Huge News!Announcing our $40M Series B led by Abstract Ventures.Learn More →
Socket
Sign inDemoInstall
Socket
Sign inDemoInstall

node-html-parser

Package Overview
Dependencies
Maintainers
1
Versions
120
Alerts
File Explorer

Advanced tools

License
Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

node-html-parser - npm Package Compare versions

Comparing version 1.2.20 to 1.2.21

.eslintrc.json
.yarnrc

18

dist/index.js
"use strict";
var __importDefault = (this && this.__importDefault) || function (mod) {
return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
exports.NodeType = exports.TextNode = exports.Node = exports.default = exports.parse = exports.HTMLElement = exports.CommentNode = void 0;
var comment_1 = require("./nodes/comment");
exports.CommentNode = comment_1.default;
Object.defineProperty(exports, "CommentNode", { enumerable: true, get: function () { return __importDefault(comment_1).default; } });
var html_1 = require("./nodes/html");
exports.HTMLElement = html_1.default;
exports.parse = html_1.parse;
exports.default = html_1.parse;
Object.defineProperty(exports, "HTMLElement", { enumerable: true, get: function () { return __importDefault(html_1).default; } });
Object.defineProperty(exports, "parse", { enumerable: true, get: function () { return html_1.parse; } });
Object.defineProperty(exports, "default", { enumerable: true, get: function () { return html_1.parse; } });
var node_1 = require("./nodes/node");
exports.Node = node_1.default;
Object.defineProperty(exports, "Node", { enumerable: true, get: function () { return __importDefault(node_1).default; } });
var text_1 = require("./nodes/text");
exports.TextNode = text_1.default;
Object.defineProperty(exports, "TextNode", { enumerable: true, get: function () { return __importDefault(text_1).default; } });
var type_1 = require("./nodes/type");
exports.NodeType = type_1.default;
Object.defineProperty(exports, "NodeType", { enumerable: true, get: function () { return __importDefault(type_1).default; } });

298

dist/main.js

@@ -5,3 +5,3 @@ var __extends = (this && this.__extends) || (function () {

({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
function (d, b) { for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]; };
function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };
return extendStatics(d, b);

@@ -18,2 +18,9 @@ };

};
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;
};
define("back", ["require", "exports"], function (require, exports) {

@@ -48,2 +55,9 @@ "use strict";

}
Object.defineProperty(Node.prototype, "innerText", {
get: function () {
return this.rawText;
},
enumerable: false,
configurable: true
});
return Node;

@@ -60,4 +74,5 @@ }());

__extends(CommentNode, _super);
function CommentNode(value) {
function CommentNode(rawText) {
var _this = _super.call(this) || this;
_this.rawText = rawText;
/**

@@ -68,3 +83,2 @@ * Node Type declaration.

_this.nodeType = type_1.default.COMMENT_NODE;
_this.rawText = value;
return _this;

@@ -80,3 +94,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -102,4 +116,5 @@ });

__extends(TextNode, _super);
function TextNode(value) {
function TextNode(rawText) {
var _this = _super.call(this) || this;
_this.rawText = rawText;
/**

@@ -110,3 +125,2 @@ * Node Type declaration.

_this.nodeType = type_2.default.TEXT_NODE;
_this.rawText = value;
return _this;

@@ -122,3 +136,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -134,3 +148,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -153,2 +167,11 @@ });

var pMatchFunctionCache = {};
function compare_tagname(tag1, tag2) {
if (!tag1) {
return !tag2;
}
if (!tag2) {
return !tag1;
}
return tag1.toLowerCase() === tag2.toLowerCase();
}
/**

@@ -253,3 +276,3 @@ * Function cache

classes = classes || [];
if (el.tagName !== tagName) {
if (!compare_tagname(el.tagName, tagName)) {
return false;

@@ -267,3 +290,3 @@ }

tagName = tagName || '';
return el.tagName === tagName;
return compare_tagname(el.tagName, tagName);
},

@@ -273,5 +296,6 @@ f3: function (el, tagName) {

tagName = tagName || '';
if (el.tagName !== tagName) {
return false;
}
// if (el.tagName !== tagName) {
// return false;
// }
return compare_tagname(el.tagName, tagName);
}

@@ -293,6 +317,6 @@ };

this.nextMatch = 0;
functionCache.f5 = functionCache.f5;
this.matchers = selector.split(' ').map(function (matcher) {
if (pMatchFunctionCache[matcher])
if (pMatchFunctionCache[matcher]) {
return pMatchFunctionCache[matcher];
}
var parts = matcher.split('.');

