react-laag provides a couple of tools to position UI elements such as tooltips and popovers with ease. It lets you focus on how your UI should look, feel and behave, by taking care of the heavy lifting such as complex calculations you would otherwise have to do yourself.
Try it out for yourself here, or see some examples in our storybook.
Click here for the v1 documentation, or read the release-notes for migrating to v2.
- 📦 Only 8kb minified & gzipped / tree-shakable / no dependencies
- 🛠 We do the positioning, you do the rest. You maintain full control over the look and feel.
- 🚀 Optimized for performance / no scroll lag whatsoever
- 🏗 Comes with sensible defaults out of the box, but you can tweak things to your liking
If you are working on your own UI library / design-system, or just struggling with some complex auto-complete-select component, react-laag might be a match for you. The flexibility react-laag provides has a small price however: you still have to do some work regarding styling and animations yourself. This pattern is also referred to as headless UI.
So, if you're looking for a full-fledged component out-of-the-box, I recommend to check out the wide range of excellent components already out there.
# NPM
npm install react-laag
# Yarn
yarn add react-laagThis library is build with TypeScript, so type-definitions are shipped out-of-the-box.
We're only scratching the surface here, but here's a quick example to get some sense what this library feels like and how to get going.
import * React from "react";
import { useLayer, useHover, Arrow } from "react-laag";
function Tooltip({ children, content }) {
const [isOver, hoverProps] = useHover();
const {
triggerProps,
layerProps,
arrowProps,
renderLayer
} = useLayer({
isOpen: isOver
});
return (
<>
<span {...triggerProps} {...hoverProps}>
{children}
</span>
{isOver &&
renderLayer(
<div className="tooltip" {...layerProps}>
{content}
<Arrow {...arrowProps} />
</div>
)}
</>
);
}In order to use this <Tooltip /> component:
const someContent = (
<div>
When you hover <Tooltip content="I'm a tooltip!">this</Tooltip> word, you
should see a tooltip
</div>
);The most important hook for positioning and rendering the layer.
import { useLayer } from "react-laag";(options: UseLayerOptions): UseLayerProps;| name | type | required | default | description |
|---|---|---|---|---|
| isOpen | boolean |
✔ | signals whether the layer is open or closed | |
| overflowContainer | boolean |
true |
should the layer be contained within the closest scroll-container (false), or is the layer allowed to overflow its closest scroll-container (true)? |
|
| placement | string |
"top-center" |
preferred placement of the layer. One of: "top-center" / "top-start" / "top-end" / "left-start" / "left-center" / "left-end" / "right-start" / "right-center" / "right-end" / "bottom-start" / "bottom-center" / "bottom-end" / "center" |
|
| possiblePlacements | string[] |
all | in case of auto: true, describes which placements are possible. Default are all placements. |
|
| preferX | "left" | "right" |
"right" |
in case of auto: true, when both left and right sides are available, which one is preferred? Note: this option only has effect when placement is "top-*" or "bottom-*". |
|
| preferY | "top" | "bottom" |
"bottom" |
in case of auto: true, when both top and bottom sides are available, which one is preferred? Note: this option only has effect when placement is "left-*" or "right-*" |
|
| auto | boolean |
false |
should we switch automatically to a placement that is more visible on the screen? | |
| snap | boolean |
false |
in case of auto: true, should we stick to the possible placements (true), or should we gradually move between two placements (false) |
|
| triggerOffset | number |
0 |
distance in pixels between layer and trigger | |
| containerOffset | number |
10 |
distance in pixels between layer and scroll-containers | |
| arrowOffset | number |
0 |
minimal distance between arrow and edges of layer and trigger | |
| layerDimensions | (layerSide: LayerSide): { width: number, height: number } |
lets you anticipate on the dimensions of the layer. Useful when the dimensions of the layer differ per side, preventing an infinite loop of re-positioning | ||
| onDisappear | (type: "partial" | "full"): void |
gets called when the layer or trigger partially or fully disappears from the screen when the layer is open. If overflowContainer is set to true, it looks at the trigger element. If overflowContainer is set to false, it looks at the layer element. |
||
| onOutsideClick | (): void |
gets called when user clicks somewhere except the trigger or layer when the layer is open | ||
| onParentClose | (): void |
