Package org.apache.lucene.codecs.lucene104

Enum Class Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding

java.lang.Object

java.lang.Enum<Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding>

org.apache.lucene.codecs.lucene104.Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding

All Implemented Interfaces:

Serializable, Comparable<Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding>, Constable

Enclosing class:

Lucene104ScalarQuantizedVectorsFormat

public static enum Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding
extends Enum<Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding>

Allowed encodings for scalar quantization.

This specifies how many bits are used per dimension and also dictates packing of dimensions into a byte stream.

Nested Class Summary

Nested classes/interfaces inherited from class java.lang.Enum

Enum.EnumDesc<E extends Enum<E>>

Enum Constant Summary

Enum Constants

Enum Constant	Description
DIBIT_QUERY_NIBBLE	Each dimension is quantized to 2 bits (dibit) and packed into bytes.
PACKED_NIBBLE	Each dimension is quantized to 4 bits two values are packed into each output byte.
SEVEN_BIT	Each dimension is quantized to 7 bits and treated as a signed value.
SINGLE_BIT_QUERY_NIBBLE	Each dimension is quantized to a single bit and packed into bytes.
UNSIGNED_BYTE	Each dimension is quantized to 8 bits and treated as an unsigned value.

Method Summary

Modifier and Type	Method	Description
static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding	fromNumBits(int bits)
static Optional<Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding>	fromWireNumber(int wireNumber)	Returns the encoding for the given wire number, or empty if unknown.
byte	getBits()	Return the number of bits used per dimension.
int	getDiscreteDimensions(int dimensions)	Return the number of dimensions rounded up to fit into whole bytes.
int	getDocBitsPerDim()	Return the number of dimensions that can be packed into a single byte.
int	getDocPackedLength(int dimensions)	Return the number of bytes required to store a packed vector of the given dimensions.
byte	getQueryBits()
int	getQueryBitsPerDim()
int	getQueryPackedLength(int dimensions)
boolean	isAsymmetric()
static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding	valueOf(String name)	Returns the enum constant of this class with the specified name.
static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding[]	values()	Returns an array containing the constants of this enum class, in the order they are declared.

Methods inherited from class java.lang.Enum

clone, compareTo, describeConstable, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Enum Constant Details

UNSIGNED_BYTE

public static final Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding UNSIGNED_BYTE

Each dimension is quantized to 8 bits and treated as an unsigned value.

PACKED_NIBBLE

public static final Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding PACKED_NIBBLE

Each dimension is quantized to 4 bits two values are packed into each output byte.

SEVEN_BIT

public static final Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding SEVEN_BIT

Each dimension is quantized to 7 bits and treated as a signed value.

This is intended for backwards compatibility with older iterations of scalar quantization. This setting will produce an index the same size as UNSIGNED_BYTE but will produce less accurate vector comparisons.

SINGLE_BIT_QUERY_NIBBLE

public static final Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding SINGLE_BIT_QUERY_NIBBLE

Each dimension is quantized to a single bit and packed into bytes. During query time, the query vector is quantized to 4 bits per dimension.

This is the most space efficient encoding, and will produce an index 8x smaller than UNSIGNED_BYTE. However, this comes at the cost of accuracy.

DIBIT_QUERY_NIBBLE

public static final Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding DIBIT_QUERY_NIBBLE

Each dimension is quantized to 2 bits (dibit) and packed into bytes. During query time, the query vector is quantized to 4 bits per dimension.

This encoding produces an index 4x smaller than UNSIGNED_BYTE, offering a balance between the compression of SINGLE_BIT_QUERY_NIBBLE and the accuracy of PACKED_NIBBLE.

Method Details

values

public static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding[] values()

Returns an array containing the constants of this enum class, in the order they are declared.

Returns:

an array containing the constants of this enum class, in the order they are declared

valueOf

public static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding valueOf(String name)

Returns the enum constant of this class with the specified name. The string must match exactly an identifier used to declare an enum constant in this class. (Extraneous whitespace characters are not permitted.)

Parameters:

name - the name of the enum constant to be returned.

Returns:

the enum constant with the specified name

Throws:

IllegalArgumentException - if this enum class has no constant with the specified name

NullPointerException - if the argument is null

fromNumBits

public static Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding fromNumBits(int bits)

isAsymmetric

public boolean isAsymmetric()

getBits

public byte getBits()

Return the number of bits used per dimension.

getQueryBits

public byte getQueryBits()

getDiscreteDimensions

public int getDiscreteDimensions(int dimensions)

Return the number of dimensions rounded up to fit into whole bytes.

getDocBitsPerDim

public int getDocBitsPerDim()

Return the number of dimensions that can be packed into a single byte.

getQueryBitsPerDim

public int getQueryBitsPerDim()

getDocPackedLength

public int getDocPackedLength(int dimensions)

Return the number of bytes required to store a packed vector of the given dimensions.

getQueryPackedLength

public int getQueryPackedLength(int dimensions)

fromWireNumber

public static Optional<Lucene104ScalarQuantizedVectorsFormat.ScalarEncoding> fromWireNumber(int wireNumber)

Returns the encoding for the given wire number, or empty if unknown.