@@ -317,2 +341,3 @@ var tagName = parts[0];

if (method !== '=' && method !== '!=') {
// eslint-disable-next-line no-template-curly-in-string
throw new Error('Selector not supported, Expect [key${op}value].op must be =,!=');

@@ -327,3 +352,3 @@ }

}
else if (reg = /^\[(.*?)\]/.exec(tagName)) {
else if ((reg = /^\[(.*?)\]/.exec(tagName))) {
attr_key = reg[1];

@@ -352,3 +377,3 @@ function_name += '5';

// source = source || '';
return pMatchFunctionCache[matcher] = obj;
return (pMatchFunctionCache[matcher] = obj);
});

@@ -383,3 +408,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -407,2 +432,3 @@ });

Object.defineProperty(exports, "__esModule", { value: true });
exports.parse = void 0;
node_3 = __importDefault(node_3);

@@ -415,12 +441,12 @@ type_3 = __importDefault(type_3);

var kBlockElements = {
div: true,
p: true,
DIV: true,
P: true,
// ul: true,
// ol: true,
li: true,
LI: true,
// table: true,
// tr: true,
td: true,
section: true,
br: true
TD: true,
SECTION: true,
BR: true
};

@@ -449,3 +475,2 @@ /**

var _this = _super.call(this) || this;
_this.tagName = tagName;
_this.rawAttrs = rawAttrs;

@@ -458,2 +483,3 @@ _this.parentNode = parentNode;

_this.nodeType = type_3.default.ELEMENT_NODE;
_this._tag_name = tagName;
_this.rawAttrs = rawAttrs || '';

@@ -506,2 +532,9 @@ _this.parentNode = parentNode || null;

};
Object.defineProperty(HTMLElement.prototype, "tagName", {
get: function () {
return this._tag_name ? this._tag_name.toUpperCase() : this._tag_name;
},
enumerable: false,
configurable: true
});
Object.defineProperty(HTMLElement.prototype, "rawText", {

@@ -514,6 +547,6 @@ /**

return this.childNodes.reduce(function (pre, cur) {
return pre += cur.rawText;
return (pre += cur.rawText);
}, '');
},
enumerable: true,
enumerable: false,
configurable: true

@@ -529,3 +562,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -564,3 +597,3 @@ });

if (currentBlock.prependWhitespace) {
text = ' ' + text;
text = " " + text;
currentBlock.prependWhitespace = false;

@@ -573,4 +606,3 @@ }

dfs(this);
return blocks
.map(function (block) {
return blocks.map(function (block) {
// Normalize each line's whitespace

@@ -581,20 +613,16 @@ return block.join('').trim().replace(/\s{2,}/g, ' ');

},
enumerable: true,
enumerable: false,
configurable: true
});
HTMLElement.prototype.toString = function () {
var tag = this.tagName;
var tag = this._tag_name;
if (tag) {
var is_void = /^(area|base|br|col|embed|hr|img|input|link|meta|param|source|track|wbr)$/i.test(tag);
var attrs = this.rawAttrs ? ' ' + this.rawAttrs : '';
var attrs = this.rawAttrs ? " " + this.rawAttrs : '';
if (is_void) {
return "<" + tag + attrs + ">";
}
else {
return "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
}
return "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
}
else {
return this.innerHTML;
}
return this.innerHTML;
};

@@ -607,3 +635,3 @@ Object.defineProperty(HTMLElement.prototype, "innerHTML", {

},
enumerable: true,
enumerable: false,
configurable: true

@@ -626,3 +654,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -664,5 +692,5 @@ });

function dfs(node) {
var idStr = node.id ? ('#' + node.id) : '';
var classStr = node.classNames.length ? ('.' + node.classNames.join('.')) : '';
write(node.tagName + idStr + classStr);
var idStr = node.id ? ("#" + node.id) : '';
var classStr = node.classNames.length ? ("." + node.classNames.join('.')) : '';
write(node._tag_name + idStr + classStr);
indention++;

@@ -674,4 +702,5 @@ node.childNodes.forEach(function (childNode) {

else if (childNode.nodeType === type_3.default.TEXT_NODE) {
if (!childNode.isWhitespace)
if (!childNode.isWhitespace) {
write('#text');
}
}

@@ -684,3 +713,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -812,4 +841,5 @@ });

else {
if (state[2])
if (state[2]) {
matcher.rewind();
}
stack.pop();

