js-dom-replace

JS-DOM-Replace by Murad Kasim is an improved version of findAndReplaceDOMText (created originally James Padosley).

JS-DOM-Replace searches the DOM for words and replaces them without affecting HTML tags, JavaScript or inline CSS. This is very useful if you have to modify the DOM text, since if you use regular find/replace it will affect tags and scripts too. With this project you are also able to add ids, titles and inline CSS styles to every replacement element.

Added functionalities: Add id, title and CSS Style to each found element and much more!:

List of additionally supported options:

• id - element ID (<div id="element-unique-id"></div>)
• title - element Title which is shown over mouse hover (<div title="This is hover title!"></div>)
• style - direct inline style (<div style="color: red;"></div>)

Example:

findAndReplaceDOMText(document.body, {
    find: 'amsterdam',
    replace: 'frankfurt',
    wrap: 'mark',
    wrapClass: 'container-replaced-word',
    id: 'unique-element-id-amsterdam',
    title: 'When you hover with the mouse over the replaced text you will read this',
    style: 'color: red; cursor: pointer;'
});

For example:

<p id="t">
  123 456 Hello
</p>

findAndReplaceDOMText(document.getElementById('t'), {
  find: /Hello/,
  wrap: 'em'
});

This would result in:

<p id="t">
  123 456 <em>Hello</em>
</p>

And it also works when matches are spread across multiple nodes! E.g.

<p id="t">
  123 456 Hell<span>o Goodbye</span>
</p>

findAndReplaceDOMText(document.getElementById('t'), {
  find: /Hello/,
  wrap: 'em'
});

This would result in:

<p id="t">
  123 456 <em>Hell</em><span><em>o</em> Goodbye</span>
</p>

The EM element has been added twice, to cover both portions of the match.

Installation

Grab the latest findAndReplaceDOMText.js (or npm install findandreplacedomtext) and include it as a <script> on your page.

API / Usage

findAndReplaceDOMText has the following argument signature:

findAndReplaceDOMText(
  element, // (Element) The element or text-node to search within
  options  // (Object) Explained below
);

API

Options

The options object includes:

• find (RegExp | String): Something to search for. A string will perform a global search by default (looking for all matches), but a RegExp will only do so if you include the global (/.../g) flag.
• replace optional (String | Function): A String of text to replace matches with, or a Function which should return replacement Node or String. If you use a string, it can contain various tokens:
• $n to represent the nth captured group of a regular expression (i.e. $1, $2, ...)
• $0 or $& to represent the entire match
• $` to represent everything to the left of the match.
• $' to represent everything to the right of the match.
• wrap optional (String | Node): A string representing the node-name of an element that will be wrapped around matches (e.g. span or em). Or a Node (i.e. a stencil node) that we will clone for each match portion.
• wrapClass optional (String): A string representing the class name to be assigned to the wrapping element (e.g. <span class="myClass">found text</span>). If the wrap option is not specified, then this option is ignored.
• portionMode optional (String, one of "retain" or "first"): Indicates whether to re-use existing node boundaries when replacing a match with text (i.e. the default, "retain"), or whether to instead place the entire replacement in the first-found match portion's node. Most of the time you'll want the default.
• filterElements optional (Function): A function to be called on every element encountered by findAndReplaceDOMText. If the function returns false the element will be altogether ignored.
• forceContext optional (Function | Boolean): A boolean or a boolean-returning function that'll be called on every element to determine if it should be considered as its own matching context. See below under Contexts for more info.
• preset optional (String): Currently there's only one preset: prose. See below.

preset:prose

The most common usage of findAndReplaceDOMText is to replace text found in regular prose, not all DOM nodes. To make this easier there is a preset that you can use to instruct it to:

• Ignore non-textual elements (E.g. <script>, <svg>, <optgroup>, <textarea>, etc.)
• Force bordered contexts on block-elements like <p> and <div> so that matches cannot cross element borders.
• Note: matches will still be able to cross textual-inline element borders (<em>, <span>, etc.)

To enable this preset:

findAndReplaceDOMText(element, {
  preset: 'prose',
  find: 'something',
  replace: 'something else'
})

What is a portion?

A portion or "match portion" is a part of a match that is delimited by node boundaries. Not all matches occur within a single text-node, so findAndReplaceDOMText has to be able to deal with cross-boundary matches (e.g. when matching /foo/ in "<em>f</em>oo").

