Run big operations in a separate thread using web workers.
Use in combination with kea-parallel-loader, worker-loader and possibly also redux-saga
To init:
npm install --save kea-parallel
npm install --save-dev kea-parallel-loader worker-loader
Then:
// ./long-worker.js
export function worker (input) {
console.log('In long worker', input)
let output = 0
// a very slow sum
for (var i = 0; i < input.count; i++) {
output += 1
}
return { status: 'done', result: output }
}without redux-saga:
// in ./index.js
import { runInParallel } from 'kea-parallel'
import longWorker from 'worker!kea-parallel!./long-worker'
const start = new Date().getTime()
runInParallel(longWorker, {count: 1000 * 1000 * 1000}).then(result => {
const end = new Date().getTime()
const time = end - start
console.log(result)
console.log(`Execution time: ${time / 1000} sec`)
})with redux-saga:
// in ./saga.js
import { runInParallel } from 'kea-parallel'
import longWorker from 'worker!kea-parallel!./long-worker'
export default function * saga () {
console.log('Starting saga')
const start = new Date().getTime()
const result = yield runInParallel(longWorker, {count: 1000 * 1000 * 1000})
const end = new Date().getTime()
const time = end - start
console.log(result)
console.log(`Execution time: ${time / 1000} sec`)
}This outputs:
Starting saga
In long worker Object {count: 1000000000}
Object {status: "done", result: 1000000000}
Execution time: 8.892 sec
Everything inside 'long-worker.js' will be run in a separate thread, making the entire app responsive despite the huge blocking for loop.
In case you choose to call all your background workers *-worker.js, feel free to add this to your webpack config, above the line with babel-loader:
// webpack 2
{
test: /\-worker\.js$/,
exclude: /node_modules/,
loaders: [
'worker',
'kea-parallel'
]
},
// webpack 1
{
test: /\-worker\.js$/,
exclude: /node_modules/,
loaders: 'worker!kea-parallel'
},Currently there is no thread pooling. Spawning a lot of workers is still untested.