@@ -842,3 +872,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -854,3 +884,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -875,3 +905,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -885,9 +915,10 @@ });

get: function () {
if (this._rawAttrs)
if (this._rawAttrs) {
return this._rawAttrs;
}
var attrs = {};
if (this.rawAttrs) {
var re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|(\S+)))?/ig;
var re = /\b([a-z][a-z0-9-]*)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|(\S+)))?/ig;
var match = void 0;
while (match = re.exec(this.rawAttrs)) {
while ((match = re.exec(this.rawAttrs))) {
attrs[match[1]] = match[2] || match[3] || match[4] || null;

@@ -899,3 +930,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -916,5 +947,3 @@ });

}
else {
return name + '=' + val;
}
return name + "=" + val;
}).join(' ');

@@ -952,5 +981,3 @@ };

}
else {
return name + '=' + val;
}
return name + "=" + val;
}).join(' ');

@@ -977,9 +1004,7 @@ };

}
else {
return name + '=' + JSON.stringify(String(val));
}
return name + "=" + JSON.stringify(String(val));
}).join(' ');
};
HTMLElement.prototype.insertAdjacentHTML = function (where, html) {
var _a, _b;
var _a, _b, _c;
var _this = this;

@@ -991,8 +1016,14 @@ if (arguments.length < 2) {

if (where === 'afterend') {
var idx = this.parentNode.childNodes.findIndex(function (child) {
return child === _this;
});
(_a = this.parentNode.childNodes).splice.apply(_a, __spreadArrays([idx + 1, 0], p.childNodes));
p.childNodes.forEach(function (n) {
_this.parentNode.appendChild(n);
if (n instanceof HTMLElement) {
n.parentNode = _this.parentNode;
}
});
}
else if (where === 'afterbegin') {
(_a = this.childNodes).unshift.apply(_a, p.childNodes);
(_b = this.childNodes).unshift.apply(_b, p.childNodes);
}

@@ -1005,3 +1036,11 @@ else if (where === 'beforeend') {

else if (where === 'beforebegin') {
(_b = this.parentNode.childNodes).unshift.apply(_b, p.childNodes);
var idx = this.parentNode.childNodes.findIndex(function (child) {
return child === _this;
});
(_c = this.parentNode.childNodes).splice.apply(_c, __spreadArrays([idx, 0], p.childNodes));
p.childNodes.forEach(function (n) {
if (n instanceof HTMLElement) {
n.parentNode = _this.parentNode;
}
});
}

@@ -1011,5 +1050,5 @@ else {

}
if (!where || html === undefined || html === null) {
return;
}
// if (!where || html === undefined || html === null) {
// return;
// }
};

@@ -1024,39 +1063,71 @@ return HTMLElement;

area: true,
AREA: true,
base: true,
BASE: true,
br: true,
BR: true,
col: true,
COL: true,
hr: true,
HR: true,
img: true,
IMG: true,
input: true,
INPUT: true,
link: true,
LINK: true,
meta: true,
source: true
META: true,
source: true,
SOURCE: true
};
var kElementsClosedByOpening = {
li: { li: true },
p: { p: true, div: true },
b: { div: true },
td: { 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 }
li: { li: true, LI: true },
LI: { li: true, LI: true },
p: { p: true, div: true, P: true, DIV: true },
P: { p: true, div: true, P: true, DIV: true },
b: { div: true, DIV: true },
B: { div: true, DIV: true },
td: { td: true, th: true, TD: true, TH: true },
TD: { td: true, th: true, TD: true, TH: true },
th: { td: true, th: true, TD: true, TH: true },
TH: { td: true, th: true, TD: true, TH: true },
h1: { h1: true, H1: true },
H1: { h1: true, H1: true },
h2: { h2: true, H2: true },
H2: { h2: true, H2: true },
h3: { h3: true, H3: true },
H3: { h3: true, H3: true },
h4: { h4: true, H4: true },
H4: { h4: true, H4: true },
h5: { h5: true, H5: true },
H5: { h5: true, H5: true },
h6: { h6: true, H6: true },
H6: { h6: true, H6: true }
};
var kElementsClosedByClosing = {
li: { ul: true, ol: true },
a: { div: true },
b: { div: true },
i: { div: true },
p: { div: true },
td: { tr: true, table: true },
th: { tr: true, table: true }
li: { ul: true, ol: true, UL: true, OL: true },
LI: { ul: true, ol: true, UL: true, OL: true },
a: { div: true, DIV: true },
A: { div: true, DIV: true },
b: { div: true, DIV: true },
B: { div: true, DIV: true },
i: { div: true, DIV: true },
I: { div: true, DIV: true },
p: { div: true, DIV: true },
P: { div: true, DIV: true },
td: { tr: true, table: true, TR: true, TABLE: true },
TD: { tr: true, table: true, TR: true, TABLE: true },
th: { tr: true, table: true, TR: true, TABLE: true },
TH: { tr: true, table: true, TR: true, TABLE: true }
};
var kBlockTextElements = {
script: true,
SCRIPT: true,
noscript: true,
NOSCRIPT: true,
style: true,
pre: true
STYLE: true,
pre: true,
PRE: true
};

