Skip to content

Borewit/bounded-queue

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Node.js CI NPM version npm downloads

bounded-queue

bounded-queue helps solves the producer–consumer problem.

graph LR;
    P(Producer);
    B[bounded-queue];
    C(Consumer);
    P-- batched item -->B;
    B-- batched item -->C;
    style B fill:#99E,stroke:#333

The bounded-queue allows the producer and consumer to operate in

Introduction

Imagine you have to read records from a database and write those to another database. A simple way to that move the records is to first read from database A and sequentially write each record to database B.

let batchNr = 0;
let items2produce = 10;

/**
 * Mockup database A, producer
 */
const dbA = {
    readRecord: async () => {
        ++batchNr;
        console.log("Producing batch #", batchNr);
        await new Promise(resolve => setTimeout(resolve, 100));
        console.log("Produced  batch #", batchNr);
        return batchNr;
    },
    moreRecordsAvailable: () => batchNr < items2produce
}

/**
 * Mockup database B, consumer
 */
const dbB = {
    async writeRecord(batchNr) {
        console.log("Consuming batch #", batchNr);
        await new Promise(resolve => setTimeout(resolve, 100));
        console.log("Consumed  batch #", batchNr);
    }
}

/**
 * Sequential conversion
 */
async function convertDatabaseRecords() {

    while(dbA.moreRecordsAvailable()) {
        const record = await dbA.readRecord();
        // Consumer
        await dbB.writeRecord(record); // expensive async write (consume) operation
    }
}

(async () => {
    console.time("no-queue");
    await convertDatabaseRecords();
    console.timeEnd("no-queue");
})();

In the previous example, we either read from database A, or write to database B. It would be faster if read from database A, while we write to database B, at the same time. As dbA.readRecord() and dbB.readRecord()areasync` functions, there is no need to introduce threading to accomplish that.

The bounded-queue helps you with that. The following example uses bounded-queue, with a maximum of 3 queued records:

import {queue} from 'bounded-queue';

let batchNr = 0;
let items2produce = 10;

/**
 * Mockup database A, producer
 */
const dbA = {
    readRecord: async () => {
        ++batchNr;
        console.log("Producing batch #", batchNr);
        await new Promise(resolve => setTimeout(resolve, 100));
        console.log("Produced  batch #", batchNr);
        return batchNr;
    },
    moreRecordsAvailable: () => batchNr < items2produce
}

/**
 * Mockup database B, consumer
 */
const dbB = {
    async writeRecord(batchNr) {
        console.log("Consuming batch #", batchNr);
        await new Promise(resolve => setTimeout(resolve, 100));
        console.log("Consumed  batch #", batchNr);
    }
}

/**
 * Conversion using bounded-queue
 */
async function convertDatabaseRecords() {

    await queue(3, () => {
        // Producer
        return dbA.moreRecordsAvailable() ? dbA.readRecord() : null; // expenive async read (produce) operation
    }, record => {
        // Consumer
        return dbB.writeRecord(record); // expensive async write (consume) operation
    });
}

(async () => {
    console.time("bounded-queue");
    await convertDatabaseRecords();
    console.timeEnd("bounded-queue");
})();

Using the bounded-queue, the conversion will complete in roughly half the time.

Installation

npm install bounded-queue

API Documentation

BoundedQueue class

Constructor

constructor(maxQueueSize: number, producer: Producer<ItemType>, consumer: Consumer<ItemType>)

  • Parameters:
    • maxQueueSize (number): Maximum number of items that can be in the queue.
    • producer (Producer): A function that produces items to be added to the queue.
    • consumer (Consumer): A function that consumes items from the queue.
  • Description: Initializes a new BoundedQueue instance with the specified maxQueueSize, producer, and consumer.

Methods

length(): number

  • Description: Returns the number of items currently in the queue.
  • Returns: The number of items in the queue.

run(): Promise<void>

  • Description: Initiates the asynchronous processing of items in the queue. It starts filling the queue with items produced by the producer function and concurrently consumes items using the consumer function.
  • Returns: A Promise that resolves when all items have been produced and consumed.

Example Usage

import { queue } from 'your-module';

// Create and run a BoundedQueue with a maximum queue size of 10
queue(10, producer, consumer)
  .then(() => {
    console.log('Queue processing completed');
  })
  .catch((error) => {
    console.error('Error during queue processing:', error);
  });

Classless usage

queue() Function

queue<ItemType>(maxQueueSize: number, producer: Producer<ItemType>, consumer: Consumer<ItemType>): Promise<void>

  • Parameters:
    • maxQueueSize (number): Maximum number of items that can be in the queue.
    • producer (Producer): A function that produces items to be added to the queue.
    • consumer (Consumer): A function that consumes items from the queue.
  • Description: A convenience function for creating and running a BoundedQueue instance. It takes the same parameters as the BoundedQueue constructor and returns a Promise that resolves when all items have been produced and consumed.
  • Returns: A Promise that resolves when all items have been produced and consumed. Resolving null indicated the end of the production.

Example Usage

import { queue } from 'your-module';

// Create and run a BoundedQueue with a maximum queue size of 10
queue(10, producer, consumer)
  .then(() => {
    console.log('Queue processing completed');
  })
  .catch((error) => {
    console.error('Error during queue processing:', error);
  });

About

Bounded batch queue, where items are produced and consumed based on user specified functions

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Sponsor this project

 

Packages

No packages published