This guide explains how to write REST Services with RESTEasy Reactive in Quarkus.
Tip
|
This is the reference guide for RESTEasy Reactive. For a more lightweight introduction, please refer to the Writing JSON REST services guides. |
RESTEasy Reactive is a new Jakarta REST (formerly known as JAX-RS) implementation written from the ground up to work on our common Vert.x layer and is thus fully reactive, while also being very tightly integrated with Quarkus and consequently moving a lot of work to build time.
You should be able to use it in place of any Jakarta REST implementation, but on top of that it has great performance for both blocking and non-blocking endpoints, and a lot of new features on top of what Jakarta REST provides.
Add the following import to your build file:
<dependency>
<groupId>io.quarkus</groupId>
<artifactId>quarkus-resteasy-reactive</artifactId>
</dependency>
implementation("io.quarkus:quarkus-resteasy-reactive")
You can now write your first endpoint in the org.acme.rest.Endpoint
class:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
@Path("")
public class Endpoint {
@GET
public String hello() {
return "Hello, World!";
}
}
- REST
- Endpoint
-
Java method which is called to serve a REST call
- URL / URI (Uniform Resource Locator / Identifier)
-
Used to identify the location of REST resources (specification)
- Resource
-
Represents your domain object. This is what your API serves and modifies. Also called an
entity
in Jakarta REST. - Representation
-
How your resource is represented on the wire, can vary depending on content types
- Content type
-
Designates a particular representation (also called a media type), for example
text/plain
orapplication/json
- HTTP
-
Underlying wire protocol for routing REST calls (see HTTP specifications)
- HTTP request
-
the request part of the HTTP call, consisting of an HTTP method, a target URI, headers and an optional message body
- HTTP response
-
the response part of the HTTP call, consisting of an HTTP response status, headers and an optional message body
Any class annotated with a @Path
annotation can have its methods exposed as REST endpoints,
provided they have an HTTP method annotation (see below).
That @Path
annotation defines the URI prefix under which those methods will be exposed. It can
be empty, or contain a prefix such as rest
or rest/V1
.
Each exposed endpoint method can in turn have another @Path
annotation which adds to its containing
class annotation. For example, this defines a rest/hello
endpoint:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
@Path("rest")
public class Endpoint {
@Path("hello")
@GET
public String hello() {
return "Hello, World!";
}
}
See URI parameters for more information about URI mapping.
You can set the root path for all rest endpoints using the @ApplicationPath
annotation, as shown below.
package org.acme.rest;
import jakarta.ws.rs.ApplicationPath;
import jakarta.ws.rs.core.Application;
@ApplicationPath("/api")
public static class MyApplication extends Application {
}
This will cause all rest endpoints to be resolve relative to /api
, so the endpoint above with @Path("rest")
would
be accessible at /api/rest/
. You can also set the quarkus.resteasy-reactive.path
build time property to set the root path if you
don’t want to use an annotation.
Each endpoint method must be annotated with one of the following annotations, which defines which HTTP method will be mapped to the method:
Annotation | Usage |
---|---|
Obtain a resource representation, should not modify state, idempotent (HTTP docs) |
|
Obtain metadata about a resource, similar to |
|
Create a resource and obtain a link to it (HTTP docs) |
|
Replace a resource or create one, should be idempotent (HTTP docs) |
|
Delete an existing resource, idempotent (HTTP docs) |
|
Obtain information about a resource, idempotent (HTTP docs) |
|
Update a resource, or create one, not idempotent (HTTP docs) |
You can also declare other HTTP methods by declaring them as an annotation with the
@HttpMethod
annotation:
package org.acme.rest;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import jakarta.ws.rs.HttpMethod;
import jakarta.ws.rs.Path;
@Retention(RetentionPolicy.RUNTIME)
@HttpMethod("CHEESE")
@interface CHEESE {
}
@Path("")
public class Endpoint {
@CHEESE
public String hello() {
return "Hello, Cheese World!";
}
}
Each endpoint method may consume or produce specific resource representations, which are indicated by
the HTTP Content-Type
header, which in turn contains
MIME (Media Type) values, such as the following:
-
text/plain
which is the default for any endpoint returning aString
. -
text/html
for HTML (such as with Qute templating) -
application/json
for a JSON REST endpoint -
text/*
which is a sub-type wildcard for any text media type -
*/*
which is a wildcard for any media type
You may annotate your endpoint class with the @Produces
or @Consumes
annotations, which
allow you to specify one or more media types that your endpoint may accept as HTTP request body
or produce as HTTP response body. Those class annotations apply to each method.
Any method may also be annotated with the @Produces
or @Consumes
annotations, in which
case they override any eventual class annotation.
The MediaType
class has many constants you
can use to point to specific pre-defined media types.
See the Negotiation section for more information.
Note
|
don’t forget to configure your compiler to generate parameter name information with -parameters (javac)
or <parameters> or <maven.compiler.parameters> (Maven).
|
The following HTTP request elements may be obtained by your endpoint method:
HTTP element | Annotation | Usage |
---|---|---|
|
URI template parameter (simplified version of the URI Template specification), see URI parameters for more information. |
|
Query parameter |
The value of a URI query parameter |
|
Header |
The value of an HTTP header |
|
Cookie |
The value of an HTTP cookie |
|
Form parameter |
The value of an HTTP URL-encoded FORM |
|
Matrix parameter |
The value of an URI path segment parameter |
For each of those annotations, you may specify the name of the element they refer to, otherwise, they will use the name of the annotated method parameter.
If a client made the following HTTP call:
POST /cheeses;variant=goat/tomme?age=matured HTTP/1.1
Content-Type: application/x-www-form-urlencoded
Cookie: level=hardcore
X-Cheese-Secret-Handshake: fist-bump
smell=strong
Then you could obtain all the various parameters with this endpoint method:
package org.acme.rest;
import jakarta.ws.rs.POST;
import jakarta.ws.rs.Path;
import org.jboss.resteasy.reactive.RestCookie;
import org.jboss.resteasy.reactive.RestForm;
import org.jboss.resteasy.reactive.RestHeader;
import org.jboss.resteasy.reactive.RestMatrix;
import org.jboss.resteasy.reactive.RestPath;
import org.jboss.resteasy.reactive.RestQuery;
@Path("/cheeses/{type}")
public class Endpoint {
@POST
public String allParams(@RestPath String type,
@RestMatrix String variant,
@RestQuery String age,
@RestCookie String level,
@RestHeader("X-Cheese-Secret-Handshake")
String secretHandshake,
@RestForm String smell) {
return type + "/" + variant + "/" + age + "/" + level + "/"
+ secretHandshake + "/" + smell;
}
}
Note
|
The @RestPath
annotation is optional: any parameter whose name matches an existing URI
template variable will be automatically assumed to have @RestPath .
|
You can also use any of the Jakarta REST annotations @PathParam
,
@QueryParam
,
@HeaderParam
,
@CookieParam
,
@FormParam
or
@MatrixParam
for this,
but they require you to specify the parameter name.
See Parameter mapping for more advanced use-cases.
Note
|
When an exception occurs in RESTEasy Reactive request parameter handling code, the exception is not printed by default to the log (for security reasons).
This can sometimes make it hard to understand why certain HTTP status codes are returned (as the Jakarta REST mandates the use of non-intuitive error codes in various cases).
In such cases, users are encouraged to set the logging level for the quarkus.log.category."org.jboss.resteasy.reactive.server.handlers.ParameterHandler".level=DEBUG |
You can group your request parameters in a container class instead of declaring them as method parameters to you endpoint, so we can rewrite the previous example like this:
package org.acme.rest;
import jakarta.ws.rs.POST;
import jakarta.ws.rs.Path;
import org.jboss.resteasy.reactive.RestCookie;
import org.jboss.resteasy.reactive.RestForm;
import org.jboss.resteasy.reactive.RestHeader;
import org.jboss.resteasy.reactive.RestMatrix;
import org.jboss.resteasy.reactive.RestPath;
import org.jboss.resteasy.reactive.RestQuery;
@Path("/cheeses/{type}")
public class Endpoint {
public static class Parameters {
@RestPath
String type;
@RestMatrix
String variant;
@RestQuery
String age;
@RestCookie
String level;
@RestHeader("X-Cheese-Secret-Handshake")
String secretHandshake;
@RestForm
String smell;
}
@POST
public String allParams(@BeanParam Parameters parameters) { (1)
return parameters.type + "/" + parameters.variant + "/" + parameters.age
+ "/" + parameters.level + "/" + parameters.secretHandshake
+ "/" + parameters.smell;
}
}
-
BeanParam
is required to comply with the Jakarta REST specification so that libraries like OpenAPI can introspect the parameters.