@@ -1100,3 +1171,3 @@ var frameflag = 'documentfragmentcontainer';

var attrs = {};
for (var attMatch = void 0; attMatch = kAttributePattern.exec(match[3]);) {
for (var attMatch = void 0; (attMatch = kAttributePattern.exec(match[3]));) {
attrs[attMatch[2]] = attMatch[4] || attMatch[5] || attMatch[6];

@@ -1117,3 +1188,3 @@ }

// a little test to find next </script> or </style> ...
var closeMarkup_1 = '</' + match[2] + '>';
var closeMarkup_1 = "</" + match[2] + ">";
var index = (function () {

@@ -1123,5 +1194,3 @@ if (options.lowerCaseTagName) {

}
else {
return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
}
return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
})();

@@ -1153,3 +1222,3 @@ if (options[match[2]]) {

while (true) {
if (currentParent.tagName === match[2]) {
if (currentParent.tagName === match[2].toUpperCase()) {
stack.pop();

@@ -1175,9 +1244,9 @@ currentParent = back_1.default(stack);

};
while (match = kMarkupPattern.exec(data)) {
while ((match = kMarkupPattern.exec(data))) {
_loop_1();
}
var valid = !!(stack.length === 1);
var valid = Boolean(stack.length === 1);
if (!options.noFix) {
var response = root;
response.valid = valid;
var response_1 = root;
response_1.valid = valid;
var _loop_2 = function () {

@@ -1211,3 +1280,3 @@ // Handle each error elements.

}
response.childNodes.forEach(function (node) {
response_1.childNodes.forEach(function (node) {
if (node instanceof HTMLElement) {

@@ -1217,9 +1286,7 @@ node.parentNode = null;

});
return response;
return response_1;
}
else {
var response = new text_1.default(data);
response.valid = valid;
return response;
}
var response = new text_1.default(data);
response.valid = valid;
return response;
}

@@ -1231,9 +1298,10 @@ exports.parse = parse;

Object.defineProperty(exports, "__esModule", { value: true });
exports.CommentNode = comment_2.default;
exports.HTMLElement = html_1.default;
exports.parse = html_1.parse;
exports.default = html_1.parse;
exports.Node = node_4.default;
exports.TextNode = text_2.default;
exports.NodeType = type_4.default;
exports.NodeType = exports.TextNode = exports.Node = exports.default = exports.parse = exports.HTMLElement = exports.CommentNode = void 0;
Object.defineProperty(exports, "CommentNode", { enumerable: true, get: function () { return __importDefault(comment_2).default; } });
Object.defineProperty(exports, "HTMLElement", { enumerable: true, get: function () { return __importDefault(html_1).default; } });
Object.defineProperty(exports, "parse", { enumerable: true, get: function () { return html_1.parse; } });
Object.defineProperty(exports, "default", { enumerable: true, get: function () { return html_1.parse; } });
Object.defineProperty(exports, "Node", { enumerable: true, get: function () { return __importDefault(node_4).default; } });
Object.defineProperty(exports, "TextNode", { enumerable: true, get: function () { return __importDefault(text_2).default; } });
Object.defineProperty(exports, "NodeType", { enumerable: true, get: function () { return __importDefault(type_4).default; } });
});

31

dist/matcher.js

@@ -8,2 +8,11 @@ "use strict";

var pMatchFunctionCache = {};
function compare_tagname(tag1, tag2) {
if (!tag1) {
return !tag2;
}
if (!tag2) {
return !tag1;
}
return tag1.toLowerCase() === tag2.toLowerCase();
}
/**

@@ -108,3 +117,3 @@ * Function cache

classes = classes || [];
if (el.tagName !== tagName) {
if (!compare_tagname(el.tagName, tagName)) {
return false;

@@ -122,3 +131,3 @@ }

tagName = tagName || '';
return el.tagName === tagName;
return compare_tagname(el.tagName, tagName);
},

@@ -128,5 +137,6 @@ f3: function (el, tagName) {

tagName = tagName || '';
if (el.tagName !== tagName) {
return false;
}
// if (el.tagName !== tagName) {
// return false;
// }
return compare_tagname(el.tagName, tagName);
}

@@ -148,6 +158,6 @@ };

this.nextMatch = 0;
functionCache.f5 = functionCache.f5;
this.matchers = selector.split(' ').map(function (matcher) {
if (pMatchFunctionCache[matcher])
if (pMatchFunctionCache[matcher]) {
return pMatchFunctionCache[matcher];
}
var parts = matcher.split('.');

