false;'; //4
                     function_name += '4';
@@ -774,9 +800,11 @@ }
     var kSelfClosingElements = {
-        meta: true,
+        area: true,
+        base: true,
+        br: true,
+        col: true,
+        hr: true,
         img: true,
+        input: true,
         link: true,
-        input: true,
-        area: true,
-        br: true,
-        hr: true
+        meta: true
     };
@@ -786,4 +814,11 @@ var kElementsClosedByOpening = {
         p: { p: true, div: true },
+        b: { div: true },
         td: { td: true, th: true },
-        th: { td: true, th: true }
+        th: { td: true, th: true },
+        h1: { h1: true },
+        h2: { h2: true },
+        h3: { h3: true },
+        h4: { h4: true },
+        h5: { h5: true },
+        h6: { h6: true }
     };
@@ -836,5 +871,5 @@ var kElementsClosedByClosing = {
                 var attrs = {};
-                for (var attMatch; attMatch = kAttributePattern.exec(match[3]);)
+                for (var attMatch = void 0; attMatch = kAttributePattern.exec(match[3]);) {
                     attrs[attMatch[2]] = attMatch[4] || attMatch[5] || attMatch[6];
-                // console.log(attrs);
+                }
                 if (!match[4] && kElementsClosedByOpening[currentParent.tagName]) {
@@ -861,4 +896,5 @@ if (kElementsClosedByOpening[currentParent.tagName][match[2]]) {
                         }
-                        if (text.length > 0)
+                        if (text.length > 0) {
                             currentParent.appendChild(new TextNode(text));
+                        }
                     }
@@ -898,5 +934,48 @@ if (index == -1) {
         }
-        return root;
+        var valid = !!(stack.length === 1);
+        if (!options.noFix) {
+            var response = root;
+            response.valid = valid;
+            var _loop_1 = function () {
+                // Handle each error elements.
+                var last = stack.pop();
+                var oneBefore = arr_back(stack);
+                if (last.parentNode && last.parentNode.parentNode) {
+                    if (last.parentNode === oneBefore && last.tagName === oneBefore.tagName) {
+                        // Pair error case <h3> <h3> handle : Fixes to <h3> </h3>
+                        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);
+                        });
+                    }
+                }
+                else {
+                    // If it's final element just skip.
+                }
+            };
+            while (stack.length > 1) {
+                _loop_1();
+            }
+            response.childNodes.forEach(function (node) {
+                if (node instanceof HTMLElement) {
+                    node.parentNode = null;
+                }
+            });
+            return response;
+        }
+        else {
+            var response = new TextNode(data);
+            response.valid = valid;
+            return response;
+        }
     }
     exports.parse = parse;
 });

package.json

 {
 	"name": "node-html-parser",
-	"version": "1.1.11",
+	"version": "1.1.12",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
@@ -9,5 +9,7 @@ "main": "dist/index.js",
 		"test": "mocha",
-		"build": "gulp",
-		"pretest": "gulp",
-		"prepare": "gulp"
+		"clean": "del-cli ./dist/",
+		"ts:cjs": "tsc -m commonjs",
+		"ts:umd": "tsc -t es5 -m umd --outDir ./dist/umd/",
+		"build": "npm run clean && npm run ts:cjs && npm run ts:umd",
+		"dev": "tsc -w"
 	},
@@ -30,6 +32,3 @@ "keywords": [
 		"blanket": "latest",
-		"del": "latest",
-		"gulp": "latest",
-		"gulp-sequence": "latest",
-		"gulp-shell": "latest",
+		"del-cli": "latest",
 		"mocha": "latest",
@@ -36,0 +35,0 @@ "should": "latest",

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

123891

increased by62.71%

Dev dependency count

10

decreased by-23.08%

Number of package files

9

increased by28.57%

Lines of code

2308

increased by9.23%

No dependency changes