You can declare URI parameters and use regular expressions in your path, so for instance
the following endpoint will serve requests for /hello/stef/23
and /hello
but not
/hello/stef/0x23
:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
@Path("hello")
public class Endpoint {
@Path("{name}/{age:\\d+}")
@GET
public String personalisedHello(String name, int age) {
return "Hello " + name + " is your age really " + age + "?";
}
@GET
public String genericHello() {
return "Hello stranger";
}
}
Any method parameter with no annotation will receive the method body.[1], after it has been mapped from its HTTP representation to the Java type of the parameter.
The following parameter types will be supported out of the box:
Type | Usage |
---|---|
The entire request body in a temporary file |
|
|
The entire request body, not decoded |
|
The entire request body, decoded |
The entire request body, decoded |
|
The request body in a blocking stream |
|
The request body in a blocking stream |
|
All Java primitives and their wrapper classes |
Java primitive types |
Large integers and decimals. |
|
JSON value types |
|
Vert.x Buffer |
|
any other type |
Will be mapped from JSON to that type |
Note
|
You can add support for more body parameter types. |
To handle HTTP requests that have multipart/form-data
as their content type, you can use the regular
@RestForm
annotation, but we have special types
that allow you to access the parts as files or as entities.
Let us look at an example of its use.
Assuming an HTTP request containing a file upload, a JSON entity and a form value containing a string description, we could write the following endpoint:
import jakarta.ws.rs.POST;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.MediaType;
import org.jboss.resteasy.reactive.PartType;
import org.jboss.resteasy.reactive.RestForm;
import org.jboss.resteasy.reactive.multipart.FileUpload;
@Path("multipart")
public class MultipartResource {
public static class Person {
public String firstName;
public String lastName;
}
@POST
public void multipart(@RestForm String description,
@RestForm("image") FileUpload file,
@RestForm @PartType(MediaType.APPLICATION_JSON) Person person) {
// do something
}
}
The description
parameter will contain the data contained in the part of HTTP request called description
(because
@RestForm
does not define a value, the field name is used),
while the file
parameter will contain data about the uploaded file in the image
part of HTTP request, and
the person
parameter will read the Person
entity using the JSON
body reader.
The size of every part in a multipart request must conform to the value of quarkus.http.limits.max-form-attribute-size
, for which the default is 2048 bytes.
Any request with a part of a size exceeding this configuration will result in HTTP status code 413.
Note
|
FileUpload
provides access to various metadata of the uploaded file. If however all you need is a handle to the uploaded file, java.nio.file.Path or java.io.File could be used.
|
If you need access to all uploaded files for all parts regardless of their names, you can do it with @RestForm(FileUpload.ALL) List<FileUpload>
.
Note
|
@PartType is used to aid
in deserialization of the corresponding part of the request into the desired Java type. It is only required if
you need to use a special body parameter type for that particular parameter.
|
Note
|
Just like for any other request parameter type, you can also group them into a container class. |
Warning
|
When handling file uploads, it is very important to move the file to permanent storage (like a database, a dedicated file system or a cloud storage) in your code that handles the POJO.
Otherwise, the file will no longer be accessible when the request terminates.
Moreover, if quarkus.http.body.delete-uploaded-files-on-end is set to true, Quarkus will delete the uploaded file when the HTTP response is sent. If the setting is disabled,
the file will reside on the file system of the server (in the directory defined by the quarkus.http.body.uploads-directory configuration option), but as the uploaded files are saved
with a UUID file name and no additional metadata is saved, these files are essentially a random dump of files.
|
Tip
|
When a Resource method needs to handle various types of multipart requests, then the The following code shows a simple example where we iterate over the parts and return a list of aggregated data: @Path("/test")
public static class Resource {
@POST
@Consumes(MediaType.MULTIPART_FORM_DATA)
@Produces(MediaType.APPLICATION_JSON)
public List<Item> hello(MultipartFormDataInput input) throws IOException {
Map<String, Collection<FormValue>> map = input.getValues();
List<Item> items = new ArrayList<>();
for (var entry : map.entrySet()) {
for (FormValue value : entry.getValue()) {
items.add(new Item(
entry.getKey(),
value.isFileItem() ? value.getFileItem().getFileSize() : value.getValue().length(),
value.getCharset(),
value.getFileName(),
value.isFileItem(),
value.getHeaders()));
}
}
return items;
}
public static class Item {
public final String name;
public final long size;
public final String charset;
public final String fileName;
public final boolean isFileItem;
public final Map<String, List<String>> headers;
public Item(String name, long size, String charset, String fileName, boolean isFileItem,
Map<String, List<String>> headers) {
this.name = name;
this.size = size;
this.charset = charset;
this.fileName = fileName;
this.isFileItem = isFileItem;
this.headers = headers;
}
}
} |
As part of reading the multipart body, RESTEasy Reactive invokes the proper MessageBodyReaderMessageBodyReader
for each part of the request.
If an IOException
occurs for one of these parts (for example if Jackson was unable to deserialize a JSON part), then a org.jboss.resteasy.reactive.server.multipart.MultipartPartReadingException
is thrown.
If this exception is not handled by the application as mentioned in Exception mapping, an HTTP 400 response is returned by default.
Similarly, RESTEasy Reactive can produce Multipart Form data to allow users download files from the server. For example, we could write a POJO that will hold the information we want to expose as:
import jakarta.ws.rs.core.MediaType;
import org.jboss.resteasy.reactive.PartType;
import org.jboss.resteasy.reactive.RestForm;
public class DownloadFormData {
@RestForm
String name;
@RestForm
@PartType(MediaType.APPLICATION_OCTET_STREAM)
File file;
}
And then expose this POJO via a Resource like so:
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;
@Path("multipart")
public class Endpoint {
@GET
@Produces(MediaType.MULTIPART_FORM_DATA)
@Path("file")
public DownloadFormData getFile() {
// return something
}
}
Additionally, you can also manually append the parts of the form using the class MultipartFormDataOutput
as:
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;
import org.jboss.resteasy.reactive.server.multipart.MultipartFormDataOutput;
@Path("multipart")
public class Endpoint {
@GET
@Produces(MediaType.MULTIPART_FORM_DATA)
@Path("file")
public MultipartFormDataOutput getFile() {
MultipartFormDataOutput form = new MultipartFormDataOutput();
form.addFormData("person", new Person("John"), MediaType.APPLICATION_JSON_TYPE);
form.addFormData("status", "a status", MediaType.TEXT_PLAIN_TYPE)
.getHeaders().putSingle("extra-header", "extra-value");
return form;
}
}
This last approach allows you adding extra headers to the output part.
Warning
|
For the time being, returning Multipart data is limited to be blocking endpoints. |
In order to return an HTTP response, simply return the resource you want from your method. The method return type and its optional content type will be used to decide how to serialise it to the HTTP response (see the Negotiation section for more advanced information).
You can return any of the pre-defined types that you can read from the HTTP response, and any other type will be mapped from that type to JSON.
In addition, the following return types are also supported:
Type | Usage |
---|---|
The contents of the file specified by the given path |
|
The partial contents of the file specified by the given path |
|
The partial contents of a file |
|
Vert.x AsyncFile, which can be in full, or partial |
Alternately, you can also return a reactive type such as Uni
,
Multi
or
CompletionStage
that resolve to one of the mentioned return types.
If you need to set more properties on the HTTP response than just the body, such as the status code
or headers, you can make your method return org.jboss.resteasy.reactive.RestResponse
from a resource method.
An example of this could look like:
package org.acme.rest;
import java.time.Duration;
import java.time.Instant;
import java.util.Date;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.MediaType;
import jakarta.ws.rs.core.NewCookie;
import org.jboss.resteasy.reactive.RestResponse;
import org.jboss.resteasy.reactive.RestResponse.ResponseBuilder;
@Path("")
public class Endpoint {
@GET
public RestResponse<String> hello() {
// HTTP OK status with text/plain content type
return ResponseBuilder.ok("Hello, World!", MediaType.TEXT_PLAIN_TYPE)
// set a response header
.header("X-Cheese", "Camembert")
// set the Expires response header to two days from now
.expires(Date.from(Instant.now().plus(Duration.ofDays(2))))
// send a new cookie
.cookie(new NewCookie("Flavour", "chocolate"))
// end of builder API
.build();
}
}
Note
|
You can also use the Jakarta REST type Response but it is
not strongly typed to your entity.
|
Alternatively, if you only need to set the status code and / or HTTP headers with static values, you can use @org.jboss.resteasy.reactive.ResponseStatus
and /or ResponseHeader
respectively.
