Nikolas Burk
Written By
Nikolas Burk
Developer @ Prisma

Nikolas is a developer and head of content at Prisma. He is excited about GraphQL as a new API technology and has a passion for learning and sharing knowledge.

Getting Started

Since this is a frontend track, you’re not going to spend any time implementing the backend. Instead, you’ll use the server from the Node tutorial.

Once you’ve created your React application, you’ll pull in the required code for the backend.

Note: The final project for this tutorial can be found on GitHub. You can always use it as a reference whenever you get lost throughout the course of the following chapters. Also note that each code block is annotated with a filename. These annotations directly link to the corresponding file on GitHub so you can clearly see where to put the code and what the end result will look like.

Need some help? No tutorial is perfect and questions always come up, so feel free to reach out to us and the community over on the urql Spectrum, if you have any questions or need some help!

Frontend

Creating the app

First, you are going to create the React project! As mentioned in the beginning, you’ll use create-react-app for that.

Note: This tutorial uses Yarn for dependency management. Find instructions for how you can install it here. If you prefer using npm, you can just run the equivalent commands.

This will create a new directory called hackernews-react-urql that contains a complete setup for a client-side only React app.

Make sure everything works by navigating into the directory and starting the app:

cd hackernews-react-urql
yarn start

This will open a browser and navigate to http://localhost:3000 where the app is running. If everything went well, you’ll see the following:

Navigating to localhost:3000

.../hackernews-react-urql/src/index.js
import React from 'react'
import ReactDOM from 'react-dom'
import './styles/index.css'
import App from './components/App'
.../hackernews-react-urql/src/components/App.js
import React from 'react'
import logo from '../logo.svg'
import '../styles/App.css'

Your project structure should now look as follows:

.
├── README.md
├── node_modules
├── package.json
├── public
│   ├── favicon.ico
│   ├── index.html
│   └── manifest.json
├── src
│   ├── App.test.js
│   ├── components
│   │   └── App.js
│   ├── index.js
│   ├── logo.svg
│   ├── reportWebVitals.js
│   └── styles
│       ├── App.css
│       └── index.css
└── yarn.lock

Prepare styling

This tutorial is about the concepts of GraphQL and how you can use it from within a React application, so we want to spend the least time possible on styling. To reduce the usage of CSS in this project, you’ll use the Tachyons library which provides a number of CSS classes.

Since we still want to have a bit more custom styling here and there, we also prepared some styles for you that you need to include in the project.

Install urql

Here’s an overview of the packages you just installed:

  • urql offers the basic urql client which includes React hooks and components, and a basic document cache by default
  • @urql/exchange-graphcache is a replacement for urql’s default cache, which supports full normalized caching, which we’ll set up later
  • graphql contains Facebook’s reference implementation of GraphQL - urql and its other packages use some of its functionality as well.
  • graphql-tag is a utility to write GraphQL query definitions using tagged template literals.

Note: You can find more information on how urql’s Exchanges work in its docs. Generally speaking, every GraphQL operation goes through a chain of middleware that can transform, filter, or fulfill them. Every core feature in urql including fetching, deduplication, and caching is implemented using Exchanges.

That’s it, you’re ready to write some code! 🚀

Configure urql’s Client

Instead of dealing with GraphQL requests directly, urql has a central Client. It controls when and how operations are made and deals with all the details of deduplication, caching, and cancellation. In contrast to working with REST APIs, you don’t have to construct any HTTP requests manually or store the results explicitly - instead you can simply write queries and mutations and send them using urql’s React bindings. Internally these bindings just use methods on the Client, for instance executeQuery and executeMutation.

The first thing you have to do when using urql is configure a Client instance. It needs to know the endpoint of your GraphQL API so it can deal with the network connections.

