Implement API for sample management #1063
Conversation
| asyncProjectService.getSamplePreviews(projectId.value(), experimentId.value(), | ||
| query.getOffset(), query.getLimit(), List.copyOf(sortOrders), filter).collectList() | ||
| .doOnError(RequestFailedException.class, this::handleRequestFailed) | ||
| .block()).stream(); |
There was a problem hiding this comment.
How about https://projectreactor.io/docs/core/release/api/reactor/core/publisher/Flux.html#toStream-- instead?
Maybe you do not even need to block then (but I don't know this).
https://projectreactor.io/docs/core/release/reference/apdx-operatorChoice.html#which.blocking
Seems to me that a stream is returned (which is expected by vaadin) so the whole collection into a list and streaming afterwards might not be needed.
...c/main/java/life/qbic/datamanager/views/projects/project/samples/SampleDetailsComponent.java
Outdated
Show resolved
Hide resolved
| .getValue(); | ||
| var result = asyncProjectService | ||
| .getSamples(context.projectId().orElseThrow().value(), experiment.experimentId().value()) | ||
| .collectList().block(); |
There was a problem hiding this comment.
Instead of waiting for all elements to be there you could only wait for the first element as it is your concern whether the list is empty.
https://projectreactor.io/docs/core/release/reference/apdx-operatorChoice.html#which.blocking
https://projectreactor.io/docs/core/release/api/reactor/core/publisher/Flux.html#blockFirst-java.time.Duration-
There was a problem hiding this comment.
i think that cheats the way Flux is working: we don't know how many samples will be send through the Flux. So maybe it is better to use a convenience method in the Service API instead?
There was a problem hiding this comment.
Why would this cheat the way flux is working?
The flux publishes elements and instead of waiting for the flux to complete, you can simply wait for the first element. No need for the flux to continue publishing afterwards as you do not want to use the result.
There was a problem hiding this comment.
Of course you could think about adding a method Mono to determine whether there are any samples in an experiment. Depending on the way the flux is created this would minimize the work the system does. But I think both are legitimate ways to handle the situation. And both are better than to wait for the flux completion.
|
sven1103
deleted the
feature/#1062-implemtation-of-sample-management-api
branch
Provides the async service implementation for the sample management related requests.