In 2019, I reinvented Common Object Request Brokerage Architecture (CORBA). If you developed a distributed system in 1997, you might have built servers that accept network requests by CORBA. You defined the types and operations implemented by a server in Interface Definition Language (IDL). An IDL compiler generated skeleton code for a server. As an application developer, you added business logic to the skeleton code to complete the server implementation.

I wrote Grapid, a framework for building GraphQL servers in Java. You define the types and operations implemented by a server in GraphQL schema definition language. The framework generates code to connect GraphQL to Java types. As an application developer, you only need to implement the business logic for the API.

This article explains how to build a GraphQL server in Java with Spring Boot.

At Spring Initializr, generate a project with these options:

Option	Value
Project	Maven Project
Language	Java
Spring Boot	2.1.3
Package Name	com.example.demo
Java Version	11

Option

Value

Project

Maven Project

Language

Java

Spring Boot

2.1.3

Package Name

com.example.demo

Java Version

Edit the pom.xml file, and add this Spring Boot starter which auto-configures a GraphQL server:

<dependency>
  <groupId>com.github.pukkaone</groupId>
  <artifactId>grapid-web-spring-boot-starter</artifactId>
  <version>0.3.0</version>
</dependency>

Also add this Maven plugin which compiles GraphQL schema definition files to Java source files:

<plugin>
  <groupId>com.github.pukkaone</groupId>
  <artifactId>grapid-maven-plugin</artifactId>
  <version>0.3.0</version>
  <configuration>
    <packagePrefix>com.example.demo.graphql</packagePrefix> (1)
  </configuration>
  <executions>
    <execution>
      <goals>
        <goal>compile</goal>
      </goals>
    </execution>
  </executions>
</plugin>

1	The compiler generates Java classes under this Java package.

The framework enforces a convention prescribing where you put GraphQL schema definition files and Java classes. Create resources directory src/main/resources/graphql/v2018_12_31/. In this example, v2018_12_31 identifies an API version. An API version must be a valid Java identifier and not a Java keyword.

Add this GraphQL schema definition file in the directory. By convention, GraphQL schema definition file names end with the extension .graphql.

Author.graphql

type Author {
  id: ID!
  name: String!
}

type Query {
  author(id: ID!): Author
}

The GraphQL schema defines the root object type Query. The compiler appends the suffix Resolver to this root object type name to derive the Java class name QueryResolver. As an application developer, you must implement the QueryResolver class. By convention, this class is in the Java package named packagePrefix.resolver where packagePrefix is the property configured in the Maven plugin.

QueryResolver.java

package com.example.demo.graphql.resolver;

import com.example.demo.graphql.v2018_12_31.type.Author; (1)
import org.springframework.stereotype.Component;

@Component
public class QueryResolver {

  public Author author(String id) { (2)
    return Author.builder()
        .id(id)
        .name("NAME")
        .build();
  }
}

1	The compiler generated the simple Java data class Author from the GraphQL object type Author.
2	The compiler translated this Java method signature from the field `author` of the GraphQL root object type Query.

Run the application:

./mvnw spring-boot:run

In GraphQL Playground, connect to http://localhost:8080/graphql/v2018_12_31 to send a GraphQL query to the server.