uk.ac.starlink.ttools.plot2

Class SplitRunner<S extends Splittable<S>>

java.lang.Object

uk.ac.starlink.ttools.plot2.SplitRunner<S>

Direct Known Subclasses:

TupleRunner

public abstract class SplitRunner<S extends Splittable<S>>
extends Object

Utility class for making use of SplitProcessor instances.

Since:

17 Sep 2019

Author:

Mark Taylor

Constructor Summary

Constructors

Constructor and Description
SplitRunner()

Method Summary

Modifier and Type	Method and Description
abstract <A> A	collect(SplitCollector<S,A> collector, java.util.function.Supplier<S> splitSupplier) Performs a collection operation.
abstract <A> A	collectPool(SplitCollector<S,A> collector, java.util.function.Supplier<S> splitSupplier) Performs a collection operation with a hint that accumulator pooling is recommended.
static <S extends Splittable<S>> SplitRunner<S>	createBenchRunner() Returns an instance that invokes multiple SplitProcessor instances of different types and reports comparative timings to stdout.
static <S extends Splittable<S>> SplitRunner<S>	createDefaultRunner() Returns an instance that uses SplitProcessor instances in the normal way with the default concurrency policy.
static <S extends Splittable<S>> SplitRunner<S>	createSequentialRunner() Returns an instance that performs single-threaded sequential execution.
static <S extends Splittable<S>> SplitRunner<S>	createStandardRunner(SplitPolicy policy) Returns an instance that uses SplitProcessor instances in the normal way with a given concurrency policy.
abstract boolean	willAttemptSplit(S content) Indicates whether an attempt will be made to split a top-level splittable in order to process it.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

SplitRunner

public SplitRunner()

Method Detail

collect

public abstract <A> A collect(SplitCollector<S,A> collector,
                              java.util.function.Supplier<S> splitSupplier)

Performs a collection operation. This usually does not do accumulator pooling.

Parameters:

collector - collector

splitSupplier - supplier for splittable object to process

Returns:

collected result

collectPool

public abstract <A> A collectPool(SplitCollector<S,A> collector,
                                  java.util.function.Supplier<S> splitSupplier)

Performs a collection operation with a hint that accumulator pooling is recommended.

Parameters:

collector - collector

splitSupplier - supplier for splittable object to process

willAttemptSplit

public abstract boolean willAttemptSplit(S content)

Indicates whether an attempt will be made to split a top-level splittable in order to process it. If it's too small for instance, false will be returned.

Parameters:

content - splittable

Returns:

true iff processing will try to split content

createDefaultRunner

public static <S extends Splittable<S>> SplitRunner<S> createDefaultRunner()

Returns an instance that uses SplitProcessor instances in the normal way with the default concurrency policy.

Returns:

default parallel instance

createStandardRunner

public static <S extends Splittable<S>> SplitRunner<S> createStandardRunner(SplitPolicy policy)

Returns an instance that uses SplitProcessor instances in the normal way with a given concurrency policy.

Parameters:

policy - concurrency policy

Returns:

standard parallel instance

createBenchRunner

public static <S extends Splittable<S>> SplitRunner<S> createBenchRunner()

Returns an instance that invokes multiple SplitProcessor instances of different types and reports comparative timings to stdout. This is not intended for production use, obviously.

Returns:

benchmarking instance

createSequentialRunner

public static <S extends Splittable<S>> SplitRunner<S> createSequentialRunner()

Returns an instance that performs single-threaded sequential execution.

Returns:

sequential execution instance