A portion object has the following properties:

• node: The DOM node pertaining to the portion. Note that this node might not fully encapsulate part of a match, e.g. the node might contain additional text.
• index: The index of the portion (0 is the first portion of the match, etc.)
• text: The text of the portion relevant to the match
• indexInMatch: The index of the portion within the match
• indexInNode: The index of the portion text within the node

The replace Function

If you pass a function to the replace option your function will be called on every portion of every match and is expected to return a DOM Node (a Text or Element node). Your function will be passed both the portion and the encapsulating match of that portion.

E.g.

Input HTML

<div id="container">
  Explaining how to write a replace <em>fun</em>ction
</div>

JS

findAndReplaceDOMText(document.getElementById('container'), {
  find: 'function',
  replace: function(portion, match) {
    return '[[' + portion.index + ']]';
  }
});

Output HTML

<div id="container">
  Explaining how to write a replace <em>[[0]]</em>[[1]]
</div>

The wrap Option

If you pass a string to the wrap option, every matching text segment will be wrapped in that element. If you also specify the wrapClass option, the wrapping element will be assigned that class after it is created. This is useful for attaching various styles from your css.

E.g.

Input HTML

<div id="container">
  Explaining how to wrap text in elements with and without classes assigned.
</div>

JS

findAndReplaceDOMText(document.getElementById('container'), {
 find: 'without',
 wrap: 'em'
});
findAndReplaceDOMText(document.getElementById('container'), {
 find: 'with ',
 wrap: 'em',
 wrapClass: 'shiny'
});

CSS

.shiny {
 background-color: yellow;
}

Output HTML

<div id="container">
  Explaining how to wrap text in elements <em class="shiny">with </em>and <em>without</em> classes assigned.
</div>

The instance

Calling findAndReplaceDOMText returns an instance of an internal Finder constructor -- the API on the object is limited, at the moment, to reverting:

var finder = findAndReplaceDOMText(...);

// Later:
finder.revert();

Note: Reversion will only work if the nodes have not been tampered with after the initial replacement -- if there have been removals, movements or normalisations then the reversion is not guarenteed to work. In this case it's best to retain your own clone of the target node(s) in order to run your own reversion.

Contexts

Matching, by default, will occur on all elements and across all element borders. E.g.

Before:

<div id="test">
  <p>ama</p><p>zing</p>
</div>

findAndReplaceDOMText(document.getElementById('test'), {
  find: 'amazing',
  wrap: 'em'
});

After:

<div id="test">
  <p><em>ama</em></p><p><em>zing</em></p>
</div>

This is a useful feature for inline elements, but is undesirable in many other cases, so to stop it from happening you can choose to "force a context" on those particular elements. In this case we want to force a context on <p> elements:

findAndReplaceDOMText(document.getElementById('test'), {
  find: 'amazing',
  wrap: 'em',
  forceContext: function(el) {
    // Using https://developer.mozilla.org/en-US/docs/Web/API/Element/matches
    return el.matches('p');
  }
});

Internally, the prose preset uses this feature:

exposed.PRESETS = {
  prose: {
    forceContext: exposed.NON_INLINE_PROSE,
    filterElements: function(el) {
      return !hasOwn.call(exposed.NON_PROSE_ELEMENTS, el.nodeName.toLowerCase());
    }
  }
};

Changelog

• 1.0.1 (04 Dec 2019): Add options to add ID, Title and Inline CSS to every replaced element (Murad Kasim).
• 0.4.6 (5 Dec 2017): Fix indexInMatch (See #60). Fix undefined being echoed in optional+empty capture groups (See #70).
• 0.4.4 (4 May 2015): Remove duplicate key from NON_CONTIGUOUS_PROSE_ELEMENTS. Expose library via UMD (See #32).
• 0.4.3 (28 Apr 2015): Add preset:prose and forceContext features. See #29.
• 0.4.2 (30 Mar 2014): Fix IE to avoid incorrectly-closed-tags issue (see #20). Thanks to shauryamal!
• 0.4.1 (11 Mar 2014): Fix portionMode:first phantom nodes (see issue #19)
• 0.4.0 (6 Oct 2013): Major API overhaul, including a new arg signature (findAndReplaceDOMText(node, options), plus the ability to replace a match with text or wrap it with a DOM Node.
• 0.3.0: Switch to semver, add node-filtering feature (as requested in Issue #11
• 0.2: Fix case where regular expression contains word bounderies and add support for specifying a capture group to replace as the fourth argument to findAndReplaceDOMText() (see issue #5)
• 0.11: Minor fix: Make sure replacement node function is called in order of matches (see issue #4)
• 0.1: Initial commit + Fix for IE's broken HTML5 cloneNode (pull request)