Have ideas to improve npm?Join in the discussion! »

simple-web-worker

1.2.0 • Public • Published

SimpleWebWorker

An utility to simplify the use of web workers

Changelog

1.2.0

Highlights:

  • Added tests
  • Fixed object URL's not being revoked when worker is not being used anymore.

See full changelog here.

Why

Create and use Web Workers can be cumbersome sometimes. This plugin aims to facilitate the use of Web Workers.

How to install and use

yarn add simple-web-worker

// or

npm install simple-web-worker --save

Then:

import SWorker from 'simple-web-worker'

Obviously, you don't have to call it SWorker. You are free to use the name you want!

API

SWorker.run(func, [args]?)

Where:

  • func is the function to be runned in worker
  • [args] is an optional array of arguments that will be used by func

This method creates a disposable web worker, runs and returns the result of given function and closes the worker.

This method works like Promise.resolve(), but in another thread.

E.g.:

SWorker.run(() => 'SWorker run 1: Function in other thread')
  .then(console.log) // logs 'SWorker run 1: Function in other thread'
  .catch(console.error) // logs any possible error
 
SWorker.run((arg1, arg2) => `SWorker run 2: ${arg1} ${arg2}`, ['Another', 'function in other thread'])
    .then(console.log) // logs 'SWorker run 2: Another function in other thread'
    .catch(console.error) // logs any possible error

SWorker.create([actions]?)

Where:

  • [actions] is an optional array of objects with two fields, message and func. Essentially, it is a messages-actions map.

If [actions] is omitted or undefined, the created <worker> will have no registered actions, so you'll have to use the method register before you can use the <worker>.

If you plan to reuse a <worker>, you should use this method. It creates a reusable <worker> (not a real Web Worker, more on this ahead) with determined actions to be runned through its postMessage() or postAll() methods.

E.g.:

const actions = [
  { message: 'func1', func: () => `Worker 1: Working on func1` },
  { message: 'func2', func: arg => `Worker 2: ${arg}` },
  { message: 'func3', func: arg => `Worker 3: ${arg}` },
  { message: 'func4', func: (arg = 'Working on func4') => `Worker 4: ${arg}` }
]
 
let worker = SWorker.create(actions)

<worker>.postMessage(message, [args]?)

Where:

  • <worker> is a worker created with SWorker.create([actions])
  • message is one of the messages in [actions]
  • [args] is an optional array of arguments that will be used by the function registered with message

When the function does not expect any arguments or the expected arguments have default values, [args] can be omitted safely.

When the expected arguments do not have default values, [args] should be provided.

This method works like Promise.resolve(), but in another thread.

E.g.:

const actions = [
  { message: 'func1', func: () => `Worker 1: Working on func1` },
  { message: 'func2', func: arg => `Worker 2: ${arg}` },
  { message: 'func3', func: arg => `Worker 3: ${arg}` },
  { message: 'func4', func: (arg = 'Working on func4') => `Worker 4: ${arg}` }
]
 
let worker = SWorker.create(actions)
 
worker.postMessage('func1')
  .then(console.log) // logs 'Worker 1: Working on func1'
  .catch(console.error) // logs any possible error
 
worker.postMessage('func1', ['Ignored', 'arguments'])
  .then(console.log) // logs 'Worker 1: Working on func1'
  .catch(console.error) // logs any possible error
 
worker.postMessage('func2')
  .then(console.log) // logs 'Worker 2: undefined'
  .catch(console.error) // logs any possible error
 
worker.postMessage('func3', ['Working on func3'])
  .then(console.log) // logs 'Worker 3: Working on func3'
  .catch(console.error) // logs any possible error
 
worker.postMessage('func4')
  .then(console.log) // logs 'Worker 4: Working on func4'
  .catch(console.error) // logs any possible error
 
worker.postMessage('func4', ['Overwrited argument'])
  .then(console.log) // logs 'Worker 4: Overwrited argument'
  .catch(console.error) // logs any possible error

<worker>.postAll([message1,... || {message: message1, args: [args1]},... || [args1],...]?)

Where:

  • <worker> is a worker created with SWorker.create([actions])
  • The argument is an optional array which accepts one of the following:
    • message1,... - strings containing one or more of the messages in [actions]
    • {message: message1, args: [args1]},... - objects containing two fields, message (a message from actions) and args (the arguments to be used by the correspondent function)
    • [args1],... - arrays of arguments to be used by the registered actions.

If [message1,...] is undefined or no argument is given, <worker> will run all registered actions without arguments.

If [{message: message1, args: [args1]},...] or [[args1],...] is used, you should use [] (an empty array) as [args] for the functions that does not expect arguments, or if the respective argument of your function has a default value and you want it to be used. If you use [null] this will be the value assumed by function argument.