Useful when working with nested layers. It is used by the parent layer to signal child layers that their layers should close also. | ||
| container | HTMLElement | (): HTMLElement | string |
Specify in which container (html-element) the layers should mount into when overflowContainer is set to true or when there's no scroll-container found. By default, in such cases the layers are mounted into a generated div#layers which gets attached to the body of the document. This prop accepts various values. When an string is passed, it is interpreted as the id of an element. |
||
| trigger | object |
This prop let's you specify information about the trigger you don't know beforehand. This is typically for situations like context-clicks (right-mouse-clicks) and text-selection. By using this prop the returning triggerProps of this hook will have no effect. |
||
getBounds: () => ClientRect |
A callback function that returns the bounds of the trigger. | |||
getParent?: () => HTMLElement |
A callback function that returns the parent element. This is optional but may be needed in cases where you'll want to prevent overflow of the layer. In other words, if you use the default option overflowContainer: true, this callback will have no effect. The returning element is used to position the layer relatively and to register event-listeners. |
|||
| environment | Window |
window |
useful when working with i-frames for instance, when things like event-listeners should be attached to another context (environment). | |
| ResizeObserver | ResizeObserverClass |
pass a polyfill when the browser does not support ResizeObserver out of the box |
| name | type | required | description |
|---|---|---|---|
| triggerProps | object |
✔ (unless the trigger-option is used) | Spread these props on the trigger-element |
ref: () => void |
Obtains a reference to the trigger-element | ||
| layerProps | object |
✔ | Spread these props on the layer-element |
ref: () => void |
Obtains a reference to the layer-element | ||
style: CSSProperties |
style-object containing positional styles | ||
| arrowProps | object |
Spread these props on the arrow-component | |
ref: () => void |
Obtains a reference to the arrow-element | ||
style: CSSProperties |
style-object containing positional styles | ||
layerSide: LayerSide |
let the arrow-component know in which direction it should point | ||
| renderLayer | (children: ReactNode) => ReactPortal |
Render the layer inside this function. Essentially, this is a wrapper around createPortal() |
|
| layerSide | "top" | "bottom" | "right" | "left" |
The side the layer is currently on relative to the trigger | |
| triggerBounds | ClientRect | null |
Bounds of the trigger when isOpen: true. Useful when sizing the layer relatively to the trigger. |
Utility hook for managing hover behavior.
import { useHover } from "react-laag";(options?: UseHoverOptions): [boolean, UseHoverProps, () => void];Example usage
const [
isOver, // should we show the layer?
hoverProps, // spread these props to the trigger-element
close // optional callback to set `isOver` to `false`
] = useHover({
delayEnter: 300, // wait 300ms before showing
delayLeave: 300, // wait 300ms before leaving
hideOnScroll: true // hide layer immediately when user starts scrolling
});| name | type | required | default | description |
|---|---|---|---|---|
| delayEnter | number |
0 |
delay in ms | |
| delayLeave | number |
0 |
delay in ms | |
| hideOnScroll | boolean |
true |
sets hovering to false when user starts scrolling |
| name | type |
|---|---|
| onMouseEnter | () => void |
| onMouseLeave | () => void |
| onTouchStart | () => void |
| onTouchMove | () => void |
| onTouchEnd | () => void |
import { Arrow } from "react-laag";
<Arrow
angle={45}
size={8}
roundness={0}
borderWidth={0}
borderColor="#000"
backgroundColor="#FFF"
layerSide="top"
/>;<Arrow /> is basically just a regular svg-element, so it will accept all default svg-props as well
| name | type | required | default | description |
|---|---|---|---|---|
| angle | number |
45 |
Angle of the triangle in degrees. A smaller angle means a more 'pointy' arrow. | |
| size | number |
8 |
distance in pixels between point of triangle and layer. | |
| roundness | number |
0 |
Roundness of the point of the arrow. Range between 0 and 1. |
|
| borderWidth | number |
0 |
Width of the border in pixels | |
| borderColor | string |
"black" |
Color of the border | |
| backgroundColor | string |
"white" |
Color of the arrow | |
| layerSide | string |
"top" |
Determines where to arrow should point to |
Utility hook that lets you use the mouse-position as source of the trigger. This is useful in scenarios like context-menus.
import { useMousePositionAsTrigger } from "react-laag";(options?: UseMousePositionAsTriggerOptions): UseMousePositionAsTriggerProps;
type UseMousePositionAsTriggerProps = {
hasMousePosition: boolean;
resetMousePosition: () => void;
handleMouseEvent: (evt: MouseEvent) => void;
trigger: {
getBounds: () => ClientRect;
getParent?: () => HTMLElement;
};
parentRef: MutableRefObject;
};Example usage
function ContextMenu() {
const {
hasMousePosition,
handleMouseEvent,
resetMousePosition,
trigger
} = useMousePositionAsTrigger();
const { layerProps, renderLayer } = useLayer({
isOpen: hasMousePosition,
onOutsideClick: resetMousePosition,
trigger
});
return (
<>
<div onContextMenu={handleMouseEvent}>Right-click to show the layer</div>
{hasMousePosition && renderLayer(<div>Layer</div>)}
</>
);
}See the context-menu example or text-selection example for more info.
| name | type | required | default | description |
|---|---|---|---|---|
| enabled | number |
true |
Should the mouse-position currently be actively tracked? | |
| preventDefault | boolean |
true |
Should handleMouseEvent preventDefault()? |
Utility function that lets you assign multiple references to a 'ref' prop.
import * as React from "react";
import { mergeRefs } from "react-laag";
const ref1 = React.useRef();
const ref2 = element => console.log(element);
<div ref={mergeRefs(ref1, ref2)} />;react-laag allows you to use to methods or modes for positioning with help of the overflowContainer option in useLayer(). When using overflowContainer: true, which is the default behavior, the layer is mounted somewhere high in the document in its own container. In such a case, the position of the layer will be fixed, meaning that it will be positioned relative to the window.
On the other hand, you can decide you don't want to overflow the container by setting overflowContainer to false. In this scenario the layer will be mounted right under the scroll-container.
So, what do we mean by the term 'scroll-container' anyways? react-laag considers a scroll-container an element which has set the overflow, overflow-x or overflow-y style to one of "auto" or "scroll". react-laag tries to find these scrollable-containers by traversing up the dom-tree, starting with the trigger-element. This way, the layer will be positioned relatively to the closest scroll-container. There's one catch though: it expects you to set the position: relative style on this scroll-container. If you accidentally forgot to set this style, react-laag will output a friendly warning in the console.
This usually is something you don't have to think about, but in some cases it may come in handy.
When setting the auto option to true in useLayer(), react-laag will create an priority-order under the hood. The preferred placement will always be on top of the list, meaning this placement will be tried first. To determine the placements after that, react-laag looks at the following things:
- the preferred placement for determining the preferred direction / axis. When using
"top-start"for instance, we can assume that although this exact placement may not fit, somewhere on top is still preferred. This direction / axis will have more priority overpreferX/preferY. preferX/preferYfor determining priority on the opposite axis regarding the preferred placement.- The next placement in line must always be as close to the previous placement as possible.
- placements which are not defined in
possiblePlacements(all by default) are skipped
Let's look at an example given placement "top-start" with a preferX of "right":
top-start -> top-center -> top-end -> right-end -> left-end -> right-center -> left-center -> right-start -> left-start -> bottom-start -> bottom-center -> bottom-end
During rendering react-laag will given the list containing priorities...
