Enhance Function<List<T>, List<R>> to Publish Individual Messages with splitMode=true

[2tsumo-hitori]

Currently, when registering a Function<List, List> as a bean in Spring Cloud Function, Spring Cloud Stream publishes the entire list as a single message instead of splitting it into individual messages, even when the producer sets splitMode=true.

I have opened a related issue in Spring Cloud Function: Issue #1242. If this feature is improved in Spring Cloud Function and subsequently applied to Spring Cloud Stream, it would enhance message processing efficiency.

With this enhancement:

• Before:
  • Consumed messages were aggregated into a single list and published as a single message.
• After:
  • By enabling splitMode=true, each element in the list can be published as an individual message automatically, eliminating the need for additional handling via StreamBridge.

This is a test in a local Kafka setup.

🖼️ Before applying splitMode:

🖼️ After applying splitMode:

Spring Boot Server Configuration
(This setup references locally modified versions of Spring Cloud Function and Spring Cloud Stream, which I updated and installed in the .m2 folder.)

spring:
  cloud:
    function:
      definition: eventStream
    stream:
      bindings:
        eventStream-in-0:
          destination: prev-topic
          group: test-group-unique
          consumer:
            batch-mode: true
        eventStream-out-0:
          destination: next-topic
          producer:
            split-mode: true
      kafka:
        binder:
          brokers: localhost:9092
          auto-create-topics: true

@Bean
public Function<List<String>, List<String>> eventStream() {
      return messages -> messages
            .stream()
            .map(String::toUpperCase)
            .toList();
}