formzilla
v3.4.2
Published
Fastify plugin for parsing multipart/form data
Downloads
2,377
Maintainers
Readme
Formzilla
Formzilla is a Fastify plugin to handle multipart/form-data
content.
Why?
Even though other plugins for the same purpose exist, like @fastify/multipart and fastify-multer, when dealing with mixed content, they don't play well with JSON schemas which are Fastify's built-in mechanism for request validation and documentation. Formzilla is intended to work seamlessly with JSON schemas and @fastify-swagger.
Example
Let's say you have an endpoint that accepts multipart/form-data
with the following schema.
const postCreateSchema = {
consumes: ["multipart/form-data"],
body: {
type: "object",
properties: {
content: {
type: "string"
},
media: {
type: "string",
format: "binary"
},
poll: {
type: "object",
properties: {
first: { type: "string" },
second: { type: "string" }
},
required: ["first", "second"]
}
}
}
};
You will find that neither @fastify/multipart
nor fastify-multer
will process this schema correctly, unless you add a preValidation
hook to convert your request body into the correct schema. I created Formzilla to solve this exact problem.
import fastify, { FastifyInstance } from "fastify";
import formDataParser from "formzilla";
import fastifySwagger from "@fastify/swagger";
import fastifySwaggerUi from "@fastify/swagger-ui";
const postCreateSchema = {
consumes: ["multipart/form-data"],
body: {
type: "object",
properties: {
content: {
type: "string"
},
media: {
type: "string",
format: "binary"
},
poll: {
type: "object",
properties: {
first: { type: "string" },
second: { type: "string" }
},
required: ["first", "second"]
}
}
}
};
const server: FastifyInstance = fastify({ logger: true });
server.register(formDataParser);
server.register(fastifySwagger, {
mode: "dynamic",
openapi: {
info: {
title: "Formzilla Demo",
version: "1.0.0"
}
}
});
server.register(fastifySwaggerUi, {
routePrefix: "/swagger"
});
server.register(async (instance, options) => {
instance.post("/create-post", {
schema: postCreateSchema,
handler: (request, reply) => {
console.log(request.body);
/*
request.body will look like this:
{
content: "Test.",
poll: { first: "Option 1", second: "Option 2" },
media: {
originalName: "flame-wolf.png",
encoding: "7bit",
mimeType: "image/png",
path?: <string>, // Only when using DiscStorage
stream?: <Readable> // Only when using StreamStorage
data?: <Buffer> // Only when using BufferStorage
error?: <Error> // Only if any errors occur during processing
}
}
*/
reply.status(200).send();
}
});
});
server.listen(
{
port: +(process.env.PORT as string) || 1024,
host: process.env.HOST || "::"
},
(err, address) => {
if (err) {
console.log(err.message);
process.exit(1);
}
console.log(`Listening on ${address}`);
}
);
Installation
npm install formzilla
Important
I guess this goes without saying, but you must register the plugin before registering your application routes.
API
Breaking changes from version 2
- Formzilla 2.x will not work with Fastify versions 4.8 and above. Use Formzilla 3.x with Fastify versions >= 4.8.
Breaking changes from version 1
- Formzilla 1.x
options
have been moved tooptions.limits
in Formzilla 2.x. - File content is stored by default in
file.stream
as aReadable
in Formzilla 2.x whereas in Formzilla 1.x it was stored infile.data
as aBuffer
.
Options
These are the valid keys for the options
object parameter accepted by Formzilla:
limits
: Same as thelimits
configuration option for busboy.const formLimits = { fieldNameSize?: number, // Max field name size (in bytes). Default: 100. fieldSize?: number, // Max field value size (in bytes). Default: 1048576 (1MB). fields?: number, // Max number of non-file fields. Default: Infinity. fileSize?: number, // For multipart forms, the max file size (in bytes). Default: Infinity. files?: number, // For multipart forms, the max number of file fields. Default: Infinity. parts?: number, // For multipart forms, the max number of parts (fields + files). Default: Infinity. headerPairs?: number // For multipart forms, the max number of header key-value pairs to parse. Default: 2000 (same as node's http module). }; server.register(formDataParser, { limits: formLimits });
storage
: Where to store the files, if any, included in the request. Formzilla provides the following built-in options. It is possible to write custom storage plugins of your own.StreamStorage
: The default storage option used by Formzilla. Stores file contents as aReadable
in thestream
property of the file. Example:server.register(formDataParser, { storage: new StreamStorage() });
BufferStorage
: Emulates Formzilla 1.x behaviour by storing file contents as aBuffer
in thedata
property of the file. Example:server.register(formDataParser, { storage: new BufferStorage() });
DiscStorage
: Saves the file to the disc. Accepts a parameter that can be either aformzilla.FileSaveTarget
or a function that accepts aformzilla.File
parameter and returns aformzilla.FileSaveTarget
. By default, Formzilla will save the file to the operating system's TEMP directory. Example:server.register(formDataParser, { storage: new DiscStorage(file => { return { directory: path.join(__dirname, "public"), fileName: file.originalName.toUpperCase() }; }) });
CallbackStorage
: For advanced users. Accepts a callback function that takes three parameters: astring
, aReadable
, and abusboy.FileInfo
. The callback function must consume theReadable
and return either aformzilla.File
or a promise that resolves to aformzilla.File
. Example:// The following example uploads the incoming stream // directly to a cloud server. The call to `resolve` is // nested inside the cloud API's callback function to ensure // that the `path` property of the `FileInternal` object // is populated correctly. server.register(formDataParser, { storage: new CallbackStorage((name, stream, info) => { return new Promise(resolve => { const file = new FileInternal(name, info); var uploader = cloudinary.v2.uploader.upload_stream((err, res) => { file.error = err; file.path = res?.secure_url; resolve(file); }); stream.pipe(uploader); }); }) });
Recommendations
Both StreamStorage
and BufferStorage
will cause files to accumulate in memory and hence make your endpoint a potential target for DDoS attacks. CallbackStorage
must cosume the stream inside the callback or it will break the application. It is recommended only if you are familiar with streams in NodeJS and want to manipulate the stream in some way before sending it to the response body. It's recommended to use DiscStorage
to temporarily store an incoming file, upload it to a cloud server like Cloudinary from your request handler, and then delete the temporary file.
Caveats
- File data will not be available in
request.body
until thepreHandler
request lifecycle stage. So if you want to access the files inside apreValidation
hook, userequest.__files__
instead. This is a temporary property that gets removed from the request object at thepreHandler
stage. It is done this way for security purposes.