When using [[args1],...], you MUST input the same number of arguments as registered actions, even if some action doesn't accept any arguments! In that case use a [], as stated above. See examples below.

If [{message: message1, args: [args1]},...] is used, every object must contain the fields message and args.
This method works like Promise.all(), but in another thread.

E.g.:

const actions = [
  { message: 'func1', func: () => `Worker 1: Working on func1` },
  { message: 'func2', func: arg => `Worker 2: ${arg}` },
  { message: 'func3', func: arg => `Worker 3: ${arg}` },
  { message: 'func4', func: (arg = 'Working on func4') => `Worker 4: ${arg}` }
]
 
let worker = SWorker.create(actions)
 
worker.postAll()
  .then(console.log) // logs ['Worker 1: Working on func1', 'Worker 2: undefined', 'Worker 3: undefined', 'Worker 4: Working on func4']
  .catch(console.error) // logs any possible error
 
worker.postAll(['func1', 'func3'])
  .then(console.log) // logs ['Worker 1: Working on func1', 'Worker 3: undefined']
  .catch(console.error) // logs any possible error
 
worker.postAll([{ message: 'func1', args: [] }, { message: 'func3', args: ['Working on func3'] })
  .then(console.log) // logs ['Worker 1: Working on func1', 'Worker 3: Working on func3']
  .catch(console.error) // logs any possible error
 
worker.postAll([[], ['Working on func2'], ['Working on func3'], []])
  .then(console.log) // logs ['Worker 1: Working on func1', 'Worker 2: Working on func2', 'Worker 3: Working on func3', 'Worker 4: Working on func4']
  .catch(console.error) // logs any possible error
 
worker.postAll([[], ['func2'], ['func3'], ['Overwriting default value of arg on func4']])
  .then(console.log) // logs ['Worker 1: Working on func1', 'Worker 2: func2', 'Worker 3: func3', 'Worker 4: Overwriting default value of arg on func4']
  .catch(console.error) // logs any possible error

<worker>.register(action || [actions])

Where:

  • <worker> is a worker created with SWorker.create([actions])
  • action is an object with two fields, message and func
  • [actions] is an array of objects, and each object is an action, as defined above

You can use action or [actions], but not both at the same time.

E.g.:

const initialActions = [
  { message: 'func1', func: () => 'Working on func1' }
]
 
let worker = SWorker.create(initialActions)
 
worker.postAll()
  .then(console.log) // logs ['Working on func1']
  .catch(console.error) // logs any possible error
 
// registering just one action
worker.register({ message: 'func2', func: () => 'Working on func2' })
 
worker.postAll()
  .then(console.log) // logs ['Working on func1', 'Working on func2']
  .catch(console.error) // logs any possible error
 
// registering multiple actions
worker.register([
  { message: 'func3', func: () => 'Working on func3' },
  { message: 'func4', func: () => 'Working on func4' }
])
 
worker.postAll()
  .then(console.log) // logs ['Working on func1', 'Working on func2', 'Working on func3', 'Working on func4']
  .catch(console.error) // logs any possible error

<worker>.unregister(message || [messages])

Where:

  • <worker> is a worker created with SWorker.create([actions])
  • message is one of the messages in [actions]
  • [messages] is an array containing one or more messages, and each message is a message, as defined above

You can use message or [messages], but not both at the same time.

E.g.:

const initialActions = [
  { message: 'func1', func: () => 'Working on func1'},
  { message: 'func2', func: () => 'Working on func2'},
  { message: 'func3', func: () => 'Working on func3'},
  { message: 'func4', func: () => 'Working on func4'}
]
 
let worker = SWorker.create(initialActions)
 
worker.postAll()
  .then(console.log) // logs ['Working on func1', 'Working on func2', 'Working on func3', 'Working on func4']
  .catch(console.error) // logs any possible error
 
// unregistering just one action
worker.unregister('func2')
 
worker.postAll()
  .then(console.log) // logs ['Working on func1', 'Working on func3', 'Working on func4']
  .catch(console.error) // logs any possible error
 
// unregistering multiple actions
worker.unregister(['func3', 'func1'])
 
worker.postAll()
  .then(console.log) // logs ['Working on func4']
  .catch(console.error) // logs any possible error

Closing workers?

You may be thinking: "How do I terminate those reusable workers if there's no close() or terminate() methods?"

Well, when you create a reusable worker, you don't receive a real Web Worker.

Instead, you get an object which holds the given messages-actions map, and when you call postMessage() or postAll() it will, under the hood, call run() with the correspondent functions.

So, to "terminate" a "worker" when it is not needed anymore, you can just do:

let worker = SWorker.create(actions)
 
// use the worker
 
worker = null

Install

npm i simple-web-worker

DownloadsWeekly Downloads

3,040

Version

1.2.0

License

MIT

Last publish

Collaborators

  • avatar