🔮 Primitives to build simple, flexible, WAI-ARIA compliant React autocomplete components

The problem

You need an autocomplete experience in your application and you want it to be accessible. You also want it to be simple and flexible to account for your use cases.

This solution

This is a collection of primitive components that you can compose together to create an autocomplete component which you can reuse in your application. It's based on ideas from the talk "Compound Components" which effectively gives you maximum flexibility with a minimal API because you are responsible for the rendering of the autocomplete components.

This differs from other solutions which render things for their use case and then expose many options to allow for extensibility causing an API that is less easy to use and less flexible as well as making the implementation more complicated and harder to contribute to.

Installation

This component is currently under development and is not yet released...

This module is distributed via npm which is bundled with node and should be installed as one of your project's dependencies:

npm install --save react-autocompletely

This package also depends on react and prop-types. Please make sure you have those installed as well.

Usage

Things are still in flux a little bit (looking for feedback).

import Autocomplete from 'react-autocompletely'

// use components together here.

Available components and relevant props:

Autocomplete

This is the main component. It renders a div and forwards props. Wrap everything in this.

onChange

function(item: any) | required

Called when the user selects an item

Autocomplete.Input

This is the input component. It renders an input and forwards props.

defaultValue

string / null | defaults to null

The initial value the input should have when it's mounted.

getValue

function(item: any) | defaults to an identity function (i => String(i))

Used to determine the inputValue for the selected item.

Autocomplete.Controller

This component allows you to receive and interact with the state of the autocomplete component.

children

function({}) | required

This is called with an object with the properties listed below:

Autocomplete.Menu

This component allows you to render the items based on the user input. It renders a div with another div for your items and a div for the menu status (for accessibility purposes)

defaultHighlightedIndex

number/null | defaults to null

This is the initial index to highlight when the menu first opens.

children

function({}) | required

This is called with the same things that the children prop is called with for Autocomplete.Controller

Autocomplete.Item

Render your items inside this component. This renders a div and forwards all props.

index

number | required

this is how react-autocompletely keeps track of your item when updating the highlightedIndex as the user keys around.

value

any | required

This is the item data that will be selected when the user selects a particular item.

Examples

Examples exist on codesandbox.io:

• react-autocompletely Apollo example

If you would like to add an example, follow these steps:

1. Fork this codesandbox
2. Update the code for your example (add some form of documentation to explain what it is)
3. Update the title and description
4. Add the tag: react-autocompletely:example

Inspiration

I was heavily inspired by Ryan Florence and his talk entitled: "Compound Components". I also took a few ideas from the code in react-autocomplete and jQuery UI's Autocomplete.

You can watch me build the first iteration of react-autocompletely on YouTube:

• Part 1
• Part 2

Other Solutions

You can implement these other solutions using react-autocompletely, but if you'd prefer to use these out of the box solutions, then that's fine too:

• react-select
• react-autocomplete

Contributors

Thanks goes to these people (emoji key):

Kent C. Dodds 💻 📖 🚇 ⚠️	Jack Moore 💡	Travis Arnold 💻 📖	Jeremy Gayed 💡	Haroen Viaene 💡	monssef 💡	Federico Zivolo 📖
Divyendu Singh 💡

This project follows the all-contributors specification. Contributions of any kind welcome!

LICENSE

MIT