README.adoc

Standalone Java Functions

These are self contained Java functions, each implements java.util.function.Function and can be used as a component that provides a value for a given input value in a variety of streaming application platforms such as Spring Cloud Data Flow, Project RIFF etc. The functions can be run as part of standalone Spring Boot applications.

Available Functions

splitter-function

This is a "one-to-many" Java function with a signature as Function<Message<?>, List<Message<?>>>. So, every inbound message is splitted into several output messages. This function is fully based on the Spring integration AbstractMessageSplitter and supports a delimiters, expression, applySequence properties. Also if one of the fileMarkers or charset properties is present, the FileSplitter implementation is used and an incoming payload is treated as file for splitting.

spel-function

Transform incoming message payload based on a SpEL expression and produces an outbound message.

filter-function

Filter an incoming message based on a boolean SpEL expression and produces that message or nothing according expression result.

header-enricher-function

Enrich the incoming message headers based on expression evaluated headers.

tasklauncher-function

This is Java function with signature Function<LaunchRequest, Optional<Long> that submits a task launch request against a Spring Cloud Data Flow server. The LaunchRequest is defined as:

public class LaunchRequest {
	@JsonProperty("args")
	private List<String> commandlineArguments = new ArrayList<>();
	@JsonProperty("deploymentProps")
	private Map<String, String> deploymentProperties = new HashMap<>();
	@JsonProperty("name")
	private String taskName;
...
}

And posted as JSON. e.g.,

{"name":"timestamp"}

or

{"name":"mytask","args":["a0=v0","a1=v1"],"deploymentProps":{"prop0":"val0"}}

The only required field is the task name which represents an existing task defined in Data Flow.

If the Data Flow server is not deployed to the default URL (http://localhost:9393), you must configure DataFlowClientProperties, including authentication if necessary.

The function will return an Optional with the taskId as the value. Before submitting the request, it first checks that the configured task platform is not at its current task execution limit (20 by default). If the task platform is at its limit, the function returns Optional.empty() and will log a warning message.