GraphQL is a modern approach to APIs that simplifies integration. This is an introduction to what GraphQL is, and we build a simple GraphQL service with the help of Vert.X.
What you’ll need to follow along
This introduction will barely scratch the surface of the potential and power that GraphQL together with Vert.X offers. Understanding of programming and the REST protocol is assumed.
Make sure to have the following installed:
- JDK 8+
- Maven
- IDE
Source code for this post can be found here: https://github.com/GustafNilstadius/intro-to-graphql-with-vertx.
Background
Imagine this, you have a book store with 3 defined data entities; Book, Author and Quote. Where every book has an author and every author has a favorite quote.
If we want to retrieve data from our book store with a traditional REST call in our newly developed SuperMegaAwesomeBookStoreTM app we have two options.
Option 1
Separate REST GET requests for each entity.
- We can make a call to
/booksto get a list of all the books. - A call to
/authors/:authorIDto get the author. - Lastly a call
/quotes/:quoteIDto get the favorite quote of the author.
Option 2
Returning a deep object for one REST GET request.
- We can make a call to
/booksto get a list of all the books complete with all the authors and quotes.
The data
Independent of the method (option 1 or 2) we end up with the same data exampled bellow.
[
{
"id": "ABCD",
"name": "Awesome book name",
"page_count": 42,
"author": {
"id": "ABC123",
"firstName": "Author",
"lastName": "Authorson",
"favoriteQuote": {
"id": "123",
"value": "Not too little, nor too much. Just right."
}
}
}
]
Either we risk getting too little data in one call e.g. if we want to display all books with the first name of the author. Alternatively we risk getting too much data in one call e.g. if we just want to list all the book names.
This is where GraphQL comes in, GraphQL enable the client to get lagom (“lagom”, Swedish, “Not too little, nor too much. Just right.”) amount of data.
GraphQL
GraphQL or Graph Query Language is as it suggests a query language for graph data. GraphQL is a communication pattern and not an implementation or API in itself. Developed by Facebook back in 2012, GraphQL is today a part of the GraphQL foundation.
GraphQL enables the clients of a GraphQL service to get just the right amount of data.
How does GraphQL solve the problems previously described and always serves lagom amount of data to its clients?
It’s actually rather simple, the client tells the backend exactly what it wants, and the backend will provide.
But how does it actually work? First we need to describe our data, in a so called schema. Then from the schema we can construct queries. With a schema we can then query the GraphQL service.
- Describe your data.
- Query.
- Get what you asked for.
Describe your data
To make your book store data query-able we first have to define it. In GraphQL this is called a schema and can either be defined programmatically depending on the implementation or in a .graphql SDL (Schema Definition Language) file.
In this example we’ll use the SDL approach. It looks like JSON, but it’s not.
So let’s start describing our data.
Note: “scalar” is what a data type is called in GraphQL.
Books
type Book {
id: ID
name: String
pageCount: Int
author: Author
}
Looking closer at the definition for book. Book is the object we are defining.
A Book has four members, id of the scalar type ID that signals to GraphQL that the field is a unique identifier and not always human readable name of the scalar String, pageCount of scalar Int and lastly author of type Author.
Author
The definition of author continues in the same fashion as for books.
type Author {
id: ID
firstName: String
lastName: String
favoriteQuote: Quote
}
Quote
type Quote {
id: ID
value: String
}
Schema
Now with all the objects defined, let’s put it all together in a schema. A GraphQL service is required to have a query type defined and may have a mutation type. mutation can be equaled to POST in REST, but is out of scope for this introduction.
schema.graphql
type Query {
bookById(id: ID): Book
getBooks: [Book]
}
type Book {
id: ID
name: String
pageCount: Int
author: Author
}
type Author {
id: ID
firstName: String
lastName: String
favoriteQuote: Quote
}
type Quote {
id: ID
value: String
}
The obligatory Query object defines two queries, these queries will later be exposed to our clients.
bookById(id: ID): Book
As the name suggests, it’s a query to find a book by id. It takes an argument of scalar ID, and it returns an object of type Book.
getBooks: [Book]
Returns an array of type Book.