@@ -172,2 +182,3 @@ var tagName = parts[0];

if (method !== '=' && method !== '!=') {
// eslint-disable-next-line no-template-curly-in-string
throw new Error('Selector not supported, Expect [key${op}value].op must be =,!=');

@@ -182,3 +193,3 @@ }

}
else if (reg = /^\[(.*?)\]/.exec(tagName)) {
else if ((reg = /^\[(.*?)\]/.exec(tagName))) {
attr_key = reg[1];

@@ -207,3 +218,3 @@ function_name += '5';

// source = source || '';
return pMatchFunctionCache[matcher] = obj;
return (pMatchFunctionCache[matcher] = obj);
});

@@ -238,3 +249,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -241,0 +252,0 @@ });

3

dist/nodes/comment.d.ts
import Node from './node';
import NodeType from './type';
export default class CommentNode extends Node {
constructor(value: string);
rawText: string;
constructor(rawText: string);
/**

@@ -6,0 +7,0 @@ * Node Type declaration.

8

dist/nodes/comment.js

@@ -6,3 +6,3 @@ "use strict";

({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
function (d, b) { for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]; };
function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };
return extendStatics(d, b);

@@ -24,4 +24,5 @@ };

__extends(CommentNode, _super);
function CommentNode(value) {
function CommentNode(rawText) {
var _this = _super.call(this) || this;
_this.rawText = rawText;
/**

@@ -32,3 +33,2 @@ * Node Type declaration.

_this.nodeType = type_1.default.COMMENT_NODE;
_this.rawText = value;
return _this;

@@ -44,3 +44,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -47,0 +47,0 @@ });

3

dist/nodes/html.d.ts

@@ -26,3 +26,2 @@ import Node from './node';

export default class HTMLElement extends Node {
tagName: string;
private rawAttrs;

@@ -32,2 +31,3 @@ parentNode: Node;

private _rawAttrs;
private _tag_name;
id: string;

@@ -58,2 +58,3 @@ classNames: string[];

exchangeChild(oldNode: Node, newNode: Node): void;
get tagName(): string;
/**

@@ -60,0 +61,0 @@ * Get escpaed (as-it) text value of current node and its children.

231

dist/nodes/html.js

@@ -6,3 +6,3 @@ "use strict";

({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
function (d, b) { for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]; };
function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };
return extendStatics(d, b);

@@ -16,2 +16,9 @@ };

})();
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 __importDefault = (this && this.__importDefault) || function (mod) {

@@ -21,2 +28,3 @@ return (mod && mod.__esModule) ? mod : { "default": mod };

Object.defineProperty(exports, "__esModule", { value: true });
exports.parse = void 0;
var he_1 = require("he");

@@ -30,12 +38,12 @@ var node_1 = __importDefault(require("./node"));

var kBlockElements = {
div: true,
p: true,
DIV: true,
P: true,
// ul: true,
// ol: true,
li: true,
LI: true,
// table: true,
// tr: true,
td: true,
section: true,
br: true
TD: true,
SECTION: true,
BR: true
};

@@ -64,3 +72,2 @@ /**

var _this = _super.call(this) || this;
_this.tagName = tagName;
_this.rawAttrs = rawAttrs;

@@ -73,2 +80,3 @@ _this.parentNode = parentNode;

_this.nodeType = type_1.default.ELEMENT_NODE;
_this._tag_name = tagName;
_this.rawAttrs = rawAttrs || '';

@@ -121,2 +129,9 @@ _this.parentNode = parentNode || null;

};
Object.defineProperty(HTMLElement.prototype, "tagName", {
get: function () {
return this._tag_name ? this._tag_name.toUpperCase() : this._tag_name;
},
enumerable: false,
configurable: true
});
Object.defineProperty(HTMLElement.prototype, "rawText", {

@@ -129,6 +144,6 @@ /**

return this.childNodes.reduce(function (pre, cur) {
return pre += cur.rawText;
return (pre += cur.rawText);
}, '');
},
enumerable: true,
enumerable: false,
configurable: true

@@ -144,3 +159,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -179,3 +194,3 @@ });

if (currentBlock.prependWhitespace) {
text = ' ' + text;
text = " " + text;
currentBlock.prependWhitespace = false;

@@ -188,4 +203,3 @@ }

dfs(this);
return blocks
.map(function (block) {
return blocks.map(function (block) {
// Normalize each line's whitespace

