at.borkowski.scovillej.impl.series

Class NumberSeriesImpl<T extends Number>

java.lang.Object

at.borkowski.scovillej.impl.series.SeriesImpl<T>

at.borkowski.scovillej.impl.series.NumberSeriesImpl<T>

Type Parameters:

T - The type of numbers measured

All Implemented Interfaces:

Series<T>, SeriesProvider<T>, SeriesResult<T>

Direct Known Subclasses:

DoubleSeriesImpl, FloatSeriesImpl, IntegerSeriesImpl, LongSeriesImpl

public abstract class NumberSeriesImpl<T extends Number>extends SeriesImpl<T>

A base class for implementing a series of numbers.

Field Summary

Fields

Modifier and Type	Field and Description
protected long	count

Constructor Summary

Constructors

Constructor and Description
NumberSeriesImpl(Comparator<T> comparator, Class<T> clazz)

Method Summary

Modifier and Type	Method and Description
protected abstract T	calcNativeMedian(T a, T b) Abstract method which must be implemented by sub-classes.
static <T> SeriesProvider<T>	createIfKnown(Class<T> clazz) Creates a SeriesProvider for a built-in series type.
protected Formatter	formatMetrics(Formatter f) Formats metrics for this class using the supplied formatter.
Map<Long,T>	getAll() Returns all measured values as a map from the tick to the value.
Double	getAverage() Returns an average (arithmetic mean) of all measured values.
Map<Long,Double>	getAveraged(long classWidth) Returns an averaged version of all values.
long	getCount() Returns the number of measured values.
Double	getDoubleMedian() Returns the median of all measured values.
static Collection<Class<?>>	getKnownSeriesClasses() Returns a list of all known series classes.
T	getMax() Returns the maximum value of all measured values.
T	getMin() Returns the minimum value of all measured values.
T	getNativeMedian() Returns the median of all measured values.
T[]	getNativeMedians() Returns the median candidates of all measured values.
Double	getStandardDeviation() Returns the standard deviation of all measured values.
Class<T>	getValueClass() Returns the runtime class type for T
boolean	hasSingleMedian() Determines whether the series has one median (because it has an odd number of values, or because it is even and two median candidates are equal).
void	initialize(Simulation sim) Initializes the series provider.
void	measure(T value) Adds a measured value to the series.
protected T	medianA()
protected T	medianB()

Methods inherited from class at.borkowski.scovillej.impl.series.SeriesImpl

toString

Methods inherited from class java.lang.Object

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

Field Detail

count

protected long count

Constructor Detail

NumberSeriesImpl

public NumberSeriesImpl(Comparator<T> comparator,                        Class<T> clazz)

Method Detail

getValueClass

public Class<T> getValueClass()

Description copied from interface: SeriesProvider

Returns the runtime class type for T

Returns:

the class object for T

initialize

public void initialize(Simulation sim)

Description copied from interface: SeriesProvider

Initializes the series provider. This method is guaranteed to be called before any methods of Series or SeriesResult.

Parameters:

sim - The simulation this series is to be performed at

getAll

public Map<Long,T> getAll()

Description copied from interface: SeriesResult

Returns all measured values as a map from the tick to the value.

Returns:

all measured values

getAveraged

public Map<Long,Double> getAveraged(long classWidth)

Description copied from interface: SeriesResult

Returns an averaged version of all values. The measured values are averaged into classes with a width of classWidth. The returned map contains all averaged results as a map from the tick (the beginning, "left border" of the class) to the average value.

Parameters:

classWidth - the width in ticks for the classes

Returns:

the resulting map

calcNativeMedian

protected abstract T calcNativeMedian(T a,                                      T b)

Abstract method which must be implemented by sub-classes. The method must calculate the median, with two values given. If this method is called, it is guaranteed that there are two median candidates and the implementation is responsible of calculating the actual median. Note that in general, mean-like calculation should be performed.

Parameters:

a - the first (left) median candidate

b - the second (right) median candidate

Returns:

the median

measure

public void measure(T value)

Description copied from interface: Series

Adds a measured value to the series. The value is added using the current tick.

Parameters:

value - the value to add

getAverage

public Double getAverage()

Description copied from interface: SeriesResult

Returns an average (arithmetic mean) of all measured values.

Returns:

the average

hasSingleMedian

public boolean hasSingleMedian()

Description copied from interface: SeriesResult

Determines whether the series has one median (because it has an odd number of values, or because it is even and two median candidates are equal).

Returns:

whether the series has one median

medianA

protected T medianA()

medianB

protected T medianB()

getDoubleMedian

public Double getDoubleMedian()

Description copied from interface: SeriesResult

Returns the median of all measured values. This method always returns a Double, even if the measured values are integer and one median is present. For a more exact version, see SeriesResult.getNativeMedian(). If there are two median values (because the number of measurements is even), the arithmetic mean of the two values is returned.

Returns:

the median

getStandardDeviation

public Double getStandardDeviation()

Description copied from interface: SeriesResult

Returns the standard deviation of all measured values.

Returns:

the standard deviation

getCount

public long getCount()

Description copied from interface: SeriesResult

Returns the number of measured values.

Returns:

the number of measured values

getMin

public T getMin()

Description copied from interface: SeriesResult

Returns the minimum value of all measured values.

Returns:

the minimum value

getMax

public T getMax()

Description copied from interface: SeriesResult

Returns the maximum value of all measured values.

Returns:

the maximum value

getNativeMedians

public T[] getNativeMedians()

Description copied from interface: SeriesResult

Returns the median candidates of all measured values.

Returns:

the native median candidates

getNativeMedian

public T getNativeMedian()

Description copied from interface: SeriesResult

Returns the median of all measured values. In contrast to SeriesResult.getDoubleMedian(), this method returns an exact median if there is one median. If there are two median values, the returned value depends on the implemented measurement series, but will be performed in a mean-like way (possible rounded towards some integer value).

Returns:

the median in native type

getKnownSeriesClasses

public static Collection<Class<?>> getKnownSeriesClasses()

Returns a list of all known series classes.

Returns:

all known series classes

createIfKnown

public static <T> SeriesProvider<T> createIfKnown(Class<T> clazz)

Creates a SeriesProvider for a built-in series type. Returns null if the provided type is not built-in.

Type Parameters:

T - the (static) type of series to create

Parameters:

clazz - the (runtime) type of series to create

Returns:

the series provider

formatMetrics

protected Formatter formatMetrics(Formatter f)

Description copied from class: SeriesImpl

Formats metrics for this class using the supplied formatter. Sub-classes should override this method to provide detailed string representations of the metrics (min, max, average, etc.)

Overrides:

formatMetrics in class SeriesImpl<T extends Number>

Parameters:

f - the formatter to use

Returns:

the formatter itself (the supplied parameter formatter)