GraphQL Server File Upload Example

This example illustrates the implementation of File API with GraphQL Server pattern. The best example for GraphQL Server basic usage can be found here: https://github.com/graphcool/graphql-server-example .

Getting Started

Initializing Prisma Database Service

graphcool deploy # choose local cluster
# copy API endpoint into the `PRISMA_ENPOINT` env var in .env

To get GRAPHCOOL_SECRET visit http://jwtbuilder.jamiekurtz.com and scroll to the bottom where you can hash your secret from graphcool.yml and get the hashed output. (sssh is used in the example.)

Setting up the S3 bucket

Head over to the AWS console and navigate to the S3 section.
Click create bucket and follow the instructions on screen.
Once you have created a bucket, add bucket name that you've picked to .env S3_BUCKET property.
Head back to the AWS Console and open Identity and Access Management (IAM) console. Navigate to Users and click Add user.
Under Access type check Programmatic access and press Next. From options, select Attach existing policies directly and a table below will open. Search for AmazonS3FullAccess and check it. Press Next to review everything and submit by pressing Create user.
Once done, copy the Access key ID to .env S3_KEY property and Secret access key to .env S3_SECRET property.
You are all set to start the server!

Starting the Server

yarn install
yarn start
# Open http://localhost:5000/

Uploading files

You can upload files to a project by doing a multipart/form-data HTTP request to the File API http://localhost:5000/upload.

It's important to use the form parameter data as seen in the example below.

Uploading workflow

Everytime you upload a file to Prisma, a new File node is created that contains information about that file.

id: the familiar system field
secret: a unique, unguessable secret that allows access to the file
name: the file name
size: the file size
url: the url of the file where it can be accessed. The url contains of the project id and the file secret, so is unguessable as well.
contentType: the contentType of the file. It is determined based on the file name (extension in the name is required!).

If you want to connect the File node to another node in a relation, you can use the id in the response.

With curl you could execute:

curl -X POST 'http://localhost:5000/upload' -F "data=@example.png; filename=coolimage.png"

This uploads the local file example.png under coolimage.png name. The response could look something like this:

[{
  "id": "cjbqvp4ii00390181b1q0dq6h",
  "name": "coolimage.png",
  "secret": "43de4b08-78b2-4b5c-a5b7-05ee350ee09a",
  "contentType": "image/png",
  "size": 36625,
  "url": "https://__S3_BUCKET__.s3-eu-west-1.amazonaws.com/43de4b08-78b2-4b5c-a5b7-05ee350ee09a"
}]

If there's no filename provided, the original name of the file is used instead.

License

MIT

maticzav/graphql-server-file-upload-example