Skip to content

teidesu/eager-async-pool

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

eager-async-pool

Simple to use, fully asynchronous and iterable-based async pool for JavaScript and TypeScript. Works with native Promises.

Features

  • Zero dependencies
  • Uses native Promise and iterators
  • Supports older environments without native iterators
  • Compatible down to ES3
  • Immediate action after processing an item
  • Cancel tokens

Quick start

npm install eager-async-pool or yarn add eager-async-pool

const { asyncPool } = require('eager-async-pool')

// exceptionally useful executor
const executor = (n) => 'Number is: ' + n
// absolutely realistic dataset
const items = [1, 2, 3, 4]

for await (const { value, error } of asyncPool(executor, items, { limit: 2 })) {
    if (error) {
        console.error(error)
    } else {
        console.log(value)
    }
}

Usage

Basic

First, write an executor function, which will do some kind of task:

const executor = (url) => fetch(url).then(i => i.text())

Then, create an array (or any iterable) that contains the data that will be passed to the executor one-by-one:

const urls = ['https://google.com', 'https://yandex.ru', 'https://bing.com']

Finally, call the asyncPool and handle the results!

for await (const { idx, item, value, error } of asyncPool(executor, urls)) {
    console.log(`${item} title: ${cheerio.load(value).find('title').text()}`)
}

Alternatively, you can use asyncPoolCallback that wraps over asyncPool and allows you to specify a callback rather than using for-await loop:

await asyncPoolCallback(executor, urls, ({ idx, item, value, error }) => {
    console.log(`${item} title: ${cheerio.load(value).find('title').text()}`)
})

Handling errors

Of course, while processing items you may very well encounter some kind of error. Luckily, it is very easy to handle them with eager-async-pool!

When executor throws an error, iterator will yield an object containing .error, which will contain that very error. To retry the failed item, just push it to the array!

for await (const { idx, item, value, error } of asyncPool(executor, urls)) {
    if (error) {
        urls.push(item)
        continue
    }
    // ...rest of the logic...
}

Cancelling

eager-async-pool also supports cancel tokens. They are extremely easy to use as well. Just use CancelToken.source(), pass the token to asyncPool and .cancel() when you feel like:

const cancel = CancelToken.source()
for await (const { idx, item, value, error } of asyncPool(executor, urls, { cancel: cancel.token })) {
    if (error) {
        cancel.cancel('An error has occurred')
        break
    }
    console.log(`${item} title: ${cheerio.load(value).find('title').text()}`)
}

Note: break in an iterable asyncPool will also effectively cancel the pool, but break is (for obvious reasons) not available in asyncPoolCallback.

Also, cancel token is compatible with other libraries that support it, so you can use it inside executors as well.

Note: cancelling an async pool only guarantees that executor will not be called anymore. It DOES NOT cancel pending operations, nor does it ignore them.
If you want to cancel pending operations as well, you must handle cancel token inside executor manually (or provide it to some library).

Modern mode

Default exported entrypoint is compatible down to ES3. If your environment supports ES6, you can instead use require('eager-async-pool/modern'), which exports the same API, but written using modern JavaScript.

API

Please see index.d.ts

Motivation

Consider the common task of batch file upload/download.
You could try to do that using built-in Promise.all(), but that's probably not the best idea, since it will probably slow down the process and use a lot of resources, and server might be limiting concurrent connections.

Also, when downloading you might encounter an error (e.g. rate limit), and you won't want that error to interrupt the entire download process, and instead you'd want to retry downloading that particular file.

You may also want to do something immediately after the file is downloaded, and not wait until all other files are downloaded as well.

About

Fully asynchronous iterable-based async pool for JS

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published