- try to find the first placement in line that fits the current screen / layout
- if none fits, it will find the placement with the most visible surface
Nesting multiple layer often occurs in large menus where items are grouped. If you're looking for an example of how to accomplish this, be sure to check out the example about nesting.
There are however some important things to consider. How do we for instance signal to the rest of the nested layers, that a layer higher up in the hierarchy has just closed? Fortunately, there's a special option for that in the userLayer() options: onParentClose. react-laag uses context under the hood to monitor which layers are related to each other. This has a couple of implications:
- When a layer closes, it will signal child-layers below to close as well through
onParentClose. onOutsideClickonly has effect on root-layer. react-laag has a kind of event-bubbling system under the hood to make sure that the root-layer doesn't close when some child-layer down below was clicked. When there was a solid click outside somewhere in the document, the root-layer will signal the rest of the layers to close as well.
react-laag doesn't do any animations for you. Why? Because we want to focus this library purely on positioning and there are a lot of libraries out there who do a far better job than react-laag could ever do.
Since renderLayer is just an abstraction over React's createPortal you can in theory use any form of animation you'd like.
Personally, I'm a big fan of framer-motion, so I will show you a quick example to get started:
import { useLayer } from "react-laag";
import { motion, AnimatePresence } from "framer-motion";
function AnimationExample() {
const [isOpen, setOpen] = React.useState(false);
const { renderLayer, triggerProps, layerProps } = useLayer({ isOpen });
return (
<>
<button {...triggerProps} onClick={() => setOpen(!isOpen)}>
Click me!
</button>
{renderLayer(
<AnimatePresence>
{isOpen && (
<motion.div
{...layerProps}
initial={{ opacity: 0 }}
animate={{ opacity: 1 }}
exit={{ opacity: 0 }}
>
Layer
</motion.div>
)}
</AnimatePresence>
)}
</>
);
}By design react-laag doesn't handle any z-indexes for you. There are too many different use-cases and scenarios possible for this library to manage. You are free to implement your own z-index strategy. However, there is a cheap fix that will probably fix 95% of your problems.
By default, react-laag renders your layers in a container right under the document's body:
<body>
<!-- React's entry -->
<div id="root"></div>
<!-- By default all layers will be rendered here -->
<div id="layers"></div>
</body>Now, nothing is stopping you to do this:
#layers {
z-index: 1000;
}All layers will now automatically inherit the z-index of this container.
If you want react-laag to mount the layers into another element, you have two options:
- use the
containeroption inuseLayer():
const {} = useLayer({
// pass in an id of the element
container: "my-own-container-id",
// pass in a callback returning an html-element
container: () => myContainer,
// pass in a html-element directly
container: myContainer
});- set the container globally with
setGlobalContainer():
// somewhere in the root of your application
import { setGlobalContainer } from "react-laag";
// works the same as the container-option above
setGlobalContainer("my-own-container-id");If you want to take full advantage of react-laag's positioning change detection, make sure your target browser(s) support ResizeObserver. To get a detailed list which browsers support this feature consult Can I use.
As of now, this sort of means all modern browsers except IE 11.
If you need to support IE 11 you can optionally provide your app with a polyfill.
If you don't want to pollute the global context you can also pass in the polyfill via the option in useLayer:
import ResizeObserver from "resize-observer-polyfill";
import { useLayer } from "react-laag";
useLayer({ ResizeObserver });Is there support for accessability?
No, unfortunately not. There are two primary reasons:
- This library is primary focussed around positioning
- Accessability not my area of expertise
I'm open to the idea in the future though. I would be happy to get some help with this!
Which browsers are supported?
react-laag works on all modern browsers. Is should also work in >= IE 11, although this may require a polyfill for stable-features
Will this work with server-side-rendering?
Yes, each build a small tests gets run in other to test compatibility.
Want to contribute to react-laag? Your help is very much appreciated! Please consult the contribution guide on how to get started.
MIT © everweij