Note: The project that was generated by create-react-app uses semicolons and double quotes for strings. All the code that you’re going to add will use no semicolons and mostly single quotes. You’re also free to delete any existing semicolons and replace double with single quotes 🔥

Let’s try to understand what’s going on in that code snippet:

  • You’re importing the Client, Provider, and defaultExchanges from urql.
  • Here you’re instantiating a new Client and are passing it your endpoint url and a list of defaultExchanges
  • Finally you render the root component of your React app. The App is wrapped with the context Provider for the urql Client.

The defaultExchanges would also be applied automatically, but in the next step you’ll set up the normalized cache!

Set up additional urql Exchanges

By default urql sets up three built-in exchanges, which provide its core functionality. These are all part of the defaultExchanges array:

  • dedupExchange deduplicates operations. If you’re sending the same queries at the same time, then it will make sure that only one of them is actually sent to your API
  • cacheExchange caches operation results. This is only a document cache, so it caches results from your GraphQL API by the unique query + variables combination that those results have been requested with.
  • fetchExchange sends GraphQL requests using fetch and supports cancellation by default.

As you can see above, by default urql only comes with a simple document cache. This cache is very useful for content-heavy sites, since it treats every query and result as documents that it can simply cache 1:1. For more complex apps you will most likely want to use normalized caching, which makes sure that data updates globally across the app, if it can be shared across queries.

Let’s set up a normalized cache!

Let’s go through the changes we’ve made to index.js in order:

  • you’re now importing dedupExchange and fetchExchange from urql and have added cacheExchange from the additional @urql/exchange-graphcache package.
  • you’re creating a new normalized cache by calling cacheExchange with a config, which is empty for now.
  • lastly, you’ve replaced defaultExchanges on the Client with a new list of exchanges that includes the normalized cache exchange. The list is in a specific order (basically: fetch last and dedup first.)

That’s it, you’re all set to start for loading some data into your app! 😎

Backend

Downloading the server code

As mentioned above, for the backend in this tutorial you’ll simply use the final project from the Node tutorial.

Note: If you are on Windows, you may want to install Git CLI to avoid potential problems with commands such as curl.

Note: You can find the server’s code on the react-urql example repository in case the command is not working.

You now have a new directory called server inside your project that contains all the code you need for your backend.

Before we start the server, let’s quickly understand the main components:

  • prisma: This directory holds all the files that relate to your Prisma setup. Prisma Client is used to access the database in your GraphQL resolvers (similar to an ORM).

    • prisma.yml is the root configuration file for your Prisma project.
    • datamodel.prisma defines your data model in the GraphQL Schema Definition Language (SDL). When using Prisma, the datamodel is used to describe the database schema.
  • src: This directory holds the source files for your GraphQL server.

    • schema.graphql contains your application schema. The application schema defines the GraphQL operations you can send from the frontend. We’ll take a closer look at this file in just a bit.
    • generated/prisma-client contains the auto-generated Prisma client, a type-safe database access library (similar to an ORM).
    • resolvers contains the resolver functions for the operations defined in the application schema.
    • index.js is the entry point for your GraphQL server.

From the mentioned files, only the application schema defined in server/src/schema.graphql is relevant for you as a frontend developer. This file contains the GraphQL schema which defines all the operations (queries, mutations and subscriptions) you can send from your frontend app.

Here is what it looks like:

.../react-apollo/blob/master/server/src/schema.graphql
# import Link, Vote, LinkSubscriptionPayload, VoteSubscriptionPayload from "./generated/prisma.graphql"

type Query {
  feed(filter: String, skip: Int, first: Int, orderBy: LinkOrderByInput): Feed!
}

type Feed {
  links: [Link!]!
  count: Int!
}

type Mutation {
  post(url: String!, description: String!): Link!
  signup(email: String!, password: String!, name: String!): AuthPayload
  login(email: String!, password: String!): AuthPayload
  vote(linkId: ID!): Vote
}

type AuthPayload {
  token: String
  user: User
}

type User {
  id: ID!
  name: String!
  email: String!
}

type Subscription {
  newLink: LinkSubscriptionPayload
  newVote: VoteSubscriptionPayload
}

