public class Runner extends ParameterizedBase
An execution has one input, one algorithm, and one output. Anything more complex can be expressed as a user program written in a JVM language.
Inputs and algorithms are decoupled by, respectively, producing and consuming a graph.
Currently only Graph
is supported but later updates may add support for new graph types
such as BipartiteGraph
.
Algorithms must explicitly support each type of output via implementation of interfaces. This is scalable as the number of outputs is small and finite.
Constructor and Description |
---|
Runner(String[] args)
Create an algorithm runner from the given arguments.
|
Modifier and Type | Method and Description |
---|---|
ExecutionEnvironment |
getExecutionEnvironment()
Get the ExecutionEnvironment.
|
String |
getName()
A unique, human-readable identifier.
|
DataSet |
getResult()
Get the result DataSet.
|
static void |
main(String[] args) |
Runner |
run()
Setup the Flink job with the graph input, algorithm, and output.
|
addParameter, configure, getUsage
public Runner(String[] args)
args
- command-line argumentspublic String getName()
Parameterized
getName
in interface Parameterized
getName
in class ParameterizedBase
public ExecutionEnvironment getExecutionEnvironment()
run()
.public DataSet getResult()
run()
.public Runner run() throws Exception
To then execute the job call execute()
.
Exception
- on errorCopyright © 2014–2021 The Apache Software Foundation. All rights reserved.