form-data-encoder

Encode FormData content into the multipart/form-data format

Installation

You can install this package using npm:

npm install form-data-encoder

Or yarn:

yarn add form-data-encoder

Or pnpm:

pnpm add form-data-encoder

ESM/CJS support

This package is targeting ESM and CJS for backwards compatibility reasons and smoothen transition period while you convert your projects to ESM only. Note that CJS support will be removed as Node.js v12 will reach its EOL. This change will be released as major version update, so you won't miss it.

Usage

To start the encoding process, you need to create a new Encoder instance with the FormData you want to encode:

import {Readable} from "stream"

import {FormData, File} from "formdata-node"
import {FormDataEncoder} from "form-data-encoder"

import fetch from "node-fetch"

const form = new FormData()

form.set("greeting", "Hello, World!")
form.set("file", new File(["On Soviet Moon landscape see binoculars through YOU"], "file.txt"))

const encoder = new FormDataEncoder(form)

const options = {
  method: "post",

  // Set request headers provided by the Encoder.
  // The `headers` property has `Content-Type` and `Content-Length` headers.
  headers: encoder.headers,

  // Create a Readable stream from the Encoder.
  // You can omit usage of `Readable.from` for HTTP clients whose support async iterables in request body.
  // The Encoder will yield FormData content portions encoded into the multipart/form-data format as node-fetch consumes the stream.
  body: Readable.from(encoder.encode()) // or just Readable.from(encoder)
}

const response = await fetch("https://httpbin.org/post", options)

console.log(await response.json())

Encoder support different spec-compatible FormData implementations. Let's try it with formdata-polyfill:

import {Readable} from "stream"

import {FormDataEncoder} from "form-data-encoder"
import {FormData} from "formdata-polyfill/esm-min.js"
import {File} from "fetch-blob" // v3

const form = new FormData()

form.set("field", "Some value")
form.set("file", new File(["File content goes here"], "file.txt"))

const encoder = new FormDataEncoder(form)

const options = {
  method: "post",
  headers: encoder.headers,
  body: Readable.from(encoder)
}

await fetch("https://httpbin.org/post", options)

Because the Encoder is iterable (it has both Symbol.asyncIterator and Symbol.iterator methods), you can use it with different targets. Let's say you want to convert FormData content into Blob, for that you can write a function like this:

import {Readable} from "stream"

import {FormDataEncoder} from "form-data-encoder"

import {FormData, File, Blob, fileFromPath} from "formdata-node"

import fetch from "node-fetch"

const form = new FormData()

form.set("field", "Just a random string")
form.set("file", new File(["Using files is class amazing"], "file.txt"))
form.set("fileFromPath", await fileFromPath("path/to/a/file.txt"))

// Note 1: When using with native Blob or fetch-blob@2 you might also need to generate boundary string for your FormDataEncoder instance
// because Blob will lowercase value of the `type` option and default boundary generator produces a string with both lower and upper cased alphabetical characters. Math.random() should be enough to fix this:
// const encoder = new FormDataEncoder(form, String(Math.random()))
const encoder = new FormDataEncoder(form)

const options = {
  method: "post",

  // Note 2: To use this approach with fetch-blob@2 you probably gonna need to convert the encoder parts output to an array first:
  // new Blob([...encoder], {type: encoder.contentType})
  body: new Blob(encoder, {type: encoder.contentType})
}

const response = await fetch("https://httpbin.org/post", options)

console.log(await response.json())

Here's FormData to Blob conversion with async-iterator approach:

import {FormData} from "formdata-polyfill/esm-min.js"
import {blobFrom} from "fetch-blob/from.js"
import {FormDataEncoder} from "form-data-encoder"

import Blob from "fetch-blob"
import fetch from "node-fetch"

// This approach may require much more RAM compared to the previous one, but it works too.
async function toBlob(form) {
  const encoder = new Encoder(form)
  const chunks = []

  for await (const chunk of encoder) {
    chunks.push(chunk)
  }

  return new Blob(chunks, {type: encoder.contentType})
}

const form = new FormData()

form.set("name", "John Doe")
form.set("avatar", await blobFrom("path/to/an/avatar.png"), "avatar.png")

const options = {
  method: "post",
  body: await toBlob(form)
}

await fetch("https://httpbin.org/post", options)

