laserdisc-io / eisner   0.1.2

MIT License GitHub

Kafka (Streams' topologies) translator

Scala versions: 2.12
sbt plugins: 1.x

Eisner

Build Status Maven Central Scala Steward badge

Eisner is a Kafka (Streams' topologies) translator

Motivating example

Kafka Stream's topologies can quickly become complex to grok without picturing them visually.

Eisner fills this gap by offering to developers an sbt plugin to create a graphical representation of Kafka Stream's Topology(ies). It took ideas and inspiration from this online tool by Joshua Koo.

For example, assuming the following topology description:

Topology
Sub-topologies:
Sub-topology: 0
	Source:  KSTREAM-SOURCE-0000000000 (topics: [conversation-meta])
	--> KSTREAM-TRANSFORM-0000000001
	Processor: KSTREAM-TRANSFORM-0000000001 (stores: [conversation-meta-state])
	--> KSTREAM-KEY-SELECT-0000000002
	<-- KSTREAM-SOURCE-0000000000
	Processor: KSTREAM-KEY-SELECT-0000000002 (stores: [])
	--> KSTREAM-FILTER-0000000005
	<-- KSTREAM-TRANSFORM-0000000001
	Processor: KSTREAM-FILTER-0000000005 (stores: [])
	--> KSTREAM-SINK-0000000004
	<-- KSTREAM-KEY-SELECT-0000000002
	Sink: KSTREAM-SINK-0000000004 (topic: count-resolved-repartition)
	<-- KSTREAM-FILTER-0000000005
Sub-topology: 1
	Source: KSTREAM-SOURCE-0000000006 (topics: [count-resolved-repartition])
	--> KSTREAM-AGGREGATE-0000000003
	Processor: KSTREAM-AGGREGATE-0000000003 (stores: [count-resolved])
	--> KTABLE-TOSTREAM-0000000007
	<-- KSTREAM-SOURCE-0000000006
	Processor: KTABLE-TOSTREAM-0000000007 (stores: [])
	--> KSTREAM-SINK-0000000008
	<-- KSTREAM-AGGREGATE-0000000003
	Sink: KSTREAM-SINK-0000000008 (topic: streams-count-resolved)
	<-- KTABLE-TOSTREAM-0000000007

Eisner will first translate it into the following DOT representation:

digraph G {
	label = "Kafka Streams Topology"
	subgraph cluster_0 {
		label = " Sub-\ntopology: 0";
		style = filled;
		color = lightgrey;
		node [style = filled, color = white];
		"KSTREAM-\nSOURCE-\n0000000000" -> "KSTREAM-\nTRANSFORM-\n0000000001";
		"KSTREAM-\nTRANSFORM-\n0000000001" -> "KSTREAM-\nKEY-\nSELECT-\n0000000002";
		"KSTREAM-\nKEY-\nSELECT-\n0000000002" -> "KSTREAM-\nFILTER-\n0000000005";
		"KSTREAM-\nFILTER-\n0000000005" -> "KSTREAM-\nSINK-\n0000000004";
	}
	subgraph cluster_1 {
		label = " Sub-\ntopology: 1";
		style = filled;
		color = lightgrey;
		node [style = filled, color = white];
		"KSTREAM-\nSOURCE-\n0000000006" -> "KSTREAM-\nAGGREGATE-\n0000000003";
		"KSTREAM-\nAGGREGATE-\n0000000003" -> "KTABLE-\nTOSTREAM-\n0000000007";
		"KTABLE-\nTOSTREAM-\n0000000007" -> "KSTREAM-\nSINK-\n0000000008";
	}
	"conversation-\nmeta" -> "KSTREAM-\nSOURCE-\n0000000000";
	"KSTREAM-\nTRANSFORM-\n0000000001" -> "conversation-\nmeta-\nstate";
	"KSTREAM-\nSINK-\n0000000004" -> "count-\nresolved-\nrepartition";
	"count-\nresolved-\nrepartition" -> "KSTREAM-\nSOURCE-\n0000000006";
	"KSTREAM-\nAGGREGATE-\n0000000003" -> "count-\nresolved";
	"KSTREAM-\nSINK-\n0000000008" -> "streams-\ncount-\nresolved";
	"conversation-\nmeta" [shape = rect];
	"count-\nresolved-\nrepartition" [shape = rect];
	"streams-\ncount-\nresolved" [shape = rect];
	"conversation-\nmeta-\nstate" [shape = cylinder];
	"count-\nresolved" [shape = cylinder];
}

then, leveraging the JS wrapper of graphviz (see here), Eisner will convert it into the following SVG:

and ultimately, it will add some "hand drawn" feeling to it using roughjs, to turn it into a new SVG like the following:

Usage

Import the plugin by placing the following line in project/plugins.sbt:

// check the current version on Maven Central (or use the badge above)
addSbtPlugin("io.laserdisc" %% "sbt-eisner" % version)

Then, in your build.sbt, enable Eisner by adding the following to your module/project where your Kafka Stream's Topology(ies) definitions are:

enablePlugins(EisnerPlugin)

Finally, trigger Eisner's Topology to SVG translation by running within sbt:

sbt> eisner

That's it (potentially)! Eisner is capable of scanning your classes searching for:

  • zero-arg constructor classes that extend org.apache.kafka.streams.Topology
  • fields or zero-arg methods that return org.apache.kafka.streams.Topology
  • fields or zero-arg methods that return scala.Function0[org.apache.kafka.streams.Topology]
  • fields or zero-arg methods that return java.util.function.Supplier[org.apache.kafka.streams.Topology]

SVGs will be created for each, named after the class(+field/method) that defines the Topology. Files are generated by default in target/eisner. As long as topologies are not changed between subsequent runs of eisner (without clean), SVGs will not be regenerated.

Tweaking with it

Currently, Eisner supports a handful of options, namely:

  • eisnerColorSink: a String representing a valid graphviz color used to draw sinks, defaults to black
  • eisnerColorSubtopology: a String representing a valid graphviz color used to frame sub topologies, defaults to lightgrey
  • eisnerColorTopic: a String representing a valid graphviz color used to draw topics, defaults to black
  • eisnerRoughSVG: a Boolean indicating whether Eisner should produce SVGs that look like hand drawings, defaults to true
  • eisnerTargetDirectory: a File which represents the directory where all SVGs will be saved into, defaults to [current project]/target/eisner
  • eisnerTopologies: a Seq[String] representing the fully qualified names of classes implementing org.apache.kafka.streams.Topology. This is useful in all cases where you need to control which SVGs get generated
  • eisnerTopologiesSnippet: a Option[String] representing a Scala snippet (inclusive of all necessary imports) that evaluates to a Seq[(String, org.apache.kafka.streams.Topology)], where the String represents the package.name you want to give to the target file (dots will be converted to path separators). This is useful in all cases where automatic classloader scanning would not work, e.g. because you define your topologies in non zero-args methods or in fields that return scala.Function1. See here for a practical example (specifically how the topology is defined here and how to get Eisner to discover it here).

License

Eisner is licensed under the MIT License (the "License"); you may not use this software except in compliance with the License.

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.