An example of this could look like:
package org.acme.rest;
import org.jboss.resteasy.reactive.Header;
import org.jboss.resteasy.reactive.ResponseHeaders;
import org.jboss.resteasy.reactive.ResponseStatus;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
@Path("")
public class Endpoint {
@ResponseStatus(201)
@ResponseHeader(name = "X-Cheese", value = "Camembert")
@GET
public String hello() {
return "Hello, World!";
}
}
When handling a @POST
, @PUT
or @DELETE
endpoint, it is common practice to redirect to a @GET
endpoint after the action has been performed to allow the user to reload the page without triggering the action a second time.
There are multiple ways to achieve this.
Using RestResponse
as the return type while making sure the proper redirection URI is created can be done as in the following example:
package org.acme.rest;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.atomic.AtomicLong;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.POST;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.Context;
import jakarta.ws.rs.core.UriInfo;
import org.jboss.resteasy.reactive.RestResponse;
@Path("/fruits")
public class FruitResource {
public static class Fruit {
public Long id;
public String name;
public String description;
public Fruit() {
}
public Fruit(Long id, String name, String description) {
this.id = id;
this.name = name;
this.description = description;
}
}
private final Map<Long, Fruit> fruits = new ConcurrentHashMap<>();
private final AtomicLong ids = new AtomicLong(0);
public FruitResource() {
Fruit apple = new Fruit(ids.incrementAndGet(), "Apple", "Winter fruit");
fruits.put(apple.id, apple);
Fruit pinneapple = new Fruit(ids.incrementAndGet(), "Pineapple", "Tropical fruit");
fruits.put(pinneapple.id, pinneapple);
}
// when invoked, this method will result in an HTTP redirect to the GET method that obtains the fruit by id
@POST
public RestResponse<Fruit> add(Fruit fruit, @Context UriInfo uriInfo) {
fruit.id = ids.incrementAndGet();
fruits.put(fruit.id, fruit);
// seeOther results in an HTTP 303 response with the Location header set to the value of the URI
return RestResponse.seeOther(uriInfo.getAbsolutePathBuilder().path(Long.toString(fruit.id)).build());
}
@GET
@Path("{id}")
public Fruit byId(Long id) {
return fruits.get(id);
}
}
If your endpoint method needs to accomplish an asynchronous or reactive task before
being able to answer, you can declare your method to return the
Uni
type (from Mutiny), in which
case the current HTTP request will be automatically suspended after your method, until
the returned Uni
instance resolves to a value,
which will be mapped to a response exactly according to the previously described rules:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import io.smallrye.mutiny.Uni;
@Path("escoffier")
public class Endpoint {
@GET
public Uni<Book> culinaryGuide() {
return Book.findByIsbn("978-2081229297");
}
}
This allows you to not block the event-loop thread while the book is being fetched from the database, and allows Quarkus to serve more requests until your book is ready to be sent to the client and terminate this request. See Execution Model documentation for more information.
The CompletionStage
return
type is also supported.
If you want to stream your response element by element, you can make your endpoint method return a
Multi
type (from Mutiny).
This is especially useful for streaming text or binary data.
This example, using Reactive Messaging HTTP shows how to stream text data:
package org.acme.rest;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import org.eclipse.microprofile.reactive.messaging.Channel;
import io.smallrye.mutiny.Multi;
@Path("logs")
public class Endpoint {
@Inject
@Channel("log-out")
Multi<String> logs;
@GET
public Multi<String> streamLogs() {
return logs;
}
}
Note
|
Response filters are not invoked on streamed responses, because they would give a false impression that you can set headers or HTTP status codes, which is not true after the initial response. Exception mappers are also not invoked because part of the response may already have been written. |
If you need to set custom HTTP headers and / or the HTTP response, then you can return org.jboss.resteasy.reactive.RestMulti
instead, like this:
package org.acme.rest;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import org.eclipse.microprofile.reactive.messaging.Channel;
import io.smallrye.mutiny.Multi;
import org.jboss.resteasy.reactive.RestMulti;
@Path("logs")
public class Endpoint {
@Inject
@Channel("log-out")
Multi<String> logs;
@GET
public Multi<String> streamLogs() {
return RestMulti.fromMultiData(logs).status(222).header("foo", "bar").build();
}
}
In more advanced cases where the headers and / or status can only be obtained from the results of an async call, the RestMulti.fromUniResponse
needs to be used.
Here is an example of such a use case:
package org.acme.rest;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import java.util.List;import java.util.Map;import org.eclipse.microprofile.reactive.messaging.Channel;
import io.smallrye.mutiny.Multi;
import org.jboss.resteasy.reactive.RestMulti;
@Path("logs")
public class Endpoint {
interface SomeService {
Uni<SomeResponse> get();
}
interface SomeResponse {
Multi<byte[]> data;
String myHeader();
}
private final SomeService someService;
public Endpoint(SomeService someService) {
this.someService = someService;
}
@GET
public Multi<String> streamLogs() {
return RestMulti.fromUniResponse(someService.get(), SomeResponse::data, (r -> Map.of("MyHeader", List.of(r.myHeader()))));
}
}
If you want to stream JSON objects in your response, you can use
Server-Sent Events
by just annotating your endpoint method with
@Produces(MediaType.SERVER_SENT_EVENTS)
and specifying that each element should be serialised to JSON with
@RestStreamElementType(MediaType.APPLICATION_JSON)
.
package org.acme.rest;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;
import org.jboss.resteasy.reactive.RestStreamElementType;
import io.smallrye.mutiny.Multi;
import org.eclipse.microprofile.reactive.messaging.Channel;
@Path("escoffier")
public class Endpoint {
// Inject our Book channel
@Inject
@Channel("book-out")
Multi<Book> books;
@GET
// Each element will be sent as JSON
@RestStreamElementType(MediaType.APPLICATION_JSON)
// by using @RestStreamElementType, we don't need to add @Produces(MediaType.SERVER_SENT_EVENTS)
public Multi<Book> stream() {
return books;
}
}
Sometimes it’s useful to create a customized SSE message, for example if you need to specify the event
field of a SSE message to distinguish various event types.
A resource method may return Multi<jakarta.ws.rs.sse.OutboundSseEvent>
and an injected jakarta.ws.rs.sse.Sse
can be used to create OutboundSseEvent
instances.
package org.acme.rest;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;
import jakarta.ws.rs.sse.OutboundSseEvent;
import jakarta.ws.rs.sse.Sse;
import org.jboss.resteasy.reactive.RestStreamElementType;
import io.smallrye.mutiny.Multi;
import org.eclipse.microprofile.reactive.messaging.Channel;
@Path("escoffier")
public class Endpoint {
@Inject
@Channel("book-out")
Multi<Book> books;
@Inject
Sse sse; (1)
@GET
@RestStreamElementType(MediaType.TEXT_PLAIN)
public Multi<OutboundSseEvent> stream() {
return books.map(book -> sse.newEventBuilder() (2)
.name("book") (3)
.data(book.title) (4)
.build());
}
}
-
Inject the server-side entry point for creating
OutboundSseEvent
s. -
Create a new outbound event builder.
-
Set the event name, i.e. the value of the
event
field of a SSE message. -
Set the data, i.e. the value of the
data
field of a SSE message.
Warning
|
Manipulation of the returned HTTP headers and status code is not possible via |
RESTEasy Reactive provides the @Cache
and @NoCache
annotations to facilitate
handling HTTP caching semantics, i.e. setting the 'Cache-Control' HTTP header.
These annotations can be placed either on a Resource Method or a Resource Class (in which case it applies to all Resource Methods of the class that do not contain the same annotation) and allow users
to return domain objects and not have to deal with building up the Cache-Control
HTTP header explicitly.
While @Cache
builds a complex Cache-Control
header, @NoCache
is a simplified notation to say that you don’t want anything cached; i.e. Cache-Control: nocache
.
