selector-observer
selector-observer allows you to monitor DOM elements that match a CSS selector. Rather than imperatively querying the DOM, register an observer for a CSS selector and trigger behavior whenever those elements appear on the page. Typical uses include: registering event handlers and initializing a component or plugin.
Why
It is a common jQuery practice to initialize code inside a jQuery ready function.
The .ready() method offers a way to run JavaScript code as soon as the page's Document Object Model (DOM) becomes safe to manipulate.
$(document).ready(function() {
$('p.hello').text('The DOM is now loaded and can be manipulated.')
})
The issue with this approach is that this initialization code is run only once per page load. Modern web pages change over time after the initial load. This example depends on a <p class="hello">
being rendered on the server and being on the initial page load. However, if it is inserted later via AJAX or dynamically through some other JS, it will not be found.
Instead of imperatively querying the DOM, we could declare the selectors we needed to operate on.
import {observe} from 'selector-observer'
observe('p.hello', {
add(el) {
$(el).text('The DOM is now loaded and can be manipulated.')
}
})
Using the observe
helper, any matching element matching p.hello
will trigger the add
hook running our initialization code.
Usage
Two types of APIs are provided: a functional singleton API and a class-based API that allows you to change the scope of observation.
The observe
function will install an observer on the current document
.
import {observe} from 'selector-observer'
observe('.foo', {
add(el) {
console.log(el, 'added to', document)
},
remove(el) {
console.log(el, 'removed from', document)
}
})
Alternatively, the class-based SelectorObserver
allows you to configure the root element. This API is similar to MutationObserver
.
import SelectorObserver from 'selector-observer'
const rootElement = document.getElementById('root')
const observer = new SelectorObserver(rootElement)
observer.observe('.foo', {
add(el) {
console.log(el, 'added to', rootElement)
},
remove(el) {
console.log(el, 'removed from', rootElement)
}
})
Use Cases
Event Handlers
selector-observer
can help automatically install event handlers on any matching elements and ensure a cleanup stage ran. This is often necessary instead of using an event delegation technique if the event does not bubble.
import {observe} from 'selector-observer'
function handleMouseEnter(event) {
event.currentTarget.classList.add('active')
}
function handleMouseLeave(event) {
event.currentTarget.classList.remove('active')
}
observe('.dropzone', {
add(el) {
el.addEventListener('mouseenter', handleMouseEnter)
el.addEventListener('mouseleave', handleMouseLeave)
},
remove(el) {
el.removeEventListener('mouseenter', handleMouseEnter)
el.removeEventListener('mouseleave', handleMouseLeave)
}
})
Initialize third-party component or plugin
Many third-party component or plugin libraries require a manual initialization step to be installed on a given element. The add and remove hooks can be used if the plugin provides a cleanup hook. Often these libraries omit that kind of API too. To work around this, observe
provides an initialize
hook that only runs once per a given element.
This example initializes the tippy
tooltip library on any btn
buttons on the page.
import {observe} from 'selector-observer'
import tippy from 'tippy'
observe('.btn', {
initialize(el) {
tippy(el)
}
})
Event Delegation vs. Direct Binding
There are two established patterns for attaching event handlers to elements: direct binding and event delegation.
Direct binding means calling addEventListener
directly on an element instance. The downside of this is you need to first find the element to attach the event handler. Often this is done on page load. However, this misses elements that may be added dynamically after the page loads.
Event delegation is a technique where event handlers are registered to a CSS selector and are matched against all triggered events. The advantage is that it matches elements that are dynamically added and removed from the page. There's also less performance overhead to registering event handlers as you do not need to query the DOM upfront.
However, there are cases where event delegation does not work, or there may be a significant performance overhead to doing so. selector-observer
can be used with direct binding to discover elements on page load and those dynamically added later.
Here's an example of using the delegated-events library to install a click handler
import {on} from 'delegated-events'
function handleClick(event) {
console.log('clicked', event.currentTarget)
}
on('click', '.foo', handleClick)
Similarly using selectors-observer
using direct binding.
import {observe} from 'selector-observer'
function handleClick(event) {
console.log('clicked', event.currentTarget)
}
observe('.foo', {
add(el) {
el.addEventListener('click', handleClick)
},
remove(el) {
el.removeEventListener('click', handleClick)
}
})
Both accomplish similar tasks. However, in this example, using delegated-events
would be preferred as there's little upfront overhead to installing the click handler. Each selector-observer
registration adds a little bit of startup overhead.
However, not all events types will work with delegated-events
. For the event to work, it must bubble. Specific event types do not bubble like mouseenter
and mouseleave
. In this case, use selector-observer
and direct bind.
Advanced Usage
constructor matching
When using the flow type checker, elements passed to add and remove hooks will be typed as Element
. CSS selectors could match any HTML element and even SVG elements.
observe('.any-element', {
add(el /*: Element */) {},
remove(el /*: Element */) {}
})
If you know the target element is a specific HTML element like a <form>
, you can specify the constructor of the element to inform the type checker.
observe('.some-form', {
constructor: HTMLFormElement,
add(form /*: HTMLFormElement */) {},
remove(form /*: HTMLFormElement */) {}
})
initialize hook
While defining add
and remove
hooks is preferred, a third initialize
hook exists to enable use cases where plugins or components do not provide proper teardown APIs. Where add
will run multiple times per Element
instance, initialize
will only run once.
import {observe} from 'selector-observer'
observe('.foo', {
initialize(el) {
console.log('initialize')
},
add(el) {
console.log('add')
},
remove(el) {
console.log('remove')
}
})
const el = document.createElement('div')
document.body.appendChild(el) // log: initialize, add
document.body.removeChild(el) // log: remove
document.body.appendChild(el) // log: add
State via initialize closure
The initialize
hooks allow for a special return value of {add: Function, remove: Function}
to dynamically set hooks for the given element instance. This enables private state to be captured in the initialize closure and shared between add and remove hooks.
import {observe} from 'selector-observer'
observe('.foo', {
initialize(el) {
let counter = 0
return {
add() {
counter++
},
remove() {
counter--
}
}
}
})
It is also useful for defining per element event handlers that can access private state inside the initialization closure.
import {observe} from 'selector-observer'
observe('.foo', {
initialize(el) {
let isMouseOver = false
function handleMouseEnter() {
isMouseOver = true
}
function handleMouseLeave() {
isMouseOver = false
}
return {
add() {
el.addEventListener('mouseenter', handleMouseEnter)
el.addEventListener('mouseleave', handleMouseLeave)
},
remove() {
el.removeEventListener('mouseenter', handleMouseEnter)
el.removeEventListener('mouseleave', handleMouseLeave)
}
}
}
})
Observables
RxJS—and other compatible observable libraries—that return a subscription object can automatically be disposed when elements are removed from the page.
import {observe} from 'selector-observer'
import {fromEvent} from 'rxjs'
observe('button', {
subscribe: button => fromEvent(button, 'click').subscribe(() => console.log('Clicked!'))
})