#IdleQueueWorker
(WARNING! Still experimental library!)
Simple async queue implementation for low priority background tasks, such as analytics sending, data collecting.
Queue processing implemented over window.requestIdleCallback (shim included) that provides task processing on the main event loop, without impacting latency-critical events.
This module has no any external dependencies and some part of code base already used in production, enjoy.
npm install --save idle-queue-worker
- Use with your favorite module bundler such (e.g. Webpack, Browserify, etc)
Module provides a few interfaces: ForEachable
, Mapable
, Foldable
.
If you want to create your own interface, use QueueWorker
and object composition concept.
For lot of cases you just need ForEachable
interface and only two methods:
forEach
→ for dequeued data processing
enqueue
→ for passing some data to the queue.
##API
QueueWorker<T>(isReady?:() => boolean, timeout?:number, queue?:Array<T>)
Params ( aka queueWorkerArgs):
isReady
— call before each attempt of queue item processing. If result is falsy, then next new work will be sheduled.timeout (given in ms)
— minimum time that must pass before calling the callback function.queue
— not recommended for use in real tasks. Only for testing usage.
Methods:
.register(callback: (item:T) => any):void
— callback function for queue processing..enqueue(item:T):void
— add item at the end of the queue.
Example:
import { QueueWorker } from 'idle-queue-worker/lib';
const worker = new QueueWorker();
worker.register(x => x * x);
worker.enqueue(2);
Abstract(...queueWorkerArgs)<T,R>
Methods:
.enqueue(item:T):void
— proxy of queueWorker method.pipe(queue:AnotherQueue):AnotherQueue
— enqueue result of callback function to next queue
You can't construct Abstact instance directly, use classes below.
ForEachable(...queueWorkerArgs)<T>
Methods:
.forEach(callback: (item:T) => any):void
— callback function applied for each dequeued element
Example:
import { ForEachable } from 'idle-queue-worker/lib';
const worker = new ForEachable();
worker.forEach(x => x * x);
Mapable(...queueWorkerArgs)<T,R>
Methods:
.map(callback: (item:T) => R):Self
— a → b callback function.take(count:number):Promise<Array<R>>
— take first N processed items of queue.takeUntil(predicate:(item:R) => boolean):Promise<Array<R>>
— collect data until predicate result is falsy, then resolve result
Example:
import { Mapable } from 'idle-queue-worker/lib';
const worker = new Mapable();
worker
.map(x => x * x)
.take(10)
.then(result => send(result));
Foldable(...queueWorkerArgs)<T,R>
Methods:
.fold(callback:(accumulator:R, item:T) => R, initialValue:R):Self
— same as Array.prototype.reduce function.foldUntil(count:number):Promise<R>
— foldResult until predicate falsy
Example:
import { Foldable } from 'idle-queue-worker/lib';
const worker = new Foldable();
worker
.fold((sum, value) => sum * value, 1)
.foldUntil(sum => sum > 10)
.then(result => send(result));
Piping:
You can pass the results of one queue to next queue just call .pipe method of ForEachable
, Mapable
, Foldable
instances. There is many opportunities to determine your data flow with queues without pain.
Example:
import { ForEachable, Mapable, Foldable } from 'idle-queue-worker/lib';
const mapableWorker = new Mapable();
const forEachableWorker = new ForEachable();
const foldableWorker = new Foldable();
foldableWorker
.fold((result, element) => result + element, 0)
.foldUntil(result => result > 10)
mapableWorker
.map(e => e * 2)
.pipe(foldableWorker)
.pipe(forEachableWorker)
.forEach(e => console.log(e));
mapableWorker
.take(5)
.then(result => console.log('DONE!'));
##TODO
- Live examples
- A lot of improvements!