Note
|
More information on the Cache-Control header and be found in RFC 7234
|
There are a number of contextual objects that the framework will give you, if your endpoint method takes parameters of the following type:
Type | Usage |
---|---|
All the request headers |
|
Information about the current endpoint method and class (requires reflection) |
|
Access to the current user and roles |
|
Information about the current endpoint method and class (no reflection required) |
|
Provides information about the current endpoint and application URI |
|
Advanced: Current Jakarta REST application class |
|
Advanced: Configuration about the deployed Jakarta REST application |
|
Advanced: Runtime access to Jakarta REST providers |
|
Advanced: Access to the current HTTP method and Preconditions |
|
Advanced: access to instances of endpoints |
|
Advanced: RESTEasy Reactive access to the current request/response |
|
Advanced: Complex SSE use-cases |
|
Advanced: Vert.x HTTP Request |
|
Advanced: Vert.x HTTP Response |
For example, here is how you can return the name of the currently logged-in user:
package org.acme.rest;
import java.security.Principal;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.SecurityContext;
@Path("user")
public class Endpoint {
@GET
public String userName(SecurityContext security) {
Principal user = security.getUserPrincipal();
return user != null ? user.getName() : "<NOT LOGGED IN>";
}
}
You can also inject those context objects using
@Inject
on fields of the same
type:
package org.acme.rest;
import java.security.Principal;
import jakarta.inject.Inject;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.SecurityContext;
@Path("user")
public class Endpoint {
@Inject
SecurityContext security;
@GET
public String userName() {
Principal user = security.getUserPrincipal();
return user != null ? user.getName() : "<NOT LOGGED IN>";
}
}
Or even on your endpoint constructor:
package org.acme.rest;
import java.security.Principal;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.SecurityContext;
@Path("user")
public class Endpoint {
SecurityContext security;
Endpoint(SecurityContext security) {
this.security = security;
}
@GET
public String userName() {
Principal user = security.getUserPrincipal();
return user != null ? user.getName() : "<NOT LOGGED IN>";
}
}
Instead of importing io.quarkus:quarkus-resteasy-reactive
, you can import either of the following modules to get support for JSON:
GAV | Usage |
---|---|
|
|
|
In both cases, importing those modules will allow HTTP message bodies to be read from JSON and serialised to JSON, for all the types not already registered with a more specific serialisation.
When using the quarkus-resteasy-reactive-jackson
extension there are some advanced features that RESTEasy Reactive supports.
When used with Jackson to perform JSON serialization, RESTEasy Reactive provides the ability to limit the set of fields that are serialized based on the roles of the current user.
This is achieved by simply annotating the fields (or getters) of the POJO being returned with @io.quarkus.resteasy.reactive.jackson.SecureField
.
A simple example could be the following:
Assume we have a POJO named Person
which looks like so:
package org.acme.rest;
import io.quarkus.resteasy.reactive.jackson.SecureField;
public class Person {
@SecureField(rolesAllowed = "admin")
private final Long id;
private final String first;
private final String last;
public Person(Long id, String first, String last) {
this.id = id;
this.first = first;
this.last = last;
}
public Long getId() {
return id;
}
public String getFirst() {
return first;
}
public String getLast() {
return last;
}
}
A very simple Jakarta REST Resource that uses Person
could be:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
@Path("person")
public class Person {
@Path("{id}")
@GET
public Person getPerson(Long id) {
return new Person(id, "foo", "bar");
}
}
Assuming security has been set up for the application (see our guide for more details), when a user with the admin
role
performs an HTTP GET on /person/1
they will receive:
{
"id": 1,
"first": "foo",
"last": "bar"
}
as the response.
Any user however that does not have the admin
role will receive:
{
"first": "foo",
"last": "bar"
}
Note
|
No additional configuration needs to be applied for this secure serialization to take place. However, users can use the @io.quarkus.resteasy.reactive.jackson.EnableSecureSerialization and @io.quarkus.resteasy.reactive.jackson.DisableSecureSerialization
annotation to opt in or out for specific Jakarta REST Resource classes or methods.
|
Jakarta REST methods can be annotated with @JsonView in order to customize the serialization of the returned POJO, on a per method-basis. This is best explained with an example.
A typical use of @JsonView
is to hide certain fields on certain methods. In that vein, let’s define two views:
public class Views {
public static class Public {
}
public static class Private extends Public {
}
}
Let’s assume we have the User
POJO on which we want to hide some field during serialization. A simple example of this is:
public class User {
@JsonView(Views.Private.class)
public int id;
@JsonView(Views.Public.class)
public String name;
}
Depending on the Jakarta REST method that returns this user, we might want to exclude the id
field from serialization.
For example, you might want an insecure method to not expose this field.
The way we can achieve that in RESTEasy Reactive is shown in the following example:
@JsonView(Views.Public.class)
@GET
@Path("/public")
public User userPublic() {
return testUser();
}
@JsonView(Views.Private.class)
@GET
@Path("/private")
public User userPrivate() {
return testUser();
}
When the result the userPublic
method is serialized, the id
field will not be contained in the response as the Public
view does not include it.
The result of userPrivate
however will include the id
as expected when serialized.
There are times when you need to completely customize the serialization of a POJO on a per Jakarta REST method basis. For such use cases, the @io.quarkus.resteasy.reactive.jackson.CustomSerialization
annotation
is a great tool, as it allows you to configure a per-method com.fasterxml.jackson.databind.ObjectWriter
which can be configured at will.
Here is an example use case:
@CustomSerialization(UnquotedFields.class)
@GET
@Path("/invalid-use-of-custom-serializer")
public User invalidUseOfCustomSerializer() {
return testUser();
}
where UnquotedFields
is a BiFunction
defined as so:
public static class UnquotedFields implements BiFunction<ObjectMapper, Type, ObjectWriter> {
@Override
public ObjectWriter apply(ObjectMapper objectMapper, Type type) {
return objectMapper.writer().without(JsonWriteFeature.QUOTE_FIELD_NAMES);
}
}
Essentially what this class does is force Jackson to not include quotes in the field names.
It is important to note that this customization is only performed for the serialization of the Jakarta REST methods that use @CustomSerialization(UnquotedFields.class)
.
To enable XML support, add the quarkus-resteasy-reactive-jaxb
extension to your project.
GAV | Usage |
---|---|
|
Importing this module will allow HTTP message bodies to be read from XML and serialised to XML, for all the types not already registered with a more specific serialisation.
The JAXB Resteasy Reactive extension will automatically detect the classes that are used in the resources and require JAXB serialization. Then, it will register these classes into the default JAXBContext
which is internally used by the JAXB message reader and writer.
However, in some situations, these classes cause the JAXBContext
to fail: for example, when you’re using the same class name in different java packages. In these cases, the application will fail at build time and print the JAXB exception that caused the issue, so you can properly fix it. Alternatively, you can also exclude the classes that cause the issue by using the property quarkus.jaxb.exclude-classes
. When excluding classes that are required by any resource, the JAXB resteasy reactive extension will create and cache a custom JAXBContext
that will include the excluded class, causing a minimal performance degradance.
Note
|
The property |
When using the quarkus-resteasy-reactive-jaxb
extension there are some advanced features that RESTEasy Reactive supports.
The JAXB resteasy reactive extension will serialize and unserialize requests and responses transparently for users. However, if you need finer grain control over JAXB components, you can inject either the JAXBContext, Marshaller, or Unmarshaller components into your beans:
@ApplicationScoped
public class MyService {
@Inject
JAXBContext jaxbContext;
@Inject
Marshaller marshaller;
@Inject
Unmarshaller unmarshaller;
// ...
}
Note
|
Quarkus will automatically find all the classes annotated with |
To customize the JAXB configuration for either the JAXB context, and/or the Marshaller/Unmarshaller components, the suggested approach is to define a CDI bean of type io.quarkus.jaxb.runtime.JaxbContextCustomizer
.
An example where a custom module needs to be registered would look like so:
@Singleton
public class RegisterCustomModuleCustomizer implements JaxbContextCustomizer {
// For JAXB context configuration
@Override
public void customizeContextProperties(Map<String, Object> properties) {
}
// For Marshaller configuration
@Override
public void customizeMarshaller(Marshaller marshaller) throws PropertyException {
marshaller.setProperty("jaxb.formatted.output", Boolean.TRUE);
}
// For Unmarshaller configuration
@Override
public void customizeUnmarshaller(Unmarshaller unmarshaller) throws PropertyException {
// ...
}
}
Note
|
It’s not necessary to implement all three methods, but only the want you need. |
Alternatively, you can provide your own JAXBContext
bean by doing:
public class CustomJaxbContext {
// Replaces the CDI producer for JAXBContext built into Quarkus
@Singleton
@Produces
JAXBContext jaxbContext() {
// ...
}
}
Important
|
Note that if you provide your custom JAXB context instance, you will need to register the classes you want to use for the XML serialization. This means that Quarkus will not update your custom JAXB context instance with the auto-discovered classes. |
To enable Web Links support, add the quarkus-resteasy-reactive-links
extension to your project.
GAV | Usage |
---|---|
|
Importing this module will allow injecting web links into the response HTTP headers by just annotating your endpoint resources with the @InjectRestLinks
annotation. To declare the web links that will be returned, you need to use the @RestLink
annotation in the linked methods. An example of this could look like:
@Path("/records")
public class RecordsResource {
@GET
@RestLink(rel = "list")
@InjectRestLinks
public List<Record> getAll() {
// ...
}
@GET
@Path("/{id}")
@RestLink(rel = "self")
@InjectRestLinks(RestLinkType.INSTANCE)
public TestRecord get(@PathParam("id") int id) {
// ...
}
@PUT
@Path("/{id}")
@RestLink
@InjectRestLinks(RestLinkType.INSTANCE)
public TestRecord update(@PathParam("id") int id) {
// ...
}
@DELETE
@Path("/{id}")
@RestLink
public TestRecord delete(@PathParam("id") int id) {
// ...
}
}
When calling the endpoint /records
which is defined by the method getAll
within the above resource using curl, you would get the web links header:
& curl -i localhost:8080/records
Link: <http://localhost:8080/records>; rel="list"
As this resource does not return a single instance of type Record
, the links for the methods get
, update
, and delete
are not injected. Now, when calling the endpoint /records/1
, you would get the following web links:
& curl -i localhost:8080/records/1
Link: <http://localhost:8080/records>; rel="list"
Link: <http://localhost:8080/records/1>; rel="self"
Link: <http://localhost:8080/records/1>; rel="update"
Link: <http://localhost:8080/records/1>; rel="delete"
The get
, update
, and delete
methods use the path param "id" and as the field "id" exists in the entity type "Record", the web link properly populates the value "1" in the returned links. In addition to this, we can also generate web links with path params that do not match with any field of the entity type. For example, the following method is using a path param "text" and the entity Record does not have any field named "text":
@Path("/records")
public class RecordsResource {
// ...
@GET
@Path("/search/{text}")
@RestLink(rel = "search records by free text")
@InjectRestLinks
public List<Record> search(@PathParam("text") String text) { (4)
// ...
}
// ...
}
The generated web link for this resource is Link: http://localhost:8080/search/{text}; rel="search records by free text"
.
Finally, when calling the delete
resource, you should not see any web links as the method delete
is not annotated with the @InjectRestLinks
annotation.
You can programmatically have access to the web links registry just by injecting the RestLinksProvider
bean:
@Path("/records")
public class RecordsResource {
@Inject
RestLinksProvider linksProvider;
// ...
}
Using this injected bean of type RestLinksProvider
, you can get the links by type using the method RestLinksProvider.getTypeLinks
or get the links by a concrete instance using the method RestLinksProvider.getInstanceLinks
.
The HAL standard is a simple format to represent web links.
To enable the HAL support, add the quarkus-hal
extension to your project. Also, as HAL needs JSON support, you need to add either the quarkus-resteasy-reactive-jsonb
or the quarkus-resteasy-reactive-jackson
extension.
GAV | Usage |
---|---|
|
After adding the extensions, we can now annotate the REST resources to produce the media type application/hal+json
(or use RestMediaType.APPLICATION_HAL_JSON). For example:
@Path("/records")
public class RecordsResource {
@GET
@Produces({ MediaType.APPLICATION_JSON, RestMediaType.APPLICATION_HAL_JSON })
@RestLink(rel = "list")
@InjectRestLinks
public List<Record> getAll() {
// ...
}
@GET
@Produces({ MediaType.APPLICATION_JSON, RestMediaType.APPLICATION_HAL_JSON })
@Path("/{id}")
@RestLink(rel = "self")
@InjectRestLinks(RestLinkType.INSTANCE)
public TestRecord get(@PathParam("id") int id) {
// ...
}
}
Now, the endpoints /records
and /records/{id}
will accept the media type both json
and hal+json
to print the records in Hal format.
For example, if we invoke the /records
endpoint using curl to return a list of records, the HAL format will look like as follows:
& curl -H "Accept:application/hal+json" -i localhost:8080/records
{
"_embedded": {
"items": [
{
"id": 1,
"slug": "first",
"value": "First value",
"_links": {
"self": {
"href": "http://localhost:8081/records/1"
},
"list": {
"href": "http://localhost:8081/records"
}
}
},
{
"id": 2,
"slug": "second",
"value": "Second value",
"_links": {
"self": {
"href": "http://localhost:8081/records/2"
},
"list": {
"href": "http://localhost:8081/records"
}
}
}
]
},
"_links": {
"list": {
"href": "http://localhost:8081/records"
}
}
}
When we call a resource /records/1
that returns only one instance, then the output is:
& curl -H "Accept:application/hal+json" -i localhost:8080/records/1
{
"id": 1,
"slug": "first",
"value": "First value",
"_links": {
"self": {
"href": "http://localhost:8081/records/1"
},
"list": {
"href": "http://localhost:8081/records"
}
}
}
Finally, you can also provide additional HAL links programmatically in your resource just by returning either HalCollectionWrapper
(to return a list of entities) or HalEntityWrapper
(to return a single object) as described in the following example:
@Path("/records")
public class RecordsResource {
@Inject
RestLinksProvider linksProvider;
@GET
@Produces({ MediaType.APPLICATION_JSON, RestMediaType.APPLICATION_HAL_JSON })
@RestLink(rel = "list")
public HalCollectionWrapper getAll() {
List<Record> list = // ...
HalCollectionWrapper halCollection = new HalCollectionWrapper(list, "collectionName", linksProvider.getTypeLinks(Record.class));
halCollection.addLinks(Link.fromPath("/records/1").rel("first-record").build());
return halCollection;
}
@GET
@Produces({ MediaType.APPLICATION_JSON, RestMediaType.APPLICATION_HAL_JSON })
@Path("/{id}")
@RestLink(rel = "self")
@InjectRestLinks(RestLinkType.INSTANCE)
public HalEntityWrapper get(@PathParam("id") int id) {
Record entity = // ...
HalEntityWrapper halEntity = new HalEntityWrapper(entity, linksProvider.getInstanceLinks(entity));
halEntity.addLinks(Link.fromPath("/records/1/parent").rel("parent-record").build());
return halEntity;
}
}
Cross-origin resource sharing (CORS) is a mechanism that allows restricted resources on a web page to be requested from another domain outside the domain from which the first resource was served.
Quarkus comes with a CORS filter at the HTTP layer level. Read the HTTP Reference Documentation to learn how to use it.
Here are some more advanced topics that you may not need to know about initially, but could prove useful for more complex use cases.
RESTEasy Reactive is implemented using two main thread types:
-
Event-loop threads: which are responsible, among other things, for reading bytes from the HTTP request and writing bytes back to the HTTP response
-
Worker threads: they are pooled and can be used to offload long-running operations
The event-loop threads (also called IO threads) are responsible for actually performing all the IO operations in an asynchronous way, and to trigger any listener interested in the completion of those IO operations.
By default, the thread RESTEasy Reactive will run endpoint methods on depends on the signature of the method. If a method returns one of the following types then it is considered non-blocking, and will be run on the IO thread by default:
-
io.smallrye.mutiny.Uni
-
io.smallrye.mutiny.Multi
-
java.util.concurrent.CompletionStage
-
org.reactivestreams.Publisher
-
Kotlin
suspended
methods
This 'best guess' approach means most operations will run on the correct thread by default. If you are writing reactive code, your method will generally return one of these types and will be executed on the IO thread. If you are writing blocking code, your methods will usually return the result directly, and these will be run on a worker thread.
You can override this behaviour using the
@Blocking
and
@NonBlocking
annotations. This can be applied at the method, class or jakarta.ws.rs.core.Application
level.
The example below will override the default behaviour and always run on a worker thread, even though it returns a Uni
.
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import io.smallrye.common.annotation.Blocking;
@Path("yawn")
public class Endpoint {
@Blocking
@GET
public Uni<String> blockingHello() throws InterruptedException {
// do a blocking operation
Thread.sleep(1000);
return Uni.createFrom().item("Yaaaawwwwnnnnnn…");
}
}
Most of the time, there are ways to achieve the same blocking operations in an asynchronous/reactive way, using Mutiny, Hibernate Reactive or any of the Quarkus Reactive extensions for example:
package org.acme.rest;
import java.time.Duration;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import io.smallrye.mutiny.Uni;
@Path("yawn")
public class Endpoint {
@GET
public Uni<String> blockingHello() throws InterruptedException {
return Uni.createFrom().item("Yaaaawwwwnnnnnn…")
// do a non-blocking sleep
.onItem().delayIt().by(Duration.ofSeconds(2));
}
}
If a method or class is annotated with jakarta.transaction.Transactional
then it will also be treated as a blocking
method. This is because JTA is a blocking technology, and is generally used with other blocking technology such as
Hibernate and JDBC. An explicit @Blocking
or @NonBlocking
on the class will override this behaviour.
If you want to override the default behavior, you can annotate a jakarta.ws.rs.core.Application
subclass in your application with @Blocking
or @NonBlocking
,
and this will set the default for every method that does not have an explicit annotation.
Behavior can still be overridden on a class or method level by annotating them directly, however, all endpoints without an annotation will now follow the default, no matter their method signature.
If your application needs to return non-nominal HTTP codes in error cases, the best is
to throw exceptions that will result in the proper HTTP response being sent by the
framework using WebApplicationException
or any of its subtypes:
package org.acme.rest;
import jakarta.ws.rs.BadRequestException;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.NotFoundException;
import jakarta.ws.rs.Path;
@Path("cheeses/{cheese}")
public class Endpoint {
@GET
public String findCheese(String cheese) {
if(cheese == null)
// send a 400
throw new BadRequestException();
if(!cheese.equals("camembert"))
// send a 404
throw new NotFoundException("Unknown cheese: " + cheese);
return "Camembert is a very nice cheese";
}
}
If your endpoint method is delegating calls to another service layer which
does not know of Jakarta REST, you need a way to turn service exceptions to an
HTTP response, and you can do that using the
@ServerExceptionMapper
annotation on a method, with one parameter of the exception type you want to handle, and turning
that exception into a RestResponse
(or a
Uni<RestResponse<?>>
):
package org.acme.rest;
import java.util.Map;
import jakarta.enterprise.context.ApplicationScoped;
import jakarta.inject.Inject;
import jakarta.ws.rs.BadRequestException;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.Response;
import org.jboss.resteasy.reactive.server.ServerExceptionMapper;
import org.jboss.resteasy.reactive.RestResponse;
class UnknownCheeseException extends RuntimeException {
public final String name;
public UnknownCheeseException(String name) {
this.name = name;
}
}
@ApplicationScoped
class CheeseService {
private static final Map<String, String> cheeses =
Map.of("camembert", "Camembert is a very nice cheese",
"gouda", "Gouda is acceptable too, especially with cumin");
public String findCheese(String name) {
String ret = cheeses.get(name);
if(ret != null)
return ret;
throw new UnknownCheeseException(name);
}
}
@Path("cheeses/{cheese}")
public class Endpoint {
@Inject
CheeseService cheeses;
@ServerExceptionMapper
public RestResponse<String> mapException(UnknownCheeseException x) {
return RestResponse.status(Response.Status.NOT_FOUND, "Unknown cheese: " + x.name);
}
@GET
public String findCheese(String cheese) {
if(cheese == null)
// send a 400
throw new BadRequestException();
return cheeses.findCheese(cheese);
}
}
Note
|
exception mappers defined in REST endpoint classes will only be called if the exception is thrown in the same class. If you want to define global exception mappers, simply define them outside a REST endpoint class: |
package org.acme.rest;
import org.jboss.resteasy.reactive.server.ServerExceptionMapper;
import org.jboss.resteasy.reactive.RestResponse;
class ExceptionMappers {
@ServerExceptionMapper
public RestResponse<String> mapException(UnknownCheeseException x) {
return RestResponse.status(Response.Status.NOT_FOUND, "Unknown cheese: " + x.name);
}
}
You can also declare exception mappers in the Jakarta REST way.
Your exception mapper may declare any of the following parameter types:
Type | Usage |
---|---|
An exception type |
Defines the exception type you want to handle |
Any of the Context objects |
|
A context object to access the current request |
It may declare any of the following return types:
Type | Usage |
---|---|
The response to send to the client when the exception occurs |
|
An asynchronous response to send to the client when the exception occurs |
Note
|
When an exception occurs, RESTEasy Reactive does not log it by default (for security reasons).
This can sometimes make it hard to understand why certain exception handling code was invoked (or not invoked).
To make RESTEasy Reactive log the actual exception before an exception mapping code is run the quarkus.log.category."org.jboss.resteasy.reactive.common.core.AbstractResteasyReactiveContext".level=DEBUG |
You can declare functions which are invoked in the following phases of the request processing:
-
Before the endpoint method is identified: pre-routing request filter
-
After routing, but before the endpoint method is called: normal request filter
-
After the endpoint method is called: response filter
These filters allow you to do various things such as examine the request URI, HTTP method, influence routing, look or change request headers, abort the request, or modify the response.
Request filters can be declared with the
@ServerRequestFilter
annotation:
import java.util.Optional;
class Filters {
@ServerRequestFilter(preMatching = true)
public void preMatchingFilter(ContainerRequestContext requestContext) {
// make sure we don't lose cheese lovers
if("yes".equals(requestContext.getHeaderString("Cheese"))) {
requestContext.setRequestUri(URI.create("/cheese"));
}
}
@ServerRequestFilter
public Optional<RestResponse<Void>> getFilter(ContainerRequestContext ctx) {
// only allow GET methods for now
if(ctx.getMethod().equals(HttpMethod.GET)) {
return Optional.of(RestResponse.status(Response.Status.METHOD_NOT_ALLOWED));
}
return Optional.empty();
}
}
Important
|
Request filters are usually executed on the same thread that the method that handles the request will be executed.
That means that if the method servicing the request is annotated with If however a filter needs to be run on the event-loop despite the fact that the method servicing the request will be
run on a worker thread, then |
Similarly, response filters can be declared with the
@ServerResponseFilter
annotation:
class Filters {
@ServerResponseFilter
public void getFilter(ContainerResponseContext responseContext) {
Object entity = responseContext.getEntity();
if(entity instanceof String) {
// make it shout
responseContext.setEntity(((String)entity).toUpperCase());
}
}
}
Your filters may declare any of the following parameter types:
Type | Usage |
---|---|
Any of the Context objects |
|
A context object to access the current request |
|
A context object to access the current response |
|
Any thrown exception, or |
It may declare any of the following return types:
Type | Usage |
---|---|
The response to send to the client instead of continuing the filter chain, or |
|
An optional response to send to the client instead of continuing the filter chain, or an empty value if the filter chain should proceed |
|
An asynchronous response to send to the client instead of continuing the filter chain, or |
Note
|
You can restrict the Resource methods for which a filter runs, by using @NameBinding meta-annotations.
|
Both HTTP request and response can be intercepted by providing ContainerRequestFilter
or ContainerResponseFilter
implementations respectively. These filters are suitable for processing the metadata associated with a message: HTTP
headers, query parameters, media type, and other metadata. They also have the capability to abort the request
processing, for instance when the user does not have the permissions to access the endpoint.
Let’s use ContainerRequestFilter
to add logging capability to our service. We can do that by implementing
ContainerRequestFilter
and annotating it with the @Provider
annotation:
package org.acme.rest.json;
import io.vertx.core.http.HttpServerRequest;
import org.jboss.logging.Logger;
import jakarta.ws.rs.container.ContainerRequestContext;
import jakarta.ws.rs.container.ContainerRequestFilter;
import jakarta.ws.rs.core.Context;
import jakarta.ws.rs.core.UriInfo;
import jakarta.ws.rs.ext.Provider;
@Provider
public class LoggingFilter implements ContainerRequestFilter {
private static final Logger LOG = Logger.getLogger(LoggingFilter.class);
@Context
UriInfo info;
@Context
HttpServerRequest request;
@Override
public void filter(ContainerRequestContext context) {
final String method = context.getMethod();
final String path = info.getPath();
final String address = request.remoteAddress().toString();
LOG.infof("Request %s %s from IP %s", method, path, address);
}
}
Now, whenever a REST method is invoked, the request will be logged into the console:
2019-06-05 12:44:26,526 INFO [org.acm.res.jso.LoggingFilter] (executor-thread-1) Request GET /legumes from IP 127.0.0.1
2019-06-05 12:49:19,623 INFO [org.acm.res.jso.LoggingFilter] (executor-thread-1) Request GET /fruits from IP 0:0:0:0:0:0:0:1
2019-06-05 12:50:44,019 INFO [org.acm.res.jso.LoggingFilter] (executor-thread-1) Request POST /fruits from IP 0:0:0:0:0:0:0:1
2019-06-05 12:51:04,485 INFO [org.acm.res.jso.LoggingFilter] (executor-thread-1) Request GET /fruits from IP 127.0.0.1
Whenever your endpoint methods return an object (of when they return a
RestResponse<?>
or Response
with
an entity), RESTEasy Reactive will look for a way to map that into an HTTP response body.
Similarly, whenever your endpoint method takes an object as parameter, we will look for a way to map the HTTP request body into that object.
This is done via a pluggable system of MessageBodyReader
and MessageBodyWriter
interfaces,
which are responsible for defining which Java type they map from/to, for which media types,
and how they turn HTTP bodies to/from Java instances of that type.
For example, if we have our own Cheese
type on our endpoint:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.PUT;
import jakarta.ws.rs.Path;
class Cheese {
public String name;
public Cheese(String name) {
this.name = name;
}
}
@Path("cheese")
public class Endpoint {
@GET
public Cheese sayCheese() {
return new Cheese("Cheeeeeese");
}
@PUT
public void addCheese(Cheese cheese) {
System.err.println("Received a new cheese: " + cheese.name);
}
}
Then we can define how to read and write it with our body reader/writers, annotated
with @Provider
:
package org.acme.rest;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.lang.annotation.Annotation;
import java.lang.reflect.Type;
import java.nio.charset.StandardCharsets;
import jakarta.ws.rs.WebApplicationException;
import jakarta.ws.rs.core.MediaType;
import jakarta.ws.rs.core.MultivaluedMap;
import jakarta.ws.rs.ext.MessageBodyReader;
import jakarta.ws.rs.ext.MessageBodyWriter;
import jakarta.ws.rs.ext.Provider;
@Provider
public class CheeseBodyHandler implements MessageBodyReader<Cheese>,
MessageBodyWriter<Cheese> {
@Override
public boolean isWriteable(Class<?> type, Type genericType,
Annotation[] annotations, MediaType mediaType) {
return type == Cheese.class;
}
@Override
public void writeTo(Cheese t, Class<?> type, Type genericType,
Annotation[] annotations, MediaType mediaType,
MultivaluedMap<String, Object> httpHeaders,
OutputStream entityStream)
throws IOException, WebApplicationException {
entityStream.write(("[CheeseV1]" + t.name)
.getBytes(StandardCharsets.UTF_8));
}
@Override
public boolean isReadable(Class<?> type, Type genericType,
Annotation[] annotations, MediaType mediaType) {
return type == Cheese.class;
}
@Override
public Cheese readFrom(Class<Cheese> type, Type genericType,
Annotation[] annotations, MediaType mediaType,
MultivaluedMap<String, String> httpHeaders,
InputStream entityStream)
throws IOException, WebApplicationException {
String body = new String(entityStream.readAllBytes(), StandardCharsets.UTF_8);
if(body.startsWith("[CheeseV1]"))
return new Cheese(body.substring(11));
throw new IOException("Invalid cheese: " + body);
}
}
If you want to get the most performance our of your writer, you can extend the
ServerMessageBodyWriter
instead of MessageBodyWriter
where you will be able to use less reflection and bypass the blocking IO layer:
package org.acme.rest;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.lang.annotation.Annotation;
import java.lang.reflect.Type;
import java.nio.charset.StandardCharsets;
import jakarta.ws.rs.WebApplicationException;
import jakarta.ws.rs.core.MediaType;
import jakarta.ws.rs.core.MultivaluedMap;
import jakarta.ws.rs.ext.MessageBodyReader;
import jakarta.ws.rs.ext.Provider;
import org.jboss.resteasy.reactive.server.spi.ResteasyReactiveResourceInfo;
import org.jboss.resteasy.reactive.server.spi.ServerMessageBodyWriter;
import org.jboss.resteasy.reactive.server.spi.ServerRequestContext;
@Provider
public class CheeseBodyHandler implements MessageBodyReader<Cheese>,
ServerMessageBodyWriter<Cheese> {
// …
@Override
public boolean isWriteable(Class<?> type, ResteasyReactiveResourceInfo target,
MediaType mediaType) {
return type == Cheese.class;
}
@Override
public void writeResponse(Cheese t, ServerRequestContext context)
throws WebApplicationException, IOException {
context.serverResponse().end("[CheeseV1]" + t.name);
}
}
Just as you can intercept requests and responses, you can also intercept readers and writers, by
extending the ReaderInterceptor
or
WriterInterceptor
on a class annotated with
@Provider
.
If we look at this endpoint:
package org.acme.rest;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.PUT;
import jakarta.ws.rs.Path;
@Path("cheese")
public class Endpoint {
@GET
public String sayCheese() {
return "Cheeeeeese";
}
@PUT
public void addCheese(String cheese) {
System.err.println("Received a new cheese: " + cheese);
}
}
We can add reader and writer interceptors like this:
package org.acme.rest;
import java.io.IOException;
import jakarta.ws.rs.WebApplicationException;
import jakarta.ws.rs.ext.Provider;
import jakarta.ws.rs.ext.ReaderInterceptor;
import jakarta.ws.rs.ext.ReaderInterceptorContext;
import jakarta.ws.rs.ext.WriterInterceptor;
import jakarta.ws.rs.ext.WriterInterceptorContext;
@Provider
public class CheeseIOInterceptor implements ReaderInterceptor, WriterInterceptor {
@Override
public void aroundWriteTo(WriterInterceptorContext context)
throws IOException, WebApplicationException {
System.err.println("Before writing " + context.getEntity());
context.proceed();
System.err.println("After writing " + context.getEntity());
}
@Override
public Object aroundReadFrom(ReaderInterceptorContext context)
throws IOException, WebApplicationException {
System.err.println("Before reading " + context.getGenericType());
Object entity = context.proceed();
System.err.println("After reading " + entity);
return entity;
}
}
In Quarkus, the RESTEasy Reactive extension and the REST Client Reactive extension share the same infrastructure. One important consequence of this consideration is that they share the same list of providers (in the Jakarta REST meaning of the word).
For instance, if you declare a WriterInterceptor
, it will by default intercept both the servers calls and the client calls,
which might not be the desired behavior.
However, you can change this default behavior and constrain a provider to:
-
only consider server calls by adding the
@ConstrainedTo(RuntimeType.SERVER)
annotation to your provider; -
only consider client calls by adding the
@ConstrainedTo(RuntimeType.CLIENT)
annotation to your provider.
All Request Parameters can be declared as String
,
but also any of the following types:
-
Types for which a
ParamConverter
is available via a registeredParamConverterProvider
. -
Primitive types.
-
Types that have a constructor that accepts a single
String
argument. -
Types that have a static method named
valueOf
orfromString
with a singleString
argument that return an instance of the type. If both methods are present thenvalueOf
will be used unless the type is anenum
in which casefromString
will be used. -
List<T>
,Set<T>
, orSortedSet<T>
, whereT
satisfies any above criterion.
The following example illustrates all those possibilities:
package org.acme.rest;
import java.lang.annotation.Annotation;
import java.lang.reflect.Type;
import java.util.List;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.ext.ParamConverter;
import jakarta.ws.rs.ext.ParamConverterProvider;
import jakarta.ws.rs.ext.Provider;
import org.jboss.resteasy.reactive.RestQuery;
@Provider
class MyConverterProvider implements ParamConverterProvider {
@Override
public <T> ParamConverter<T> getConverter(Class<T> rawType, Type genericType,
Annotation[] annotations) {
// declare a converter for this type
if(rawType == Converter.class) {
return (ParamConverter<T>) new MyConverter();
}
return null;
}
}
// this is my custom converter
class MyConverter implements ParamConverter<Converter> {
@Override
public Converter fromString(String value) {
return new Converter(value);
}
@Override
public String toString(Converter value) {
return value.value;
}
}
// this uses a converter
class Converter {
String value;
Converter(String value) {
this.value = value;
}
}
class Constructor {
String value;
// this will use the constructor
public Constructor(String value) {
this.value = value;
}
}
class ValueOf {
String value;
private ValueOf(String value) {
this.value = value;
}
// this will use the valueOf method
public static ValueOf valueOf(String value) {
return new ValueOf(value);
}
}
@Path("hello")
public class Endpoint {
@Path("{converter}/{constructor}/{primitive}/{valueOf}")
@GET
public String conversions(Converter converter, Constructor constructor,
int primitive, ValueOf valueOf,
@RestQuery List<Constructor> list) {
return converter + "/" + constructor + "/" + primitive
+ "/" + valueOf + "/" + list;
}
}
RESTEasy Reactive supports the use of the implementations of java.time.Temporal
(like java.time.LocalDateTime
) as query, path, or form params.
Furthermore, it provides the @org.jboss.resteasy.reactive.DateFormat
annotation, which can be used to set a custom expected pattern.
Otherwise, the JDK’s default format for each type is used implicitly.
HTTP allows requests to be conditional, based on a number of conditions, such as:
-
Date of last resource modification
-
A resource tag, similar to a hash code of the resource to designate its state or version
Let’s see how you can do conditional request validation using the
Request
context object:
package org.acme.rest;
import java.time.Instant;
import java.time.temporal.ChronoUnit;
import java.time.temporal.TemporalUnit;
import java.util.Date;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.PUT;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.core.EntityTag;
import jakarta.ws.rs.core.Request;
import jakarta.ws.rs.core.Response;
import jakarta.ws.rs.core.Response.ResponseBuilder;
@Path("conditional")
public class Endpoint {
// It's important to keep our date on seconds because that's how it's sent to the
// user in the Last-Modified header
private Date date = Date.from(Instant.now().truncatedTo(ChronoUnit.SECONDS));
private int version = 1;
private EntityTag tag = new EntityTag("v1");
private String resource = "Some resource";
@GET
public Response get(Request request) {
// first evaluate preconditions
ResponseBuilder conditionalResponse = request.evaluatePreconditions(date, tag);
if(conditionalResponse != null)
return conditionalResponse.build();
// preconditions are OK
return Response.ok(resource)
.lastModified(date)
.tag(tag)
.build();
}
@PUT
public Response put(Request request, String body) {
// first evaluate preconditions
ResponseBuilder conditionalResponse = request.evaluatePreconditions(date, tag);
if(conditionalResponse != null)
return conditionalResponse.build();
// preconditions are OK, we can update our resource
resource = body;
date = Date.from(Instant.now().truncatedTo(ChronoUnit.SECONDS));
version++;
tag = new EntityTag("v" + version);
return Response.ok(resource)
.lastModified(date)
.tag(tag)
.build();
}
}
When we call GET /conditional
the first time, we will get this response:
HTTP/1.1 200 OK
Content-Type: text/plain;charset=UTF-8
ETag: "v1"
Last-Modified: Wed, 09 Dec 2020 16:10:19 GMT
Content-Length: 13
Some resource
So now if we want to check if we need to fetch a new version, we can make the following request:
GET /conditional HTTP/1.1
Host: localhost:8080
If-Modified-Since: Wed, 09 Dec 2020 16:10:19 GMT
And we would get the following response:
HTTP/1.1 304 Not Modified
Because the resource has not been modified since that date, this saves on sending the resource but can also help your users detect the concurrent modification.
For example, one client wants to update the resource, but another user has modified it since.
You can follow the previous GET
request with this update:
PUT /conditional HTTP/1.1
Host: localhost:8080
If-Unmodified-Since: Wed, 09 Dec 2020 16:25:43 GMT
If-Match: v1
Content-Length: 8
Content-Type: text/plain
newstuff
And if some other user has modified the resource between your GET
and your PUT
you would
get this answer back:
HTTP/1.1 412 Precondition Failed
ETag: "v2"
Content-Length: 0
One of the main ideas of REST (and HTTP) is that your resource is independent of its representation, and that both the client and server are free to represent their resources in as many media types as they want. This allows the server to declare support for multiple representations and let the client declare which ones it supports and get served something appropriate.
The following endpoint supports serving cheese in plain text or JSON:
package org.acme.rest;
import jakarta.ws.rs.Consumes;
import jakarta.ws.rs.GET;
import jakarta.ws.rs.PUT;
import jakarta.ws.rs.Path;
import jakarta.ws.rs.Produces;
import jakarta.ws.rs.core.MediaType;
import com.fasterxml.jackson.annotation.JsonCreator;
class Cheese {
public String name;
@JsonCreator
public Cheese(String name) {
this.name = name;
}
@Override
public String toString() {
return "Cheese: " + name;
}
}
@Path("negotiated")
public class Endpoint {
@Produces({MediaType.APPLICATION_JSON, MediaType.TEXT_PLAIN})
@GET
public Cheese get() {
return new Cheese("Morbier");
}
@Consumes(MediaType.TEXT_PLAIN)
@PUT
public Cheese putString(String cheese) {
return new Cheese(cheese);
}
@Consumes(MediaType.APPLICATION_JSON)
@PUT
public Cheese putJson(Cheese cheese) {
return cheese;
}
}
The user will be able to select which representation it gets with the
Accept
header, in the case of JSON:
> GET /negotiated HTTP/1.1
> Host: localhost:8080
> Accept: application/json
< HTTP/1.1 200 OK
< Content-Type: application/json
< Content-Length: 18
<
< {"name":"Morbier"}
And for text:
> GET /negotiated HTTP/1.1
> Host: localhost:8080
> Accept: text/plain
>
< HTTP/1.1 200 OK
< Content-Type: text/plain
< Content-Length: 15
<
< Cheese: Morbier
Similarly, you can PUT
two different representations. JSON:
> PUT /negotiated HTTP/1.1
> Host: localhost:8080
> Content-Type: application/json
> Content-Length: 16
>
> {"name": "brie"}
< HTTP/1.1 200 OK
< Content-Type: application/json;charset=UTF-8
< Content-Length: 15
<
< {"name":"brie"}
Or plain text:
> PUT /negotiated HTTP/1.1
> Host: localhost:8080
> Content-Type: text/plain
> Content-Length: 9
>
> roquefort
< HTTP/1.1 200 OK
< Content-Type: application/json;charset=UTF-8
< Content-Length: 20
<
< {"name":"roquefort"}
The body of an HTTP response is not compressed by default.
You can enable the HTTP compression support by means of quarkus.http.enable-compression=true
.
If compression support is enabled then the response body is compressed if:
-
the resource method is annotated with
@io.quarkus.vertx.http.Compressed
, or -
the
Content-Type
header is set and the value is a compressed media type as configured viaquarkus.http.compress-media-types
.
The response body is never compressed if:
-
the resource method is annotated with
@io.quarkus.vertx.http.Uncompressed
, or -
the
Content-Type
header is not set.
Tip
|
By default, the following list of media types is compressed: text/html , text/plain , text/xml , text/css , text/javascript and application/javascript .
|
Note
|
If the client does not support HTTP compression then the response body is not compressed. |
Quarkus enables the inclusion or exclusion of Jakarta REST Resources, Providers and Features directly thanks to build time conditions in the same that it does for CDI beans.
Thus, the various Jakarta REST classes can be annotated with profile conditions (@io.quarkus.arc.profile.IfBuildProfile
or @io.quarkus.arc.profile.UnlessBuildProfile
) and/or with property conditions (io.quarkus.arc.properties.IfBuildProperty
or io.quarkus.arc.properties.UnlessBuildProperty
) to indicate to Quarkus at build time under which conditions these Jakarta REST classes should be included.
In the following example, Quarkus includes the ResourceForApp1Only
Resource class if and only if the build profile app1
has been enabled.
@IfBuildProfile("app1")
public class ResourceForApp1Only {
@GET
@Path("sayHello")
public String sayHello() {
return "hello";
}
}
Please note that if a Jakarta REST Application has been detected and the method getClasses()
and/or getSingletons()
has/have been overridden, Quarkus will ignore the build time conditions and consider only what has been defined in the Jakarta REST Application.
Quarkus can also conditionally disable Jakarta REST Resources based on the value of runtime properties using the @io.quarkus.resteasy.reactive.server.EndpointDisabled
annotation.
In the following example, Quarkus will exclude RuntimeResource
at runtime if the application has some.property
configured to "disable"
.
@EndpointDisabled(name = "some.property", stringValue = "disable")
public class RuntimeResource {
@GET
@Path("sayHello")
public String sayHello() {
return "hello";
}
}
In addition to the Server side, RESTEasy Reactive comes with a new MicroProfile REST Client implementation that is non-blocking at its core.
Please note that the quarkus-rest-client
extension may not be used with RESTEasy Reactive, use quarkus-rest-client-reactive
instead.
See the REST Client Reactive Guide for more information about the reactive REST client.