@@ -196,20 +210,16 @@ return block.join('').trim().replace(/\s{2,}/g, ' ');

},
enumerable: true,
enumerable: false,
configurable: true
});
HTMLElement.prototype.toString = function () {
var tag = this.tagName;
var tag = this._tag_name;
if (tag) {
var is_void = /^(area|base|br|col|embed|hr|img|input|link|meta|param|source|track|wbr)$/i.test(tag);
var attrs = this.rawAttrs ? ' ' + this.rawAttrs : '';
var attrs = this.rawAttrs ? " " + this.rawAttrs : '';
if (is_void) {
return "<" + tag + attrs + ">";
}
else {
return "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
}
return "<" + tag + attrs + ">" + this.innerHTML + "</" + tag + ">";
}
else {
return this.innerHTML;
}
return this.innerHTML;
};

@@ -222,3 +232,3 @@ Object.defineProperty(HTMLElement.prototype, "innerHTML", {

},
enumerable: true,
enumerable: false,
configurable: true

@@ -241,3 +251,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -279,5 +289,5 @@ });

function dfs(node) {
var idStr = node.id ? ('#' + node.id) : '';
var classStr = node.classNames.length ? ('.' + node.classNames.join('.')) : '';
write(node.tagName + idStr + classStr);
var idStr = node.id ? ("#" + node.id) : '';
var classStr = node.classNames.length ? ("." + node.classNames.join('.')) : '';
write(node._tag_name + idStr + classStr);
indention++;

@@ -289,4 +299,5 @@ node.childNodes.forEach(function (childNode) {

else if (childNode.nodeType === type_1.default.TEXT_NODE) {
if (!childNode.isWhitespace)
if (!childNode.isWhitespace) {
write('#text');
}
}

@@ -299,3 +310,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -427,4 +438,5 @@ });

else {
if (state[2])
if (state[2]) {
matcher.rewind();
}
stack.pop();

@@ -457,3 +469,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -469,3 +481,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -490,3 +502,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -500,9 +512,10 @@ });

get: function () {
if (this._rawAttrs)
if (this._rawAttrs) {
return this._rawAttrs;
}
var attrs = {};
if (this.rawAttrs) {
var re = /\b([a-z][a-z0-9\-]*)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|(\S+)))?/ig;
var re = /\b([a-z][a-z0-9-]*)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|(\S+)))?/ig;
var match = void 0;
while (match = re.exec(this.rawAttrs)) {
while ((match = re.exec(this.rawAttrs))) {
attrs[match[1]] = match[2] || match[3] || match[4] || null;

@@ -514,3 +527,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -531,5 +544,3 @@ });

}
else {
return name + '=' + val;
}
return name + "=" + val;
}).join(' ');

@@ -567,5 +578,3 @@ };

}
else {
return name + '=' + val;
}
return name + "=" + val;
}).join(' ');

@@ -592,9 +601,7 @@ };

}
else {
return name + '=' + JSON.stringify(String(val));
}
return name + "=" + JSON.stringify(String(val));
}).join(' ');
};
HTMLElement.prototype.insertAdjacentHTML = function (where, html) {
var _a, _b;
var _a, _b, _c;
var _this = this;

@@ -606,8 +613,14 @@ if (arguments.length < 2) {

if (where === 'afterend') {
var idx = this.parentNode.childNodes.findIndex(function (child) {
return child === _this;
});
(_a = this.parentNode.childNodes).splice.apply(_a, __spreadArrays([idx + 1, 0], p.childNodes));
p.childNodes.forEach(function (n) {
_this.parentNode.appendChild(n);
if (n instanceof HTMLElement) {
n.parentNode = _this.parentNode;
}
});
}
else if (where === 'afterbegin') {
(_a = this.childNodes).unshift.apply(_a, p.childNodes);
(_b = this.childNodes).unshift.apply(_b, p.childNodes);
}

@@ -620,3 +633,11 @@ else if (where === 'beforeend') {

else if (where === 'beforebegin') {
(_b = this.parentNode.childNodes).unshift.apply(_b, p.childNodes);
var idx = this.parentNode.childNodes.findIndex(function (child) {
return child === _this;
});
(_c = this.parentNode.childNodes).splice.apply(_c, __spreadArrays([idx, 0], p.childNodes));
p.childNodes.forEach(function (n) {
if (n instanceof HTMLElement) {
n.parentNode = _this.parentNode;
}
});
}

@@ -626,5 +647,5 @@ else {

}
if (!where || html === undefined || html === null) {
return;
}
// if (!where || html === undefined || html === null) {
// return;
// }
};