We have now defined a GraphQL schema, the schema can be updated at any time, this makes versioning in a GraphQL service easy. As long as no fields are removed, it can be updated without breaking the clients.
Query
Let’s construct our first query. We want to get all books, and for each book we want to get the number of pages and the author’s first name.
query {
getBooks {
pageCount,
author {
firstName
}
}
}
Note: operation type query can be omitted in most cases. Added for clarity and best practice
Example result for above query:
{
"data": {
"getBooks": [
{
"pageCount": 276,
"author": {
"firstName": "William"
}
},
{
"pageCount": 635,
"author": {
"firstName": "Herman"
}
}
]
}
}
Not that exactly what was requested is also returned. Nothing more, nothing less.
We not have a client that wants to get the favorite quote of each author in addition to the above data, how would we change the query to do that?
Just ask for it! By adding the favoriteQuote element to the author we can get the favorite quote for each of the authors.
query {
getBooks {
pageCount,
author {
firstName
favoriteQuote {
value
}
}
}
}
Example result:
{
"data": {
"getBooks": [
{
"pageCount": 276,
"author": {
"firstName": "William",
"favoriteQuote": {
"value": "Chuck Norris calculated the square root of negative one while eating a bowl full of rusty fishhooks."
}
}
},
{
"pageCount": 635,
"author": {
"firstName": "Herman",
"favoriteQuote": {
"value": "The only way to make saturday every day is to kill Chuck Norris. Which is 99999999999999999999999999999999% impossible."
}
}
}
]
}
}
Vert.X
Vert.X describes itself as follows;
Vert.x is a tool-kit for building reactive applications on the JVM. Reactive applications are both scaleable as workloads grow, and resilient when failures arise. A reactive application is responsive as it keeps latency under control by making efficient usage of system resources, and by protecting itself from errors.
Vert.X can best be described as a modern and asynchronous alternative to Spring-boot or a good version of NodeJS. I can go in depth about Vert.X, but it’s out of scope for this introduction.
The golden rule for Vert.X is to never block the event loop. Every module in Vert.X runs inside what’s called a verticle.
public class MainVerticle extends AbstractVerticle {
private BookService bookService;
private AuthorService authorService;
private QuoteService quoteService;
@Override
public void start(Promise<Void> startPromise) {
/* Do stuff, like init GraphQL, deploy microservices and start HTTP servers */
/* Don't forget to complete the startPromise */
}
}
GraphQL with micro-services in Vert.X (Java)
Due to the highly asynchronous nature of Vert.X and GraphQL they are a great match. Let’s put together what we learnt about GraphQL and make a practical demo in Vert.X.
Remember that GraphQL is not an implementation in itself, therefore we are using the implementation GraphQL-Java in this example.
Init GraphQL-Java
Let’s initiate our GraphQL instance.
src/main/java/com/redpilllinpro/graphqldemo/MainVerticle.java
private GraphQL setupGraphQL() {
/* Read the schema file from the file system. */
String schema = vertx.fileSystem().readFileBlocking("schema/schema.graphql").toString();
/* (1) Parse schema and create a TypeDefinitionRegistry */
SchemaParser schemaParser = new SchemaParser();
TypeDefinitionRegistry typeDefinitionRegistry = schemaParser.parse(schema);
/* (2) RuntimeWiring linking our schema/TypeDefinitionRegistry to our services */
RuntimeWiring runtimeWiring = newRuntimeWiring()
.type(newTypeWiring("Query")
.dataFetcher("bookById", bookByIdDataFetcher())
.dataFetcher("getBooks", booksDataFetcher()))
.type(newTypeWiring("Book")
.dataFetcher("author", authorDataFetcher()))
.type(newTypeWiring("Author")
.dataFetcher("favoriteQuote", quoteDataFetcher()))
.build();
SchemaGenerator schemaGenerator = new SchemaGenerator();
GraphQLSchema graphQLSchema = schemaGenerator.makeExecutableSchema(typeDefinitionRegistry, runtimeWiring);
return GraphQL.newGraphQL(graphQLSchema).build();
}
- TypeDefinitionRegistry The TypeDefinitionRegistry is the GraphQL-Java representation of our schema file.
