Minimalistic and developer friendly middleware and an Upload scalar to add support for GraphQL multipart requests (file uploads via queries and mutations) to various Node.js GraphQL servers.
This module was graphql-upload. The original module is exceptionally well documented and well written. It was very easy to fork and amend. Thanks Jayden!
I needed something simpler which won't attempt doing any disk I/O. There were no server-side JavaScript alternative modules for GraphQL file uploads. Thus, this fork was born.
Differences to graphql-upload
- Results in 9 less production dependencies.
- And 6 less MB in your
node_modules. - And using a bit less memory.
- And a bit faster.
- Most importantly, less risk that one of the dependencies would break your server.
Using ASCII-only text. Direct developers to resolve common mistakes.
- Thus works faster.
- Does not have a risk of clogging your file system. Even on high load.
- No need to manually destroy the programmatically aborted streams.
Does not follow strict specification
You can't have same file referenced twice in a GraphQL query/mutation.
- Does not accept any arguments to
createReadStream(). Will throw if any provided. - Calling
createReadStream()more than once per file is not allowed. Will throw.
Otherwise, this module is a drop-in replacement for the graphql-upload.
The following environments are known to be compatible:
- Node.js versions 12, 14, 16, and 18. It works in Node 10 even though the unit tests fail.
- AWS Lambda. Reported to be working.
- Google Cloud Functions (GCF) Experimental. Untested.
- Azure Functions Working.
- Koa
- Express.js
See also GraphQL multipart request spec server implementations.
To install graphql-upload-minimal and the graphql peer dependency from npm run:
npm install graphql-upload-minimal graphqlUse the graphqlUploadKoa or graphqlUploadExpress middleware just before GraphQL middleware. Alternatively, use processRequest to create a custom middleware.
A schema built with separate SDL and resolvers (e.g. using makeExecutableSchema) requires the Upload scalar to be setup.
Clients implementing the GraphQL multipart request spec upload files as Upload scalar query or mutation variables. Their resolver values are promises that resolve file upload details for processing and storage. Files are typically streamed into cloud storage but may also be stored in the filesystem.
Minimalistic code example showing how to upload a file along with arbitrary GraphQL data and save it to an S3 bucket.
Express.js middleware. You must put it before the main GraphQL sever middleware. Also, make sure there is no other Express.js middleware which parses multipart/form-data HTTP requests before the graphqlUploadExpress middleware!
const express = require("express");
const expressGraphql = require("express-graphql");
const { graphqlUploadExpress } = require("graphql-upload-minimal");
express()
.use(
"/graphql",
graphqlUploadExpress({ maxFileSize: 10000000, maxFiles: 10 }),
expressGraphql({ schema: require("./my-schema") })
)
.listen(3000);GraphQL schema:
scalar Upload
input DocumentUploadInput {
docType: String!
file: Upload!
}
type SuccessResult {
success: Boolean!
message: String
}
type Mutations {
uploadDocuments(docs: [DocumentUploadInput!]!): SuccessResult
}GraphQL resolvers:
const { S3 } = require("aws-sdk");
const resolvers = {
Upload: require("graphql-upload-minimal").GraphQLUpload,
Mutations: {
async uploadDocuments(root, { docs }, ctx) {
try {
const s3 = new S3({ apiVersion: "2006-03-01", params: { Bucket: "my-bucket" } });
for (const doc of docs) {
const { createReadStream, filename /*, fieldName, mimetype, encoding */ } = await doc.file;
const Key = `${ctx.user.id}/${doc.docType}-${filename}`;
await s3.upload({ Key, Body: createReadStream() }).promise();
}
return { success: true };
} catch (error) {
console.log("File upload failed", error);
return { success: false, message: error.message };
}
},
},
};See the example Koa server and client.
Reported to be working.
const { processRequest } = require("graphql-upload-minimal");
module.exports.processRequest = function (event) {
return processRequest(event, null, { environment: "lambda" });
};Possible example. Experimental. Untested.
const { processRequest } = require("graphql-upload-minimal");
exports.uploadFile = function (req, res) {
return processRequest(req, res, { environment: "gcf" });
};Possible example. Working.
const { processRequest } = require("graphql-upload-minimal");
exports.uploadFile = function (context, req) {
return processRequest(context, req, { environment: "azure" });
};When uploading multiple files you can make use of the fieldName property to keep track of an identifier of the uploaded files. The fieldName is equal to the passed name property of the file in the multipart/form-data request. This can be modified to contain an identifier (like a UUID), for example using the formDataAppendFile in the commonly used apollo-upload-link library.
- Only use
createReadStream()before the resolver returns; late calls (e.g. in an unawaited async function or callback) throw an error.
The GraphQL multipart request spec allows a file to be used for multiple query or mutation variables (file deduplication), and for variables to be used in multiple places. GraphQL resolvers need to be able to manage independent file streams.
busboy parses multipart request streams. Once the operations and map fields have been parsed, Upload scalar values in the GraphQL operations are populated with promises, and the operations are passed down the middleware chain to GraphQL resolvers.