@@ -639,39 +660,71 @@ return HTMLElement;

area: true,
AREA: true,
base: true,
BASE: true,
br: true,
BR: true,
col: true,
COL: true,
hr: true,
HR: true,
img: true,
IMG: true,
input: true,
INPUT: true,
link: true,
LINK: true,
meta: true,
source: true
META: true,
source: true,
SOURCE: true
};
var kElementsClosedByOpening = {
li: { li: true },
p: { p: true, div: true },
b: { div: true },
td: { 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 }
li: { li: true, LI: true },
LI: { li: true, LI: true },
p: { p: true, div: true, P: true, DIV: true },
P: { p: true, div: true, P: true, DIV: true },
b: { div: true, DIV: true },
B: { div: true, DIV: true },
td: { td: true, th: true, TD: true, TH: true },
TD: { td: true, th: true, TD: true, TH: true },
th: { td: true, th: true, TD: true, TH: true },
TH: { td: true, th: true, TD: true, TH: true },
h1: { h1: true, H1: true },
H1: { h1: true, H1: true },
h2: { h2: true, H2: true },
H2: { h2: true, H2: true },
h3: { h3: true, H3: true },
H3: { h3: true, H3: true },
h4: { h4: true, H4: true },
H4: { h4: true, H4: true },
h5: { h5: true, H5: true },
H5: { h5: true, H5: true },
h6: { h6: true, H6: true },
H6: { h6: true, H6: true }
};
var kElementsClosedByClosing = {
li: { ul: true, ol: true },
a: { div: true },
b: { div: true },
i: { div: true },
p: { div: true },
td: { tr: true, table: true },
th: { tr: true, table: true }
li: { ul: true, ol: true, UL: true, OL: true },
LI: { ul: true, ol: true, UL: true, OL: true },
a: { div: true, DIV: true },
A: { div: true, DIV: true },
b: { div: true, DIV: true },
B: { div: true, DIV: true },
i: { div: true, DIV: true },
I: { div: true, DIV: true },
p: { div: true, DIV: true },
P: { div: true, DIV: true },
td: { tr: true, table: true, TR: true, TABLE: true },
TD: { tr: true, table: true, TR: true, TABLE: true },
th: { tr: true, table: true, TR: true, TABLE: true },
TH: { tr: true, table: true, TR: true, TABLE: true }
};
var kBlockTextElements = {
script: true,
SCRIPT: true,
noscript: true,
NOSCRIPT: true,
style: true,
pre: true
STYLE: true,
pre: true,
PRE: true
};

@@ -715,3 +768,3 @@ var frameflag = 'documentfragmentcontainer';

var attrs = {};
for (var attMatch = void 0; attMatch = kAttributePattern.exec(match[3]);) {
for (var attMatch = void 0; (attMatch = kAttributePattern.exec(match[3]));) {
attrs[attMatch[2]] = attMatch[4] || attMatch[5] || attMatch[6];

@@ -732,3 +785,3 @@ }

// a little test to find next </script> or </style> ...
var closeMarkup_1 = '</' + match[2] + '>';
var closeMarkup_1 = "</" + match[2] + ">";
var index = (function () {

@@ -738,5 +791,3 @@ if (options.lowerCaseTagName) {

}
else {
return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
}
return data.indexOf(closeMarkup_1, kMarkupPattern.lastIndex);
})();

@@ -768,3 +819,3 @@ if (options[match[2]]) {

while (true) {
if (currentParent.tagName === match[2]) {
if (currentParent.tagName === match[2].toUpperCase()) {
stack.pop();

@@ -790,9 +841,9 @@ currentParent = back_1.default(stack);

};
while (match = kMarkupPattern.exec(data)) {
while ((match = kMarkupPattern.exec(data))) {
_loop_1();
}
var valid = !!(stack.length === 1);
var valid = Boolean(stack.length === 1);
if (!options.noFix) {
var response = root;
response.valid = valid;
var response_1 = root;
response_1.valid = valid;
var _loop_2 = function () {

@@ -826,3 +877,3 @@ // Handle each error elements.