Another way to convert FormData parts to blob using form-data-encoder is making a Blob-ish class:

import {Readable} from "stream"

import {FormDataEncoder} from "form-data-encoder"
import {FormData} from "formdata-polyfill/esm-min.js"
import {blobFrom} from "fetch-blob/from.js"

import Blob from "fetch-blob"
import fetch from "node-fetch"

class BlobDataItem {
  constructor(encoder) {
    this.#encoder = encoder
    this.#size = encoder.headers["Content-Length"]
    this.#type = encoder.headers["Content-Type"]
  }

  get type() {
    return this.#type
  }

  get size() {
    return this.#size
  }

  stream() {
    return Readable.from(this.#encoder)
  }

  get [Symbol.toStringTag]() {
    return "Blob"
  }
}

const form = new FormData()

form.set("name", "John Doe")
form.set("avatar", await blobFrom("path/to/an/avatar.png"), "avatar.png")

const encoder = new FormDataEncoder(form)

// Note that node-fetch@2 performs more strictness tests for Blob objects, so you may need to do extra steps before you set up request body (like, maybe you'll need to instaniate a Blob with BlobDataItem as one of its blobPart)
const blob = new BlobDataItem(enocoder) // or new Blob([new BlobDataItem(enocoder)], {type: encoder.contentType})

const options = {
  method: "post",
  body: blob
}

await fetch("https://httpbin.org/post", options)

In this example we will pull FormData content into the ReadableStream:

 // This module is only necessary when you targeting Node.js or need web streams that implement Symbol.asyncIterator
import {ReadableStream} from "web-streams-polyfill/ponyfill/es2018"

import {FormDataEncoder} from "form-data-encoder"
import {FormData} from "formdata-node"

import fetch from "node-fetch"

function toReadableStream(encoder) {
  const iterator = encoder.encode()

  return new ReadableStream({
    async pull(controller) {
      const {value, done} = await iterator.next()

      if (done) {
        return controller.close()
      }

      controller.enqueue(value)
    }
  })
}

const form = new FormData()

form.set("field", "My hovercraft is full of eels")

const encoder = new FormDataEncoder(form)

const options = {
  method: "post",
  headers: encoder.headers,
  body: toReadableStream(encoder)
}

// Note that this example requires `fetch` to support Symbol.asyncIterator, which node-fetch lacks of (but will support eventually)
await fetch("https://httpbin.org/post", options)

Speaking of async iterables - if HTTP client supports them, you can use encoder like this:

import {FormDataEncoder} from "form-data-encoder"
import {FormData} from "formdata-node"

import fetch from "node-fetch"

const form = new FormData()

form.set("field", "My hovercraft is full of eels")

const encoder = new FormDataEncoder(form)

const options = {
  method: "post",
  headers: encoder.headers,
  body: encoder
}

await fetch("https://httpbin.org/post", options)

...And for those client whose supporting form-data-encoder out of the box, the usage will be much, much more simpler:

import {FormData} from "formdata-node" // Or any other spec-compatible implementation

import fetch from "node-fetch"

const form = new FormData()

form.set("field", "My hovercraft is full of eels")

const options = {
  method: "post",
  body: form
}

// Note that node-fetch does NOT support form-data-encoder
await fetch("https://httpbin.org/post", options)

API

`class FormDataEncoder`

`constructor(form[, boundary, options]) -> {Encoder}`

{FormDataLike} form - FormData object to encode. This object must be a spec-compatible FormData implementation.
{string} [boundary] - An optional boundary string that will be used by the encoder. If there's no boundary string is present, Encoder will generate it automatically.
{object} [options] - Encoder options.
{boolean} [options.enableAdditionalHeaders = false] - When enabled, the encoder will emit additional per part headers, such as Content-Length. Please note that the web clients do not include these, so when enabled this option might cause an error if multipart/form-data does not consider additional headers.

Creates a multipart/form-data encoder.

Instance properties

`boundary -> {string}`

Returns boundary string.

`contentType -> {string}`

Returns Content-Type header.

`contentLength -> {string}`

Return Content-Length header.

`headers -> {object}`

Returns headers object with Content-Type and Content-Length header.

Instance methods

`values() -> {Generator<Uint8Array | FileLike, void, undefined>}`

Creates an iterator allowing to go through form-data parts (with metadata). This method will not read the files.

`encode() -> {AsyncGenerator<Uint8Array, void, undefined>}`

Creates an async iterator allowing to perform the encoding by portions. This method will also read files.

`[Symbol.iterator]() -> {Generator<Uint8Array | FileLike, void, undefined>}`

An alias for Encoder#values() method.

`[Symbol.asyncIterator]() -> {AsyncGenerator<Uint8Array, void, undefined>}`

An alias for Encoder#encode() method.

`isFileLike(value) -> {boolean}`

Check if a value is File-ish object.

{unknown} value - a value to test

`isFormDataLike(value) -> {boolean}`

Check if a value is FormData-ish object.

{unknown} value - a value to test