- RuntimeWiring links our TypeDefinitionRegistry (schema) to our DataFetchers for each of the objects.
The rest of the initiation inside the Vert.X verticle starter.
src/main/java/com/redpilllinpro/graphqldemo/MainVerticle.java
public void start(Promise<Void> startPromise) {
/* Omitted, data service initiation */
GraphQLHandlerOptions graphQLHandlerOptions = new GraphQLHandlerOptions()
.setRequestBatchingEnabled(true);
GraphQL graphQL = setupGraphQL();
GraphQLHandler graphQLHandler = GraphQLHandler.create(graphQL, graphQLHandlerOptions);
/* Omitted */
}
Start HTTP server in Vert.X
src/main/java/com/redpilllinpro/graphqldemo/MainVerticle.java
public void start(Promise<Void> startPromise) {
/* Omitted, data service initiation */
/* GraphQL-Java stuff */
/* GraphQL graphical interface, GraphiQL. */
GraphiQLHandlerOptions options = new GraphiQLHandlerOptions()
.setEnabled(true);
/* Creating a HTTP router */
Router router = Router.router(vertx);
router.route().handler(LoggerHandler.create());
router.post().handler(BodyHandler.create());
/* Binding our GraphQL handler to a endpoint, can be any endpoint. One server can have muliple GraphQL endpoints. */
router.route("/graphql").handler(graphQLHandler);
/* Binding the GraphiQL interface to a endpoint */
router.route("/graphiql/*").handler(GraphiQLHandler.create(options));
/* Error handling */
router.errorHandler(500, ctx -> {
ctx.failure().printStackTrace();
ctx.response().setStatusCode(500).end();
});
/* Strat server on port 8888 */
vertx.createHttpServer().requestHandler(router).listen(8888, http -> {
if (http.succeeded()) {
startPromise.complete();
System.out.println("HTTP server started on port 8888");
} else {
startPromise.fail(http.cause());
}
});
}
DataFetchers
src/main/java/com/redpilllinpro/graphqldemo/MainVerticle.java
public void start(Promise<Void> startPromise) {
bookService = BookService.create(vertx);
authorService = AuthorService.create(vertx);
quoteService = QuoteService.create(vertx);
}
See sources for details on the data services. QuoteService is a little extra interesting.
Build the code and try our GraphQL service
Build and run our Vert.X service.
$ mvn clean package
$ java -jar target/graphqldemo-1.0.0-SNAPSHOT-fat.jar
HTTP server started on port 8888
Head over to http://localhost:8888/graphiql/index.html and try out our service!
GraphQL is a great tool to try out a GraphQL service, a good starting point is to try the query we wrote before.
This will get all books and some select data members of every book.
query {
getBooks {
id,
pageCount,
author {
firstName
favoriteQuote {
value
}
}
}
}
Playing around in GraphiQL is a good way get an idea of what GraphQL is and the power of the protocol. If we now want to query just one book the query might look something like this:
query {
bookById(id: "book-3") {
pageCount
}
}
Note that we are giving the id of the book as an argument to the query.
GraphQL also supports multiple queries in one request, that means that we can combine the two queries above into one query.
query {
getBooks {
id,
pageCount,
author {
firstName
favoriteQuote {
value
}
}
}
bookById(id: "book-3") {
pageCount
}
}
Thanks for reading through this brief introduction to GraphQL. The official documentation on both GraphQL and Vert.X is a great resource, and a good next step if this introduction sparked an interest.
