Skip to content

Highly customizable client-side 'auto-complete' component for large datasets

License

Notifications You must be signed in to change notification settings

jgliner/react-universal-search

Repository files navigation

react-universal-search

Highly customizable client-side 'auto-complete' component for large datasets


Development

PRs are more than welcome!

@TODO: Throw up a CONTRIBUTING.md In the meantime, all I ask is ES6 only, please!

  1. git clone
  2. cd react-universal-search
  3. npm install
  4. npm install --no-save react react-dom
  5. npm run dev (starts webpack-dev-server)
  6. Navigate to localhost:9000

Once you are done making changes, run npm run build. This will create the main minified distribution bundle, react-universal-search.js. Please commit this file.


Options

API

Property Type Default
customComponent func(item, iterator) undefined
customMatchCountComponent func(count) undefined
customNoMatchComponent element null
focusedOnly bool false
hasCategories bool detected
limitResults int 0 (all)
listToSearch obj undefined
parseMethod string ['greedy', 'strict', 'symbol-permissive'] 'greedy'
placeholder string undefined
showMatchCount bool false
showWhenNoMatches bool false

Details

Property Description
customComponent Component called for each individual item passed in via listToSearch. If you want to add conditional rendering logic/styles to each entry and/or category header, this would be the place to do it. If no function is specified, the default layout is an unstyled list of <div> elements. -- More info in Custom Components below...
customMatchCountComponent Called on every keystroke. Callback includes the number of results that match the given query.
customNoMatchComponent A React element that displays when exactly zero results match the given query string.
focusedOnly If set to true, will hide the results box onBlur and show onFocus
hasCategories This library will try and detect categories in your formatting automatically, to use as headers. It will always detect false for non-Object datatypes, as well as Arrays. Pass a boolean in here to override either of these behaviors.
limitResults If listToSearch has categories, this will limit the number of results displayed per category. If listToSearch is an array, it will limit all results
listToSearch Accepted Input Formats
IMPORTANT: one property MUST be called name. This will be the main search string. If this gets to be inconvenient, I will add another parameter for custom Key names
parseMethod How the query string should be interpreted. See Below for details.
placeholder Placeholder text that a user will see before typing (same as in an HTML5 <input> element)
showMatchCount Show total results after each keystroke. If this is true without a corresponding customMatchCountComponent, a default component will be used. NOTE: This option will be overridden if a valid customMatchCountComponent is passed in.
showWhenNoMatches Display a message when exactly zero results match the given query string. If this is true without a corresponding customNoMatchComponent, a default component will be used. NOTE: This option will be overridden if a valid customNoMatchComponent is passed in.

Accepted Input Formats

Plain Array of Objects (i.e. No Categories):

[
  { name: 'foo' },
  { name: 'bar' },
  { name: 'baz' },
  ...
]

Keyed Objects Containing Arrays:

{
  foo: [
    {name: 'bar'},
    ...
  ],
  baz: [
    {name: 'qux'},
    ...
  ],
  ...
}

As long as this general structure is valid, anything else can be passed in at the inner-object level (e.g. strings, numbers, elements, even more <UniversalSearch> components... though I can't think of a single use case for that...)


Custom Components

You'll likely want to use your own React components instead of the ugly default pile of <div>s. Below is some key information on how I simplified this in order to make react-universal-search as customizable as possible while still retaining relative speed and space efficiency:

This library temporarily injects two properties into each entry within your listToSearch only if a valid category format is detected/enabled:

entry._category - string - A copy of the item's containing Key

entry._firstInCategory - bool - Will be true if this is the first entry in a given category, or undefined if false. Use this to conditionally render your headers.

this.props.listToSearch = {
  foo: [
    { name: 'bar' },
    { name: 'baz' },
  ],
  qux: [
    { name: 'quux' },
    { name: 'corge' },
  ],
};

// ...in your customComponent

console.log(foo[0].name)                 // 'bar'
console.log(foo[0]._category)            // 'foo'
console.log(foo[0]._firstInCategory)     // true

console.log(foo[1].name)                 // 'baz'
console.log(foo[1]._category)            // 'foo'
console.log(foo[1]._firstInCategory)     // undefined

console.log(qux[0].name)                 // 'quux'
console.log(qux[0]._category)            // 'qux'
console.log(qux[0]._firstInCategory)     // true

console.log(qux[1].name)                 // 'corge'
console.log(qux[1]._category)            // 'qux'
console.log(qux[1]._firstInCategory)     // undefined

Example


Parsing Methods

greedy (Default):

Attempts to match a string from any starting point. Beyond that, the results will take on the order in which they were passed in.

strict:

Attempts to match a string starting from the beginning. All other matches are discarded

symbol-permissive:

Attempts to match a string starting from the beginning, but will also match if there are symbols or whitespaces (/[\W\s]/) leading up to the matching string.

this.props.listToSearch = {
  foo: [
    { name: 'bar' },
    { name: 'baz' },
  ],
  qux: [
    { name: 'quux' },
    { name: 'corge' },
    { name: 'foobar' },
  ],
  grault: [
    { name: '[bar]bazqux' },
    { name: 'bazqux' },
  ],
};

// <input />
e.target.value = 'ba';

// RESULTS //

// 'greedy'

// spread operator to represent Set()
[...this.state.results] = [
  { name: 'bar', _category: 'foo', _firstInCategory: true },
  { name: 'baz', _category: 'foo' },
  { name: 'foobar', _category: 'qux' },
  { name: '[bar]bazqux', _category: 'grault', _firstInCategory: true },
  { name: 'bazqux', _category: 'grault' },
]; // 5 matches

// 'strict'

// spread operator to represent Set()
[...this.state.results] = [
  { name: 'bar', _category: 'foo', _firstInCategory: true },
  { name: 'baz', _category: 'foo' },
  { name: 'bazqux', _category: 'grault' },
]; // 3 matches

// 'symbol-permissive'

// spread operator to represent Set()
[...this.state.results] = [
  { name: 'bar', _category: 'foo', _firstInCategory: true },
  { name: 'baz', _category: 'foo' },
  { name: '[bar]bazqux', _category: 'grault', _firstInCategory: true },
  { name: 'bazqux', _category: 'grault' },
]; // 4 matches

@TODO:

  • Docs
  • Optimize
  • Testing Suite
  • Publish to npm

Data Source for Example: https://www.kaggle.com/nickhould/craft-cans

Database released under Open Database License, individual contents under Database Contents License

About

Highly customizable client-side 'auto-complete' component for large datasets

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages