Step 1: Create Node.js Project

During this step, you want to ensure that at least the following things are taken care of:

• You can use the latest (modern) JavaScript syntax (most likely using Babel )

• You can launch the app by running yarn start which is a commonly used convention.

• When you make changes to the source code, the app (API) automatically restarts (see Nodemon).

• Node.js debugger is configured and ready to be used when needed.

Assuming that you already have Node.js and Yarn (or, NPM) installed, bootstrap a new Node.js project by running:

# Creates package.json file in the root of the project's folder
$ yarn init
​
# Installs Express.js and GraphQL.js NPM modules (runtime dependencies)
$ yarn add graphql express express-graphql
​
# Installs Nodemon and Babel (dev dependencies)
$ yarn add nodemon @babel/core @babel/cli @babel/preset-env --dev

Create .babelrc file containing Babel configuration. Add "start" script to the package.json file:

{
  "name": "api",
  "version": "1.0.0",
  "private": true,
  "dependencies": {
    "express": "^4.17.1",
    "express-graphql": "^0.9.0",
    "graphql": "^14.5.6"
  },
  "devDependencies": {
    "@babel/cli": "^7.6.0",
    "@babel/core": "^7.6.0",
    "@babel/preset-env": "^7.6.0",
    "nodemon": "^1.19.2"
  },
  "scripts": {
    "build": "babel src --out-dir build --source-maps=inline --delete-dir-on-start --copy-files --verbose",
    "start": "yarn build --watch & sleep 1 && nodemon --watch build build/index.js"
  }
}

{
  "presets": [
    ["@babel/preset-env", { "targets": { "node": "10.15.3" } }]
  ]
}

Finally, create src/index.js file that will serve as an entry point to the (Express.js) app:

src/index.js
import express from "express";
import graphql from "express-graphql";
​
const app = express();
const port = process.env.PORT || 8080;
​
app.use("/", graphql({
  schema: null, // TODO: Implement GraphQL schema
}));
​
app.listen(port, () => {
  console.log(`GraphQL API listening on http://localhost:${port}/`);
});

At this point, when you launch the app by running yarn start and navigate to http://localhost:8080/ in the browser's window, you must be able to see the following:

Step 2: Create GraphQL Schema

Basically, you create an GraphQL API by describing the hierarchy (schema) of all the (query) fields and mutations that the API must support.

Imagine that we need to fetch some OS-related info from the server using this query:

query {
  environment {
    arch
    platform
    uptime
  }
}

{
  data: {
    environment: {
      arch: 'x64',
      platform: 'linux',
      uptime: 11.256
    }
  }
}

It's asking for arch, platform, and uptime values grouped under the top-level envrionment field.

For this particular API we would need to implement just one GraphQL type (Environment), and one top-level query field environment. But since, we're planning to add more types and query fields later on, it would be a good idea to group them under src/types and src/queires folders. Plus, you would add src/schema.js file exporting the "schema" GraphQL object type.

​

The Environment GraphQL type is going to list arch, platform, and uptime fields, alongside their types and resolve() methods:

src/types/environment.js
import { GraphQLObjectType, GraphQLString, GraphQLFloat } from "graphql";
​
export const EnvironmentType = new GraphQLObjectType({
  name: "Environment",
​
  fields: {
    arch: {
      type: GraphQLString,
      resolve: () => process.arch
    },
​
    platform: {
      type: GraphQLString,
      resolve: () => process.platform
    },
​
    uptime: {
      type: GraphQLFloat,
      resolve: () => process.uptime()
    }
  }
});

Whenever a certain field is requested by the client, the GraphQL API runtime would call the corresponding resolve() function that would return the actual value for that field. Note that these resolve methods can be async (returning a Promise).

For the top-level fields, like environment field in our example, we're going to introduce yet another convention ⁠— placing them in multiple files under the src/queries folder. In many cases, those top-level fields would contain large resolve() functions and most likely you won't like having all of them within the same file. So, the environment field is going to be exported from src/queires/environment.js:

src/queries/environment.js
import { EnvironmentType } from "../types";
​
export const environment = {
  type: EnvironmentType,
  resolve: () => ({})
};

Note, that the field resolves to an empty object {}. If it would resolve to null or undefined the query traversal would stop right there, and the GraphQL query (from the example above) would resolve to:

{
  "data": {
    "environment": null
  }
}

You would also need src/queries/index.js and src/types/index.js re-exporting everything from the sibling files:

export * from './envrionment';

export * from './environment';

The environment field declaration we just created is going to be used in the root GraphQL object type:

src/schema.js
import { GraphQLSchema, GraphQLObjectType } from "graphql";
import * as queries from "./queries";
​
export default new GraphQLSchema({
  query: new GraphQLObjectType({
    name: "Query",
    fields: {
      ...queries
    }
  })
});

Finally, we're going to pass this schema type to the express-graphql middleware inside src/index.js :

src/index.js
import express from "express";
import graphql from "express-graphql";
import schema from "./schema";
​
const app = express();
const port = process.env.PORT || 8080;
​
app.use("/", graphql({
  schema,
  graphiql: process.env.NODE_ENV !== "production"
}));
​
app.listen(port, () => {
  console.log(`GraphQL API listening on http://localhost:${port}/`);
});

With all that in place, you must be able to test our first GraphQL query using GraphiQL IDE that express-graphql middleware provides out of the box:

Step 3: Create GraphQL Context

Inside of the resolve() methods we would often need access to the (request) context data such the currently logged-in user, data loaders (more on that later), etc.

For this purpose, let's create Context class and pass it to the express-graphql middleware alongside the schema:

src/context.js
export class Context {
  constructor(req) {
    this._req = req;
  }
​
  get user() {
    return this._req.user;
  }
​
  get hostname() {
    return this._req.hostname;
  }
}

src/index.js
import express from "express";
import graphql from "express-graphql";
import schema from "./schema";
import { Context } from "./context";
​
const app = express();
const port = process.env.PORT || 8080;
​
app.use("/", graphql(req => ({
  schema,
  context: new Context(req),
  graphiql: process.env.NODE_ENV !== "production"
})));
​
app.listen(port, () => {
  console.log(`GraphQL API listening on http://localhost:${port}/`);
});

Just for quick demonstration, let's see how to use this context object inside of a resolve() method:

src/types/environment.js
import { GraphQLObjectType, GraphQLString, GraphQLFloat } from "graphql";
​
export const EnvironmentType = new GraphQLObjectType({
  name: "Environment",
​
  fields: {
    ...,
  
    hostname: {
      type: GraphQLString,
      resolve(self, args, ctx) {
        return ctx.hostname;
      }
    }
  }
});

Step 4: Configure a Debugger

Sooner or later you may bump into a situation where you would need to debug your code. It would be wise to ensure that debugging is working OK, even before you actually need it. In case with VS Code, setting up the debugger takes just two simple steps:

• Pass --insepect argument to nodemon.

• Create a launch configuration that would attach to an existing Node.js process.

For the first step, you can copy and paste the start script inside of package.json file as follows:

package.json
{
  ...,
  "scripts": {
    "build": "babel src --out-dir build --source-maps=inline --delete-dir-on-start --copy-files --verbose",
    "start": "yarn build --watch & sleep 1 && nodemon --watch build build/index.js",
    "debug": "yarn build --watch & sleep 1 && nodemon --inspect --watch build build/index.js"
  }
}

And then create .vscode/launch.json file instructing VS Code how it should launch the debugger:

.vscode/launch.json
{
  "version": "0.2.0",
  "configurations": [
    {
      "type": "node",
      "request": "attach",
      "name": "Attach to Node.js",
      "processId": "${command:PickProcess}",
      "restart": true,
      "protocol": "inspector"
    }
  ]
}
​