@teamawesome/tiny-batch

Installation

npm install @teamawesome/tiny-batch

tiny-batch

tiny-batch is a utility to create functions of which the execution is batched. This can be very useful for instance to limit the amount of queries or http requests while still having a single, easy to use function.

Usage

Call tinybatch to create an async function that adds to the batch. The first argument is a callback that will handle the batching.

import tinybatch from "@teamawesome/tiny-batch";

const batchedFunc = tinybatch((batchedArgs) => {
  // code
});

For example, fetch users from different components with a single request:

import tinybatch from "@teamawesome/tiny-batch";

const getUserById = tinybatch((batchedArgs: [number][]): User[] => {
  // batchedArgs equals [[1], [2]]
  const userIds = batchedArgs.flat();

  return fetch(`api/${userIds}`)
    .then((response) => response.json())
    .then((json) => json.users);
});

const user1 = await getUserById(1);
const user2 = await getUserById(2);

Callback

Each call of the batched function adds its arguments to the queue as-is. The callback then gets an array of all these arguments. The callback must return an array or a promise of an array. The return value will be used to resolve the batched function calls in the same order. If an entry is instanceof Error, the call will be rejected.

import tinybatch from '@teamawesome/tiny-batch';

const batchedFunc = tinybatch((batchedArgs: unknown[][]): string[] => {
    // batchedArgs equals
    // [
    //  [1, 2, 3],
    //  ["a", "b", "c"]
    // ]

    return batchedArgs.map((_, index) => `${index} done!`);
});

await first = batchedFunc(1, 2, 3); // 0 done!
await second = batchedFunc("a", "b", "c"); // 1 done!

Scheduler

tinybatch has a second argument to specify a scheduler. A scheduler determines when to execute the callback. The scheduler is called each time an entry is added to the batch. tinybatch comes with some scheduler factories out of the box:

name	description
microtaskScheduler()	(default) Queues a flush in the microtask queue at the first call.
intervalScheduler(ms)	Flushes every given ms, regardless of the queue. The timer can be cleared with the stop() method.
timeoutScheduler(ms)	Waits the given amount of ms after the first call to flush. The timer can be cleared with the stop() method.
amountScheduler(amount)	Flushes after the given amount of calls.

import { tinybatch, amountScheduler } from "@teamawesome/tiny-batch";

// Get users in batches of 10.
const getUserById = tinybatch((batchedArgs) => {
  // code
}, amountScheduler(10));

Batched Function

The queue can be manually flushed. This will execute the callback regardless of the scheduler. Note that the callback is never called if the queue is empty.

batchedFunc.flush();

The queue can also be inspected.

console.log(batchedFunc.queue);

The scheduler of a tinybatch is available. Some schedulers have extra methods, for instance to clear timers.

console.log(batchedFunc.scheduler);

Caching

To reduce overhead even more, caching can be introduced. While this is not supported directly by tiny-batch, it is very simple to achieve. Use any of the memoization libraries available. For example, memoizee;

import memoizee from "memoizee";

const batched = tinybatch((args) => {
  // code
});
const batchedAndCached = memoizee(batched, {
  // Set the amount of arguments that "batchedAndCached" will receive.
  length: 1,
});

await batchedAndCached("once");
await batchedAndCached("once");

The second call is not added to the queue but will resolve with the same value.