This schema allows for the following operations:

  • Queries:

    • feed: Retrieves all links from the backend, note that this query also allows for filter, sorting and pagination arguments
  • Mutations:

    • post: Allows authenticated users to create a new link
    • signup: Create an account for a new user
    • login: Login an existing user
    • vote: Allows authenticated users to vote for an existing link
  • Subscriptions:

    • newLink: Receive realtime updates when a new link is created
    • newVote: Receive realtime updates when a vote was submitted

For example, you can send the following feed query to retrieve the first 10 links from the server:

{
  feed(skip: 0, first: 10) {
    links {
      description
      url
      postedBy {
        name
      }
    }
  }
}

Or the signup mutation to create a new user:

mutation {
  signup(
    name: "Sarah",
    email: "sarah@graph.cool",
    password: "graphql"
  ) {
    token
    user {
      id
    }
  }
}

Deploying the Prisma database service

There is one thing left to do before you can start your server and begin sending queries and mutations to it. The Prisma project needs to be deployed so the GraphQL server can access it.

To deploy the service all you need to do is install the server’s dependencies and invoke the prisma deploy command inside the server directory.

Note that you can also omit yarn prisma in the above command if you have the prisma CLI installed globally on your machine (which you can do with yarn global add prisma). In that case, you can simply run prisma deploy.

Note: Once the command has finished running, the CLI writes the endpoint for the Prisma API to your prisma.yml. It will look similar to this: https://eu1.prisma.sh/john-doe/hackernews-node/dev.

Exploring the server

With the proper Prisma endpoint in place, you can now explore the server!

The yarn start executes the start script defined in package.json. The script first starts the server (which is then running on http://localhost:4000) and then opens up a GraphQL Playground for you to explore and work with the API.

carrying the authentication token

A Playground is a “GraphQL IDE”, providing an interactive environment that allows to send queries, mutations and subscriptions to your GraphQL API. It is similar to a tool like Postman which you might know from working with REST APIs, but comes with a lot of additional benefits.

The first thing to note about the Playground is that it has built-in documentation for its GraphQL API. This documentation is generated based on the GraphQL schema and can be opened by clicking the green SCHEMA-button on the right edge of the Playground. Consequently, it shows you the same information you saw in the application schema above:

The GraphQL Playground

The left pane of the Playground is the editor that you can use to write your queries, mutations and subscriptions. Once you click the play button in the middle, your request is sent and the server’s response will be displayed in the results pane on the right.

Since you’re adding two mutations to the editor at once, the mutations need to have operation names. In your case, these are CreatePrismaLink and CreateUrqlLink.

Clicking the play button to select the mutation

This creates two new Link records in the database. You can verify that the mutations actually worked by sending the following query in the already open Playground:

{
  feed {
    links {
      id
      description
      url
    }
  }
}

Note: You can also send the feed query in the default Playground in the app section.

If everything went well, the query will return the following data (the ids will of course be different in your case since they were generated by Prisma and are globally unique):

{
  "data": {
    "feed": {
      "links": [
        {
          "id": "cjcnfwjeif1rx012483nh6utk",
          "description": "The best GraphQL client",
          "url": "https://github.com/FormidableLabs/urql"
        },
        {
          "id": "cjcnfznzff1w601247iili50x",
          "description": "Prisma turns your database into a GraphQL API 😎",
          "url": "https://www.prismagraphql.com"
        }
      ]
    }
  }
}

Fantastic, your server works! 👏

Unlock the next chapter
Why are there two GraphQL API layers in a backend architecture with Prisma?
To increase robustness and stability of the GraphQL server (if one layer fails, the server is backed by the second one).
To increase performance of the GraphQL server (requests are accelerated by going through multiple layers).
Prisma provides the database layer which offers CRUD operations. The second layer is the application layer for business logic and common workflows (like authentication).
Having two GraphQL layers is a hard requirement by the GraphQL specification.