Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Sep 16 16:56
    roman-garcia commented #1608
  • Sep 16 14:22
    odrotbohm labeled #1636
  • Sep 16 14:22
    odrotbohm unlabeled #1636
  • Sep 16 14:22
    odrotbohm commented #1636
  • Sep 16 14:17
    roman-garcia commented #1608
  • Sep 16 13:41
    reda-alaoui closed #1636
  • Sep 16 13:41
    reda-alaoui commented #1636
  • Sep 16 13:37
    reda-alaoui edited #1636
  • Sep 16 13:36
    odrotbohm labeled #1636
  • Sep 16 13:36
    odrotbohm labeled #1636
  • Sep 16 13:36
    odrotbohm labeled #1636
  • Sep 16 13:36
    odrotbohm assigned #1636
  • Sep 16 13:36
    odrotbohm commented #1636
  • Sep 16 13:35
    reda-alaoui edited #1636
  • Sep 16 13:34
    reda-alaoui edited #1636
  • Sep 16 13:34
    reda-alaoui edited #1636
  • Sep 16 13:34
    reda-alaoui edited #1636
  • Sep 16 13:33
    reda-alaoui opened #1636
  • Sep 16 13:18

    odrotbohm on release

    #1595 - Continue development on… #1598 - Fully prepare handler m… #1402 - Properly create Propert… and 19 more (compare)

  • Sep 16 13:08
    odrotbohm closed #1635
Greg L. Turnquist
@gregturn
But nothing as direct as ObjectMapper settings.
If we were to add a hook, I'd imagine it being there.
To me, a ObjectMapperCustomizer sort of lambda function added to HalConfig would be the simplest way. But I'd also question granting users that much access to the internals.
Kai Toedter
@toedter
Thx Greg, I think that many users want to use additional Jackson modules, like the JavaTimeModule. I would prefer the possibility that a user could add some configuration, like how to serialize LocalDateTime or Instant on a Spring HATEOAS level, and then apply this to all media types.
Greg L. Turnquist
@gregturn

Something like...

public interface ObjectMapperCustomizer extends Consumer<ObjectMapper> {
}

You provide new HalConfig().customizeObjectMapper(objectMapper -> { /* insert customizations here */}), and the serializer, after getting/creating the ObjectMapper, could then invoke halConfig.objectMapperCustomizer.accept(theObjectMapper).

Yeah?

I could see this applied to the others as well.

I've opened spring-projects/spring-hateoas#1382 to track this discussion.
Greg L. Turnquist
@gregturn
Greg L. Turnquist
@gregturn
Woot!
Patrick Hahn
@xorph
Has someone an example how to build page links for PagedModel with a pageable in a reactive application using the WebFluxLinkBuilder? The resulting link does not contain paging parameters.
@GetMapping(produces = [MediaTypes.HAL_JSON_VALUE])
fun findAllPaged(
        @PageableDefault(
                size = DEFAULT_PAGE_SIZE,
                page = FIRST_PAGE,
                sort = [DEFAULT_SORT_BY],
                direction = Sort.Direction.ASC
        )
        pageable: Pageable
    ): Mono<PagedModel<Foo>>
linkTo(methodOn(FooController::class.java).findAllPaged(pageable))
                        .withSelfRel()
                        .toMono()
Greg L. Turnquist
@gregturn
To back up a bit, "pageable" is still a fuzzy concept in reactive. If you pass in a Pageable to a reactive repostiory, you get back a Flux, not a page. So essentially, you have to build it up yourself.
We need to hammer out with the Reactor team on what makes for good, reactive "paged" systems.
Patrick Hahn
@xorph

Right now we added the paging query params via a kotlin extension on Link. So we can do something like this:

linkTo(methodOn(FooController::class.java).findAllPaged(pageable))
                        .withSelfRel()
                        .toMono()
                        .map { it.toPagedLink(pageable) }

Under the hood we had to transform the link to URI, extend with paging query params and create the link with the rel again. Is there a way to directly add query params to the Link?

