Awesome

This project contains sample implementation of custom components for Nussknacker.

Components

Common

We provide simple enricher provider which generates random string, using commons-text library to demonstrate how to add additional dependencies.

Streaming-Flink only

Flink provider contains sample Flink components to give an idea of how to implement Flink based sources, sinks and custom transformations.

• cdr (call detail record) - a source that emits specific case class
• csvSource - a source that emits a CSV record according to the row definition, shares the same Flink source function as cdr source
• csvSink - a sink that writes a row represented as a list of strings to the given file

Usage

To use custom component, you have to put fat jar with the component on model classpath. If you use default configuration, the easiest way is to put the jar in components/common directory, as this directory is added to classpath.

If you are using setup based on docker-compose, you can mount jar as additional volume, adding

- /(...)/sampleComponents.jar:/opt/nussknacker/components/common/sampleComponents.jar

in docker-compose

For more production-ready setups (e.g. Kubernetes) it is better to create image based on official Nussknacker image with jar added. See gradle project and docker/Dockerfile for a sample.

docker/runDocker.sh provides minimal script allowing to run Designer with added library (it doesn't connect with Kafka, Flink etc., but can be used to check if component is configured properly)

Gradle

./gradlew shadowJar to build fat jar ./gradlew buildImages to build Docker images of Designer and Lite runtime.

Sbt

sbt assembly to build fat jar. At the moment we don't provide scripts for image generation, use docker/Dockerfile as a base.