Package weka.knowledgeflow

Class StepTask<T>

java.lang.Object
weka.knowledgeflow.StepTask<T>
Type Parameters:
T - the type of the result stored in the returned ExecutionResult object
All Implemented Interfaces:
Serializable, Callable<ExecutionResult<T>>
public abstract class StepTask<T> extends Object implements Callable<ExecutionResult<T>>, Serializable
A task that can be executed by the ExecutionEnvironment's submitTask() service. Step's wanting to use this to execute units of work in parallel must extends it and can work with it in one of two ways:

1. By using the Future returned by submitTask(), or
2. By registering an implementation of StepCallback when creating a subclass of StepTask.
Subclasses of StepTask should store their results (and potentially errors) in the provided ExecutionResult member variable (obtainable by calling getExecutionResult()).
Author:
Mark Hall (mhall{[at]}pentaho{[dot]}com)
See Also:

  • Constructor Details

    • StepTask

      public StepTask(Step source)
      Constructor. Use this constructor if you are going to access the Future returned by ExecutionEnvironment.submitTask().
      Parameters:
      source - the source step producing this task

    • StepTask

      public StepTask(Step source, boolean resourceIntensive)
      Constructor. Use this constructor if you are going to access the Future returned by ExecutionEnvironment.submitTask()
      Parameters:
      source - the source step producing this task
      resourceIntensive - true if this task is cpu/memory intensive

    • StepTask

      public StepTask(Step source, StepTaskCallback<T> callback)
      Constructor with supplied callback. Use this constructor to be notified via the supplied callback when a the task has completed processing
      Parameters:
      source - the source step producing this task
      callback - the callback to use

    • StepTask

      public StepTask(Step source, StepTaskCallback<T> callback, boolean resourceIntensive)
      Constructor with supplied callback. Use this constructor to be notified via the supplied callback when a task has completed processing
      Parameters:
      source - the source step producing this task
      callback - the callback to use
      resourceIntensive - true if this task is cpu/memory intensive

  • Method Details

    • setResourceIntensive

      public void setResourceIntensive(boolean resourceIntensive)
      Set whether this StepTask is resource intensive (cpu/memory) or not. By default, a StepTask is resource intensive
      Parameters:
      resourceIntensive - false if this StepTask is not resource intensive

    • isResourceIntensive

      public boolean isResourceIntensive()
      Get whether this StepTask is resource intensive (cpu/memory) or not. By default, a StepTask is resource intensive
      Returns:
      false if this StepTask is not resource intensive

    • setMustRunSingleThreaded

      public void setMustRunSingleThreaded(boolean singleThreaded)
      Set whether this StepTask must run single threaded - i.e. only one of these tasks is executing at any one time in the JVM. The Knowledge Flow uses a special executor service with a single worker thread to execute these tasks. This property, if true, overrides isResourceIntensive().
      Parameters:
      singleThreaded - true if this task must run single threaded

    • getMustRunSingleThreaded

      public boolean getMustRunSingleThreaded()
      Get whether this StepTask must run single threaded - i.e. only one of these tasks is executing at any one time in the JVM. The Knowledge Flow uses a special executor service with a single worker thread to execute these tasks. This property, if true, overrides isResourceIntensive().
      Returns:
      true if this task must run single threaded

    • call

      public ExecutionResult<T> call() throws Exception
      Executor service calls this method to do the work
      Specified by:
      call in interface Callable<T>
      Returns:
      the results of execution in an ExecutionResult
      Throws:
      Exception

    • process

      public abstract void process() throws Exception
      The actual work gets done here. Subclasses to override. Subclasses can use getExecutionResult() to obtain an ExecutionResult object to store their results in
      Throws:
      Exception