hardened-fetch
v1.2.3
Published
Hardened Fetch is a tiny wrapper for `global.fetch` adding request timeouts, request throttling, retries with backoff and rate limit support, pagination, and descriptive errors.
Downloads
1,412
Maintainers
Readme
hardened-fetch
Hardened Fetch is a tiny wrapper for global.fetch
adding request timeouts, request throttling, retries with backoff and rate limit support, pagination, and descriptive errors. It makes working with APIs without SDKs and web scraping easier.
import { HardenedFetch } from 'hardened-fetch'
const client = new HardenedFetch({
baseUrl: 'https://swapi.dev/api/',
})
const response = await client.fetch('/species/1/')
Installation
This is a Node.js module available through the npm registry. Node.js 18 or higher is required.
$ npm install hardened-fetch
Features
- Request timeouts
- Throttle concurrent requests and request rate
- Retries failed requests with exponential back off
- Retries rate-limited requests according to reset time
- Easily step-through paginated API responses
API
new HardenedFetch([options] = {})
Creates a new Hardened Fetch client.
Constructor Options:
| Name | Type | Description |
| ----------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------ |
| baseUrl
| string
| A base URL to prepend to each request, optional. |
| defaultHeaders
| HeadersInit
| Default headers to add to each request, optional. |
| maxConcurrency
| number
| How many requests can be running at the same time. |
| minRequestTime
| number
| How long to wait after launching a request before launching another one. |
| maxRetries
| number
| Number of retry attempts for failed requests. |
| doNotRetry
| number[]
| List of HTTP status codes that will not trigger a retry attempt. |
| rateLimitHeader
| string
| The name of the rate limit reset header, usually one of "Retry-After"
, "X-RateLimit-Reset"
, or"X-Rate-Limit-Reset"
|
| resetFormat
| string
| The format of the rate limit reset header, must be one of "datetime"
, "seconds"
or "milliseconds"
. |
All of the options and their defaults are shown below:
const client = new HardenedFetch({
// Request options
baseUrl: undefined,
defaultHeaders: undefined,
// Throttle options
maxConcurrency: 10,
minRequestTime: 0,
// Retry options
maxRetries: 3,
doNotRetry: [400, 401, 403, 404, 422, 451],
// Rate limit options
rateLimitHeader: 'Retry-After',
resetFormat: 'seconds',
})
client.fetch(url, [init] = {}, [timeout] = 30000)
Expects a url
to the resource that you wish to fetch and optionally custom settings to apply to the request, and a timeout in milliseconds. Returns a promise which will resolve with the Response
object when successful. Rejects with a relevant HTTP error on failure.
const response = await client.fetch('https://swapi.dev/api/species/1/')
const json = await response.json()
client.paginatedFetch(url, [options] = {}, [timeout] = 30000)
Expects a url
to the resource that you wish to fetch and optionally custom settings to apply to the request, and a timeout in milliseconds. Returns an AsyncIterator
which will resolve with a Response
object and done
property on each successful iteration. Rejects with a relevant HTTP error on failure.
const pages = client.paginatedFetch('https://swapi.dev/api/species')
for await (const { response, done } of pages) {
const json = await response.json()
if (done) {
console.log('Done!')
}
}
License
This package is MIT licensed.