Greg L. Turnquist
@gregturn
It's not the most common use case, but you can do Link.of(UriTemplate, LinkRelation). UriTemplate has helper methods to build up all kinds of stuff.
Patrick Hahn
@xorph
I will have a look at it. Thanks!
DonHamzovic
@DonHamzovic
This message was deleted

I'm developing a microservice using SPRING DATA REST where I need to sub-resources to be accessible only thru main resource like the following structure :

http://localhost:8080/posts
http://localhost:8080/posts/1/comments

and block direct access to sub-resource directly like this http://localhost:8080/comments/*.

where comments must be accessible only thru related poste , I did the following in my repository :

@RepositoryRestResource(collectionResourceRel = "posts", path = "posts")
public interface PostRepository extends PagingAndSortingRepository<Post, Long> {
   ...
}

for comments :

 @RepositoryRestResource(collectionResourceRel = "comments", path = "comments")
    public interface CommentRepository extends PagingAndSortingRepository<Comment, Long> {
       ...
    }

now by default SPRING DATA REST returns the following results when i goto to : http://localhost:8080

{
  "id": 1,
  "content": "some post text .........................",
  "author":"abc",

  "_links": {
    "self": {
      "href": "http://localhost:8080/posts/1"
    },
    "attributes": {
      "href": "http://localhost:8080/posts/1/comments"
    }
  }
}

now if i want to post a comment i need to do the following :

http://{server:port}/comment METHOD:POST

```
{"author":"abc","content":"PQROHSFHFSHOFSHOSF", "post":"http://{server:port}/post/1"}

```
but what i need to achienve is to POST to an url like this http://{server:port}/posts/1/comment where the POST is the root resource NOT like the previous path http://{server:port}/comment

i now that this is possible if a create a custom comment @controller but i want to use the already building features of SPRING DATA REST and Hateoas support .

Greg L. Turnquist
@gregturn
Why do you need the CommentRepository? SDR, by default, tries to create root aggregates for every repository you define. If, as you are saying, don't need Comments to be a root aggregate, why not ditch it?
DonHamzovic
@DonHamzovic
no i need it if i don't add CommentRepository SPRING DATA REST will not create the link for the sub-resource
Greg L. Turnquist
@gregturn
@DonHamzovic I'm not sure how this fits, but have you tried applying @RestResource(exported = false) on CommenRepository? That should stop exporting it as an aggregate root. But I don't know if that ALSO stops making it available as a subresource.
1 reply
Greg L. Turnquist
@gregturn
@xorph Have you seen @toedter 's deck? It shows up the UriTemplate APIs => https://speakerdeck.com/toedter/hello-rest-api-what-can-i-do-next-c47a5e7c-4a1f-491a-8337-bdc28c382f0d?slide=28
Ingo Griebsch
@ingogriebsch
A collection of examples which showcase how to use the features of Spring HATEOAS Siren is available now! Included are topics like setup, serialization, deserialization and internationalization.
Every feedback is very welcome! :)
https://github.com/ingogriebsch/spring-hateoas-siren-samples
Ingo Griebsch
@ingogriebsch
I’m pleased to announce the release of Spring HATEOAS Siren 1.0.0-M3! The library is accessible through Maven Central.
Source: https://github.com/ingogriebsch/spring-hateoas-siren
Documentation: https://ingogriebsch.github.io/spring-hateoas-siren
You can expect the release of the first version soon! :)
Ingo Griebsch
@ingogriebsch
I’m really happy to announce the first final release of Spring HATEOAS Siren!
Check it out, version 1.0.0 is accessible through Maven Central!
Source: https://github.com/ingogriebsch/spring-hateoas-siren
Documentation: https://ingogriebsch.github.io/spring-hateoas-siren
Examples: https://github.com/ingogriebsch/spring-hateoas-siren-samples
Every feedback is very welcome! :)
ctwoolsey
@ctwoolsey

I'm struggling to read data from a webClient query.
I've configured SpringConfig:

@EnableHypermediaSupport(type = EnableHypermediaSupport.HypermediaType.HAL)
class SpringConfig {
@Bean
WebClientCustomizer webClientCustomizer(HypermediaWebClientConfigurer configurer) {
return {webClientBuilder ->
configurer.registerHypermediaTypes(webClientBuilder)
}
}
}

Then in my repository I have:

@Repository ('PersonRepository')
class PersonRepository {
private final WebClient webClient

PersonRepository(final WebClient.Builder webClientBuilder) {
this.webClient = webClientBuilder.build()
}

List<person> getPeople() {
CollectionModel<person> personCollectionModel = this.webClient.get().uri('localhost:8080/api/people')
.retrieve()
.bodyToMono(new TypeReferences.CollectionModelType<person>())
.block()

return personCollectionModel.content
}

The problem is that "personCollectionModel" is empty. It is like the HAL format is unable to be read into the structure.

This is the data that is returned from localhost:8080/api/people

{
"_embedded": {
"people": [
{
"id": 2,
"firstName": "Mark",
"lastName": "Hamil"
}
]
}
}

Any light you can shine on why this isn't returning data as expected would be appreciated. Thank you.

ctwoolsey
@ctwoolsey
@gregturn In your article: https://spring.io/blog/2020/04/22/spring-hateoas-brings-you-new-ways-to-configure-clients, do these new ways to configure allow for reading HAL json with "_embedded" tags?
Greg L. Turnquist
@gregturn
They should.
        client.get().uri("http://localhost/employees").accept(HAL_JSON) //
                .exchange() //
                .expectStatus().isOk() //
                .expectBody(new TypeReferences.CollectionModelType<EntityModel<Employee>>() {}) // <3>
                .consumeWith(result -> {
                    CollectionModel<EntityModel<Employee>> model = result.getResponseBody(); // <4>

                    // Assert against the hypermedia model.
                    assertThat(model.getRequiredLink(IanaLinkRelations.SELF)).isEqualTo(Link.of("http://localhost/employees"));
                    assertThat(model.getContent()).hasSize(2);
                });
Greg L. Turnquist
@gregturn
The inner type needs to be EntityModel<Employee> and you also need the {} as well.
ctwoolsey
@ctwoolsey

@gregturn Thank you for the pointer. I am receiving a status of 406 Not accepted when I use accept(HAL_JSON). Here is the modified HAL output:

    "_embedded": {
        "people": [
            {
                "id": 2,
                "firstName": "Mark",
                "lastName": "Hamil",
                "_links": {
                    "self": {
                        "href": "http://localhost:8080/api/people/2"
                    }
                }
            }
        ]
    },
    "_links": {
        "self": {
            "href": "http://localhost:8080/api/people"
        }
    }
}

And here is my code to receive it:

Mono<CollectionModel<EntityModel<Person>>> personCollectionModel = this.webClient.get().uri('localhost:8080/api/people')
                  .accept(MediaTypes.HAL_JSON)
                  .exchange()
                  .flatMap { response ->
                       if (response.statusCode().isError()) {
                         System.out.println('Error---------------------------------')
                       } else {
                            response.bodyToMono(new ParameterizedTypeReference<CollectionModel<EntityModel<Person>>>() {})
                       }
                  }

I'm not sure why the HAL format is not being recognized as valid HAL.

ctwoolsey
@ctwoolsey
Thank you for your help. I figured it out, this is now working, I needed to set produces = "application/hal+json"
Greg L. Turnquist
@gregturn
If the Content-Type header is something else, you can always drop the Accept request header. If you are getting HAL, you should be able to consume it. The rest of simply HTTP negotation.
It's possible, if you're using Boot, that you're getting served application/json. Don't remember it NOT serving up application/hal+json when requested, though.
Ravish Rathod
@ravishrathod
Hello, I recently upgrade my app from a very old version of spring-boot (and hateoas) to the latest boot and hateoas v1.1.2.RELEASE. After this I am seeing a spike in memory usage. Heap dump points to WebHandler.AFFORDANCES_CACHE consuming most of memory. Looking at WebHandler it seems that the cache is never cleared and is unbounded... has anyone faced this issue ? Thanks
Greg L. Turnquist
@gregturn
@ravishrathod The last fix that involved that cache was in 1.0. If you'll open a new ticket and provide reproducible details, we can certainly investigate.
Ravish Rathod
@ravishrathod
Thanks @gregturn , I will add a link to git project
Here is the link to a simple project which exposes a /users/{id}. If we call this api with enough distinct id values, it causes AFFORDANCES_CACHE to fill up. https://github.com/ravishrathod/spring-hateoas-memory
Ravish Rathod
@ravishrathod
ctruzzi
@ctruzzi
We have pre-existing contracts, is it possible to change the mixin of links from _links to just links?
Greg L. Turnquist
@gregturn

@ctruzzi Not without violating the HAL spec. In essence, you can write your own media type => https://docs.spring.io/spring-hateoas/docs/1.2.0/reference/html/#mediatypes.custom

You can also look at the community-provided media types as templates:

Ingo Griebsch
@ingogriebsch
Hello @gregturn it looks like the Spring HATEOAS project page is not up to date. The page points to version 1.1.2 and 1.2.0. It does not point to version 1.1.3 and 1.2.2. I found corresponding milestones on GitHub and artifacts on Maven Central for these versions. :)
Greg L. Turnquist
@gregturn
I believe Ollie recently put out a new release. We'll get it updated shortly.
Ingo Griebsch
@ingogriebsch
:thumbsup:
Okay, I only wanted to make sure I can use them to update my project to these versions.
Greg L. Turnquist
@gregturn
@ingogriebsch I updated all the versions.
Ingo Griebsch
@ingogriebsch
:thumbsup:
Ingo Griebsch
@ingogriebsch
Hello @gregturn, I would like to ask if there is somewhere explained how the Spring HATEOAS library decides which RepresentationModelProcessor is executed on which specific RepresentationModel? I just struggle a little bit with building complex RepresentationModel's (with a builder) and therefore not having a concrete type I could use to define a processor for.
1 reply
koehnj
@koehnj
Dear Spring HATEOAS team, i just noticed that you've recently added support for the "type" attribute for HAL-FORMS properties. Are there any plans to support the remaining additional attributes as well? Placeholder, minLength and maxLength are especially useful.
2 replies
Marcel Widmer
@marzelwidmer

Hello what is the best practice wo handle in the links with combination of DDD with validation on it... If maybe make a abstraction again on REST entpoint or not... or if there a better way to create the _link

eg. when I use a valid object from a customer. afford<CustomerResource> { methodOn(CustomerResource::class.java).add(customer = customer) }

val customer = Customer(id= null, firstName = FirstName("firstName"), lastName = LastName("lastName"))

    @GetMapping(value = ["/customers/{id}"], produces = [MediaTypes.HAL_JSON_VALUE])
    fun one(@PathVariable id: String): EntityModel<Customer> {
        // TODO: 09.02.21  Validation problems
        val customer = Customer(id= Any().toString(), firstName = FirstName(Any().toString()), lastName = LastName(Any().toString()))
        val create: Affordance = afford<CustomerResource> { methodOn(CustomerResource::class.java).add(customer = customer) }

or should I create the Domain objects eg on the service level... sound not nice for me... because I pass already the controller
maybe I miss understood something....

@odrotbohm will be nice update on yout spring-restbucks ;) how handle everything together..

Greg L. Turnquist
@gregturn

@marzelwidmer I can understand having a DTO that targets the serialized JSON, and having it carry validation rules embedded in either the constructors or the setters. This is the format used to "present" your API. After validating things in the controller, I would then transform into the internal type used at the service level. And I would do that transformation in the controller. The service doesn't have to be aware of this API-level of things.

If the validation rules you have in mind apply at the service level, put your validation rules there. Otherwise, keep them in the DTO. Then, when you transform stuff in the controller, you can have suitable error handling to report to the user of the API what went wrong.

Oliver Drotbohm
@odrotbohm
I don't quite follow. I.e. I don't actually get what the goal is.