Class BiasedNumbers

java.lang.Object

com.carrotsearch.randomizedtesting.generators.BiasedNumbers

public final class BiasedNumbers
extends Object

Utility classes for selecting numbers at random, but not necessarily in an uniform way. The implementation will try to pick "evil" numbers more often than uniform selection would. This includes exact range boundaries, numbers very close to range boundaries, numbers very close (or equal) to zero, etc. The exact method of selection is implementation-dependent and may change (if we find even more evil ways).

Field Summary

Fields

Modifier and Type	Field	Description
private static final int	EVIL_RANDOM_REPRESENTATION_BITS
private static final int	EVIL_RANGE_LEFT
private static final int	EVIL_RANGE_RIGHT
private static final int	EVIL_SIMPLE_PROPORTION
private static final int	EVIL_VERY_CLOSE_RANGE_ENDS
private static final int	EVIL_ZERO_OR_NEAR

Constructor Summary

Constructors

Constructor	Description
BiasedNumbers()

Method Summary

Modifier and Type	Method	Description
private static int	flip(int floatBits)
private static long	flip(long bits)
private static float	fromSortable(int sortable)
private static double	fromSortable(long sortable)
static double	fuzzDown(Random r, double v, double min)	Fuzzify the input value by decreasing it by a few ulps, but never past min.
static float	fuzzDown(Random r, float v, float min)	Fuzzify the input value by decreasing it by a few ulps, but never past min.
static double	fuzzUp(Random r, double v, double max)	Fuzzify the input value by increasing it by a few ulps, but never past max.
static float	fuzzUp(Random r, float v, float max)	Fuzzify the input value by increasing it by a few ulps, but never past max.
static double	randomDoubleBetween(Random r, double min, double max)	A random double between min (inclusive) and max (inclusive).
static float	randomFloatBetween(Random r, float min, float max)	A random float between min (inclusive) and max (inclusive).
private static long	toSortable(double value)
private static int	toSortable(float value)

Methods inherited from class Object

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

Field Details

EVIL_RANGE_LEFT

private static final int EVIL_RANGE_LEFT

See Also:

• Constant Field Values

EVIL_RANGE_RIGHT

private static final int EVIL_RANGE_RIGHT

See Also:

• Constant Field Values

EVIL_VERY_CLOSE_RANGE_ENDS

private static final int EVIL_VERY_CLOSE_RANGE_ENDS

See Also:

• Constant Field Values

EVIL_ZERO_OR_NEAR

private static final int EVIL_ZERO_OR_NEAR

See Also:

• Constant Field Values

EVIL_SIMPLE_PROPORTION

private static final int EVIL_SIMPLE_PROPORTION

See Also:

• Constant Field Values

EVIL_RANDOM_REPRESENTATION_BITS

private static final int EVIL_RANDOM_REPRESENTATION_BITS

See Also:

• Constant Field Values

Constructor Details

BiasedNumbers

public BiasedNumbers()

Method Details

randomDoubleBetween

public static double randomDoubleBetween(Random r,
 double min,
 double max)

A random double between min (inclusive) and max (inclusive). If you wish to have an exclusive range, use Math.nextAfter(double, double) to adjust the range. The code was inspired by GeoTestUtil from Apache Lucene.

Parameters:

min - Left range boundary, inclusive. May be Double.NEGATIVE_INFINITY, but not NaN.

max - Right range boundary, inclusive. May be Double.POSITIVE_INFINITY, but not NaN.

fuzzDown

public static double fuzzDown(Random r,
 double v,
 double min)

Fuzzify the input value by decreasing it by a few ulps, but never past min.

fuzzUp

public static double fuzzUp(Random r,
 double v,
 double max)

Fuzzify the input value by increasing it by a few ulps, but never past max.

fromSortable

private static double fromSortable(long sortable)

toSortable

private static long toSortable(double value)

flip

private static long flip(long bits)

randomFloatBetween

public static float randomFloatBetween(Random r,
 float min,
 float max)

A random float between min (inclusive) and max (inclusive). If you wish to have an exclusive range, use Math.nextAfter(float, double) to adjust the range. The code was inspired by GeoTestUtil from Apache Lucene.

Parameters:

min - Left range boundary, inclusive. May be Float.NEGATIVE_INFINITY, but not NaN.

max - Right range boundary, inclusive. May be Float.POSITIVE_INFINITY, but not NaN.

fuzzDown

public static float fuzzDown(Random r,
 float v,
 float min)

Fuzzify the input value by decreasing it by a few ulps, but never past min.

fuzzUp

public static float fuzzUp(Random r,
 float v,
 float max)

Fuzzify the input value by increasing it by a few ulps, but never past max.

fromSortable

private static float fromSortable(int sortable)

toSortable

private static int toSortable(float value)

flip

private static int flip(int floatBits)