Setting up our GraphQL server with Apollo Server

5. Apollo Server

🛠 Backend first steps

On the backend side, our first goal is to create a GraphQL server that can:

Receive an incoming GraphQL query from our client
Validate that query against our newly created schema
Populate the queried schema fields with mocked data
Return the populated fields as a response

The Apollo Server library helps us implement this server quickly, painlessly, and in a production-ready way.

In the server/src/ folder, open index.js.

To create our server, we'll use the @apollo/server package that we installed previously. From that package, we'll only need the named export ApolloServer, so we'll declare that constant between curly braces.

We'll also need to use the startStandaloneServer function, which we can import from the @apollo/server/standalone package.

Just below, we'll import our typeDefs from our schema.js file:

const { ApolloServer } = require("@apollo/server");
const { startStandaloneServer } = require("@apollo/server/standalone");
const typeDefs = require("./schema");

Next, let's set up an async function called startApolloServer. Inside, we'll create an instance of the ApolloServer class and pass it our typeDefs in its options object:

async function startApolloServer() {
  const server = new ApolloServer({ typeDefs });
}

Note: We're using shorthand property notation with implied keys, because we've named our constant with the matching key (typeDefs).

To start the server, we'll use the startStandaloneServer function, passing it the server we just initialized.

async function startApolloServer() {
  const server = new ApolloServer({ typeDefs });
  startStandaloneServer(server);
}

The startStandaloneServer function returns a Promise, so we'll await the results of that call, and pull out the url property from the result.

async function startApolloServer() {
  const server = new ApolloServer({ typeDefs });
  const { url } = await startStandaloneServer(server);
}

We'll also log a nice little message letting us know that our server is indeed up and running!

async function startApolloServer() {
  const server = new ApolloServer({ typeDefs });
  const { url } = await startStandaloneServer(server);
  console.log(`
    🚀  Server is running!
    📭  Query at ${url}
  `);
}

Finally, let's not forget to actually call the startApolloServer function at the bottom of the file!

startApolloServer();

Save your changes. From the terminal, we'll launch our server with npm run start (make sure you're in the server/ folder).

We get the log message and...not much else! We have a running server, but that's it. Floating in the vacuum of localhost space without access to any data, it's a sad and lonely server for now. 😿

Which of these are purposes of a GraphQL server?

Creating GraphQL queriesExposing a separate endpoint for each schema typeValidating GraphQL queries against our schemaReturning populated schema fields as a responseReceiving incoming GraphQL queries

Even though our server isn't connected to any data sources yet, it would be great to be able to send the server a test query and get a valid response. Fortunately, ApolloServer provides a way to do exactly that, using mocked data.

Mocking data

To enable mocked data, we'll need to use two new packages: @graphql-tools/mock and @graphql-tools/schema.

Let's go ahead and install them.

npm install @graphql-tools/mock @graphql-tools/schema

At the top, we'll import addMocksToSchema and makeExecutableSchema.

const { addMocksToSchema } = require("@graphql-tools/mock");
const { makeExecutableSchema } = require("@graphql-tools/schema");

Then, we'll need to tweak the ApolloServer initialization. Instead of passing it the typeDefs directly, we'll be using the schema property. This property is another way of initializing an Apollo Server, which is useful for building federated subgraphs (more on that in our Voyage series) or if we're using functions like makeExecutableSchema (which we are!).

As the value of the schema property, we'll call the addMocksToSchema function and pass it an object. This object defines its own schema property, and here we'll call the makeExecutableSchema function. Then, we'll pass this function an object containing our typeDefs.

const server = new ApolloServer({
  schema: addMocksToSchema({
    schema: makeExecutableSchema({ typeDefs }),
  }),
});

With this code, we're generating an executable schema from our typeDefs, and instructing Apollo Server to populate every queried schema field with a placeholder value (such as Hello World for String fields).

However, Hello World isn't a very realistic value for the title of a track or the URL of an author's picture! To serve mocked data that's closer to reality, we'll define a mocks object. This object contains functions that provide the mocked data we want the server to return for each queried field.

Here's our mocks object:

const mocks = {
  Track: () => ({
    id: () => "track_01",
    title: () => "Astro Kitty, Space Explorer",
    author: () => {
      return {
        name: "Grumpy Cat",
        photo:
          "https://res.cloudinary.com/dety84pbu/image/upload/v1606816219/kitty-veyron-sm_mctf3c.jpg",
      };
    },
    thumbnail: () =>
      "https://res.cloudinary.com/dety84pbu/image/upload/v1598465568/nebula_cat_djkt9r.jpg",
    length: () => 1210,
    modulesCount: () => 6,
  }),
};

This object defines mock values for all of the fields of a Track object (including the Author object it contains). We pass this object to the ApolloServer constructor like so:

const server = new ApolloServer({
  schema: addMocksToSchema({
    schema: makeExecutableSchema({ typeDefs }),
    mocks,
  }),
});

With mocks enabled, Apollo Server always returns exactly two entries for every list field.
To get more entries at a time, let's say 6, we'll add a Query.tracksForHome to our mocks object and return an Array of that given length like so: [...new Array(6)].

1
const mocks = {
2
  Query: () => ({
3
    tracksForHome: () => [...new Array(6)],
4
  }),
5
  Track: () => ({
6
    id: () => "track_01",
7
    title: () => "Astro Kitty, Space Explorer",
8
    author: () => {
9
      return {
10
        name: "Grumpy Cat",
11
        photo:
12
          "https://res.cloudinary.com/dety84pbu/image/upload/v1606816219/kitty-veyron-sm_mctf3c.jpg",
13
      };
14
    },
15
    thumbnail: () =>
16
      "https://res.cloudinary.com/dety84pbu/image/upload/v1598465568/nebula_cat_djkt9r.jpg",
17
    length: () => 1210,
18
    modulesCount: () => 6,
19
  }),
20
};

Which of these are true about querying Apollo Server without a connected data source?

You can enable default mocked responses for every schema field.Apollo Server rejects incoming requests if it is not connected to a data source.You can configure custom mocked responses for every schema field.

Code Challenge!

Create a mock object with a type SpaceCat, an id of spacecat_01, and a title of 'spacecat pioneer'

Loading progress

Now, with our server loaded with mocked data, how can we run a query on it to test if everything works as expected? In the next lesson, we'll use the Apollo Explorer to build and run test queries seamlessly.

Share your questions and comments about this lesson

Your feedback helps us improve! If you're stuck or confused, let us know and we'll help you out. All comments are public and must follow the Apollo Code of Conduct. Note that comments that have been resolved or addressed may be removed.

You'll need a GitHub account to post below. Don't have one? Post in our Odyssey forum instead.