CODEWITHSUNDEEP
javaspringrestspring-integration
Stefano Veloccia
Stefano Veloccia

Reputation: 190

Read and download from a paginated REST-Services with spring integration

Currently I am working on a Spring Integration application which has a following scenario:

  1. An int-http:outbound-gateway read from a REST-Services a list of paginated elements: about in
  2. Each page content splitted and stored in a folder to be processed later by a spring batch job.

I'm quite new with spring-integration and I don't know if it's possibile to create a kind of loop with `int-http:outbound-gateway' to read all pages until the last one.

We're talking about 66254 elements splitted in 2651 pages. What I'm looking for is a best practice to read and download all pages and collecting data without have any memory issue.

Any suggestion will be appreciated

Thanks

Upvotes: 2

Views: 987

Answers (2)

Chris
Chris

Reputation: 4648

I did something similar using the Java DSL for Spring Integration:

@Bean
IntegrationFlow servicePointIntegrationFlow() {
    return IntegrationFlows.from("servicePointExportChannel")
            .enrichHeaders(headerEnricherSpec -> headerEnricherSpec.header("pageNumber", new AtomicInteger()))
            .channel("singlePageRequestChannel")
            .get();
}
/**
 * Get a single page, and persist the contents of that page to the DB
 * Then, check if there are more pages.
 *   * If yes, request the next page by adding another message to the same channel (but with incremented pageNumber in the message header)
 *   * If no, publish a message to the next channel
 */
@Bean
IntegrationFlow paginatedDataFlow(ServicePointTransformer servicePointTransformer) {
    return IntegrationFlows.from("singlePageRequestChannel")
            .log(INFO, m -> "paginatedDataFlow called with pageNumber: " + m.getHeaders().get("pageNumber"))
            .handle(Http.outboundGateway(m -> myUrl + "&pageNumber={pageNumber}")
                    .uriVariable("pageNumber", m -> m.getHeaders().get("pageNumber", AtomicInteger.class).getAndIncrement())
                    .httpMethod(HttpMethod.GET)
                    .expectedResponseType(MyPage.class)
                    .mappedResponseHeaders(""))
            .routeToRecipients(route -> route
                    .recipientFlow(flow -> flow
                            .transform(MyPage::getServicePoints)
                            .transform(servicePointTransformer)
                            .handle(Jpa.updatingGateway(myEntityManagerFactory),
                                    spec -> spec.transactional(myTransactionManager))
                    .recipientFlow("payload.isNextPageAvailable", f -> f.channel("singlePageRequestChannel"))
                    .recipientFlow("!payload.isNextPageAvailable", f -> f.channel("someOtherChannel")))
            .get();
}

The payload/page is JSON and contains a top level field nextPageAvailable which can be access in its deserialized form as MyPage.isNextPageAvailable().

Upvotes: 1

Artem Bilan
Artem Bilan

Reputation: 121177

Yes, it is possible, although a bit tricky.

Assume your REST service require page as request param, so, you would like to make a request from the page #1 and loop (increment page param) until the service returns empty result.

So, you may have configuration for the REST service like:

<int-http:outbound-gateway url="http://service/elements?page={page}">
    <int-http:uri-variable name="page" expression="headers['page']"/>
</int-http:outbound-gateway>

Pay attention to that <int-http:uri-variable> definition. From the beginning you have to send the message to this <int-http:outbound-gateway> with the page header as a 1.

The reply from this gateway you should send to something like <recipient-list-router>, or <publish-subscribe-channel>, where one of the subscriber is still your splitter to to store items into the folder.

Another subscriber is a bit smart. It starts from <filter> to check if the payload (a result from the REST call) is empty, meaning that we have done and no more pages on the service to retrieve. Otherwise you use <header-enricher> to increment and replace the page header and send the result into that our first <int-http:outbound-gateway>.

Upvotes: 3

Related Questions