}
response.childNodes.forEach(function (node) {
response_1.childNodes.forEach(function (node) {
if (node instanceof HTMLElement) {

@@ -832,10 +883,8 @@ node.parentNode = null;

});
return response;
return response_1;
}
else {
var response = new text_1.default(data);
response.valid = valid;
return response;
}
var response = new text_1.default(data);
response.valid = valid;
return response;
}
exports.parse = parse;

7

dist/nodes/node.d.ts

@@ -6,7 +6,8 @@ import NodeType from './type';

export default abstract class Node {
nodeType: NodeType;
abstract nodeType: NodeType;
childNodes: Node[];
text: string;
rawText: string;
abstract text: string;
abstract rawText: string;
abstract toString(): string;
get innerText(): string;
}

7

dist/nodes/node.js

@@ -10,4 +10,11 @@ "use strict";

}
Object.defineProperty(Node.prototype, "innerText", {
get: function () {
return this.rawText;
},
enumerable: false,
configurable: true
});
return Node;
}());
exports.default = Node;

3

dist/nodes/text.d.ts

@@ -8,3 +8,4 @@ import NodeType from './type';

export default class TextNode extends Node {
constructor(value: string);
rawText: string;
constructor(rawText: string);
/**

@@ -11,0 +12,0 @@ * Node Type declaration.

10

dist/nodes/text.js

@@ -6,3 +6,3 @@ "use strict";

({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
function (d, b) { for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]; };
function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };
return extendStatics(d, b);

@@ -28,4 +28,5 @@ };

__extends(TextNode, _super);
function TextNode(value) {
function TextNode(rawText) {
var _this = _super.call(this) || this;
_this.rawText = rawText;
/**

@@ -36,3 +37,2 @@ * Node Type declaration.

_this.nodeType = type_1.default.TEXT_NODE;
_this.rawText = value;
return _this;

@@ -48,3 +48,3 @@ }

},
enumerable: true,
enumerable: false,
configurable: true

@@ -60,3 +60,3 @@ });

},
enumerable: true,
enumerable: false,
configurable: true

@@ -63,0 +63,0 @@ });

10

package.json
{
"name": "node-html-parser",
"version": "1.2.20",
"version": "1.2.21",
"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",

@@ -26,3 +26,3 @@ "main": "dist/index.js",

"dependencies": {
"he": "1.1.1"
"he": "1.2.0"
},

@@ -39,2 +39,4 @@ "devDependencies": {

"eslint": "latest",
"eslint-config-prettier": "latest",
"eslint-plugin-import": "latest",
"high5": "^1.0.0",

@@ -44,9 +46,7 @@ "htmlparser": "^1.7.7",

"htmlparser2": "^4.1.0",
"libxmljs": "^0.19.7",
"mocha": "latest",
"parse5": "^6.0.0",
"parse5": "^6.0.1",
"should": "latest",
"spec": "latest",
"travis-cov": "latest",
"tslint": "^5.0.0",
"typescript": "next"

@@ -53,0 +53,0 @@ },

74

README.md

@@ -66,3 +66,3 @@ # 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://travis-ci.org/taoqf/node-html-parser.svg?branch=master)](https://travis-ci.org/taoqf/node-html-parser)

## API
## HTMLElement Methods

@@ -86,16 +86,2 @@ ### parse(data[, options])

### HTMLElement#text
Get unescaped text value of current node and its children. Like `innerText`.
(slow for the first time)
### HTMLElement#rawText
Get escpaed (as-it) text value of current node and its children. May have
`&amp;` in it. (fast)
### HTMLElement#structuredText
Get structured Text
### HTMLElement#trimRight()

@@ -105,6 +91,2 @@

### HTMLElement#structure
Get DOM structure
### HTMLElement#removeWhitespace()

@@ -134,10 +116,2 @@

### HTMLElement#firstChild
Get first child node
### HTMLElement#lastChild
Get last child node
### HTMLElement#setAttribute(key: string, value: string)

@@ -153,4 +127,12 @@

Get `key` attrubte.
Get `key` attribute.
### HTMLElement#exchangeChild(oldNode: Node, newNode: Node)
Exchanges given child with new child.
### HTMLElement#removeChild(node: Node)
Remove child node.
### HTMLElement#toString()

@@ -160,2 +142,35 @@

### HTMLElement#set_content(content: string | Node | Node[])
Set content. **Notice**: Do not set content of the **root** node.
## HTMLElement Properties
### HTMLElement#text
Get unescaped text value of current node and its children. Like `innerText`.
(slow for the first time)
### HTMLElement#rawText
Get escpaed (as-it) text value of current node and its children. May have
`&amp;` in it. (fast)
### HTMLElement#structuredText
Get structured Text
### HTMLElement#structure
Get DOM structure
### HTMLElement#firstChild
Get first child node
### HTMLElement#lastChild
Get last child node
### HTMLElement#innerHTML

@@ -169,4 +184,1 @@

### HTMLElement#set_content(content: string | Node | Node[])
Set content. **Notice**: Do not set content of the **root** node.
.eslintrc.js
yarn-error.log
SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

About

Packages

npm

Go

Maven

PyPI

Rubygems

Stay in touch

Get open source security insights delivered straight into your inbox.

  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc