-
Notifications
You must be signed in to change notification settings - Fork 77
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
POC for Polymorphism #890
POC for Polymorphism #890
Conversation
✅ Deploy Preview for springwolf-ui ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
springwolf-examples/springwolf-kafka-example/src/test/resources/asyncapi.yaml
Outdated
Show resolved
Hide resolved
...java/io/github/springwolf/core/asyncapi/scanners/common/payload/internal/PayloadService.java
Outdated
Show resolved
Hide resolved
.../main/java/io/github/springwolf/core/asyncapi/scanners/common/payload/NamedSchemaObject.java
Outdated
Show resolved
Hide resolved
...java/io/github/springwolf/core/asyncapi/components/examples/walkers/DefaultSchemaWalker.java
Outdated
Show resolved
Hide resolved
...in/java/io/github/springwolf/core/asyncapi/schemas/converters/SchemaTitleModelConverter.java
Outdated
Show resolved
Hide resolved
...java/io/github/springwolf/core/asyncapi/scanners/common/payload/internal/PayloadService.java
Show resolved
Hide resolved
...ringwolf/core/asyncapi/scanners/channels/annotations/AsyncAnnotationChannelsScannerTest.java
Show resolved
Hide resolved
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
TypeToClassConverter#extractClass is only used by the cloudstream plugin now.
A next step (possibly outside of the PR) can be to apply the same Class -> Type change to the cloudstream plugin and so that TypeToClassConverter can be fully removed, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yes, that was what I was thinking as well
springwolf-examples/springwolf-kafka-example/src/test/resources/asyncapi.yaml
Show resolved
Hide resolved
springwolf-core/src/test/resources/schemas/xml/annotation-definitions-xml.json
Show resolved
Hide resolved
@@ -227,7 +227,7 @@ | |||
}, | |||
"name": "io.github.springwolf.examples.amqp.dtos.AnotherPayloadDto", | |||
"title": "AnotherPayloadDto", | |||
"description": "Another payload model", | |||
"description": "", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Were you able to find out, why the description disappears?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The description that is shown is the description of the schema, so the logical place to show this description is on the Schema-component, as it is already.
The description that is on that place, should be the one of the @asyncmessage, in my opinion.
The reason why the description is not anymore shown, is that Swagger creates now the reference to the schema-object, whereas before it was a hardcoded reference by Springwolf itself. Due to removing that hardcoded reference, we do not receive anymore the full schema, but an already referenced schema-object on the payload, without a description.
If this is a blocking matter, what could be done is:
- Either, fetching the schema from the swagger-reference and fetching the description via that way.
- Either, setting the description on the reference-schema with a custom modelConverter.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good point regarding the location of the description and it seems that the description does not belong here.
Springwolf has two ways to specify an optional description via the Swagger Schema annotation (belongs to the schema) and via the AsyncOperation annotation (belongs to the operation).
Can the description be null in that case instead of an empty string?
(I am currently unsure, whether the ui selects the description already from the correct location)
EDIT: Just compared the PR preview https://deploy-preview-890--springwolf-ui.netlify.app/springwolf-ui/asyncapi-ui.html#channel-string-topic and the current state https://demo.springwolf.dev/springwolf-ui/asyncapi-ui.html#channel-string-topic for the string topic and the ui seems to handle channel Operation description
and Message description
correctly.
@@ -25,8 +25,8 @@ | |||
"another-topic": { | |||
"address": "another-topic", | |||
"messages": { | |||
"io.github.springwolf.examples.kafka.dtos.AnotherPayloadDto": { | |||
"$ref": "#/components/messages/io.github.springwolf.examples.kafka.dtos.AnotherPayloadDto" | |||
"java.util.ListIo.github.springwolf.examples.kafka.dtos.AnotherPayloadDto": { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The case of List
in combination with batch listeners may be debatable.
I've added a ConsumerRecord listener to the kafka example. Using the extractable-classes, do you see an easy way to auto-extract the actual class?
Another modelConverter?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
hmm, I must admit, that indeed the consumerRecord does not get extracted the way it should. A "modelConverter" is always a good idea to adapt the schema to the way we would like. I think this is a same case as with the MonetaryExample, where Swagger does not see the defined fields of the class
I haven't tried the example on the current code of Springwolf, how it generates. I suppose quite the same.
An idea that comes to my mind is that maybe for each of the different "support-bindings", some custom modelconverters could be added on the classpath. So for example this case: The class ConsumerRecord
belongs to the Kafka framework, so only when that binding is selected, the support for consumerRecord could be added with a custom modelconverter. You could define that modelconverter bean on the SpringwolfKafkaBindingAutoConfiguration
.
I suppose each framework has its special quircks, so adding custom modelconverters for each of them to the binding-autoconfiguration seems like the way to go.
I don't know what you think.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I added some code as mentioned in #891 (comment)
If you don't feel strongly about it, I would like to minimize breaking changes and continue to extract List
Schema payloadSchema = resolvedSchema.schema; | ||
preProcessSchemas(payloadSchema, preProcessSchemas, type); | ||
HashMap<String, Schema> postProcessSchemas = new HashMap<>(preProcessSchemas); | ||
postProcessSchema(preProcessSchemas, postProcessSchemas, actualContentType); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am probably missing something:
Is this part of the duplication, you mentioned? It seems that the same schemas are post processed again and/or the resolvedSchema is not post-processed.
Side-note: I didnt know about ResolvedSchema. Great improvement to use it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The duplication is more in the methods, I simply duplicated, like:
postProcessSchema
processAsyncApiPayloadAnnotation
processSchemaAnnotation
processCommonModelConverters
Those 4 methods are twice in the swaggerSchemaService. Once for the old way of defining schemas (still in use for headers) and once for the new way of defining schemas (for defining the payload).
So, as part of a next refactoring the header-generation could be adapted to make also use of "resolvedSchema" and then those duplicate methods can go away.
extracted from springwolf#890 Co-authored-by: David Beck <[email protected]>
extracted from springwolf#890 Co-authored-by: David Beck <[email protected]>
extracted from #890 Co-authored-by: David Beck <[email protected]>
extracted from springwolf#890 Co-authored-by: David Beck <[email protected]>
* refactor(core): rename NamedSchemaObject to PayloadSchemaObject extracted from #890 Co-authored-by: David Beck <[email protected]> * refactor(core): use SchemaType constants --------- Co-authored-by: David Beck <[email protected]>
Merge from Timon's repo
also add to schemas section to allow publishing
POC for Polymorphism