async-batch

Asynchronously process task batches.

Usage

async-batch shines when you need to process several potentially taxing tasks in terms of computing resources. Some sample use cases:

• Automated calls to a rate-limited API.
• I/O heavy operations.
• Throttling calls to a service running on devices with a small amount of computing resources.

The package exports a single function, asyncBatch, which expects three parameters:

• A list of tasks. Tasks can be anything (numbers, objects, lists, etc.)
• A handler function to run the tasks.
• The number of concurrent workers to run tasks on.

The following code snippet showcases how to use it:

import asyncBatch from "async-batch";

async function squares() {
  const input = [10, 2, 3, 8, 1, 7, 4];
  const processingOrder: number[] = [];

  const result = await asyncBatch(
    input,
    (task: number, taskIndex: number, workerIndex: number) => new Promise(
      (resolve) => setTimeout(
        () => processingOrder.push(task) && resolve(task * task),
        task * 25,
      ),
    ),
    2,
  );

  console.log(processingOrder);   // [  2,  3, 10,  1,  8,  4,  7];
  console.log(result);            // [100,  4,  9, 64,  1, 49, 16];

  return result;
}

squares();

asyncBatch returns a Promise that resolves with the results of all tasks once all of them have been processed.

Tasks in the input list are grabbed in order, however, since each task can take different times to complete, the completion order is not guaranteed to be ordered in any way. Despite of this, the results list returned once asyncBatch resolves is guaranteed to preserve the order of the elements in a way that the result of processing the task at index X will always be found at index X in the results list.