org.locationtech.geowave.core.index.sfc.hilbert

Class UnboundedHilbertSFCOperations

java.lang.Object

org.locationtech.geowave.core.index.sfc.hilbert.UnboundedHilbertSFCOperations

All Implemented Interfaces:

HilbertSFCOperations

public class UnboundedHilbertSFCOperations
extends Object
implements HilbertSFCOperations

This supports Compact Hilbert SFC operations using a BigInteger internally to represent intermediate results. This can be significantly slower than using a primitive long for intermediate results but can support arbitrarily many bits of precision.

Field Summary

Fields

Modifier and Type	Field and Description
protected BigDecimal[]	binsPerDimension
protected BigInteger	maxHilbertValue
protected BigInteger	minHilbertValue
protected static BigInteger	UNIT_CELL_SIZE

Constructor Summary

Constructors

Constructor and Description
UnboundedHilbertSFCOperations()

Method Summary

Modifier and Type	Method and Description
MultiDimensionalNumericData	convertFromHilbert(byte[] hilbertValue, com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve, SFCDimensionDefinition[] dimensionDefinitions) Convert the single SFC value to the ranges of raw values that it represents
byte[]	convertToHilbert(Double[] values, com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve, SFCDimensionDefinition[] dimensionDefinitions) Convert the raw values (ordered per dimension) to a single SFC value
RangeDecomposition	decomposeRange(NumericData[] rangePerDimension, com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve, SFCDimensionDefinition[] dimensionDefinitions, int totalPrecision, int maxFilteredIndexedRanges, boolean removeVacuum, boolean overInclusiveOnEdge) Decompose the raw range per dimension values into an optimal set of compact Hilbert SFC ranges
BigInteger	getEstimatedIdCount(MultiDimensionalNumericData data, SFCDimensionDefinition[] dimensionDefinitions) Get a quick (minimal complexity calculation) estimate of the total row IDs a particular data would require to fully cover with SFC values
double[]	getInsertionIdRangePerDimension(SFCDimensionDefinition[] dimensionDefinitions) * Get the range/size of a single insertion ID for each dimension
long[]	indicesFromHilbert(byte[] hilbertValue, com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve, SFCDimensionDefinition[] dimensionDefinitions) Convert the single SFC value to the per dimension SFC coordinates that it represents
void	init(SFCDimensionDefinition[] dimensionDefs) initialize this set of operations with the given dimension definitions
long[]	normalizeRange(double minValue, double maxValue, int dimension, SFCDimensionDefinition boundedDimensionDefinition) Determines the coordinates a given range will span within this space filling curve.

Methods inherited from class java.lang.Object

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

Field Detail

UNIT_CELL_SIZE

protected static final BigInteger UNIT_CELL_SIZE

binsPerDimension

protected BigDecimal[] binsPerDimension

minHilbertValue

protected BigInteger minHilbertValue

maxHilbertValue

protected BigInteger maxHilbertValue

Constructor Detail

UnboundedHilbertSFCOperations

public UnboundedHilbertSFCOperations()

Method Detail

init

public void init(SFCDimensionDefinition[] dimensionDefs)

Description copied from interface: HilbertSFCOperations

initialize this set of operations with the given dimension definitions

Specified by:

init in interface HilbertSFCOperations

Parameters:

dimensionDefs - the dimension definitions to use

convertToHilbert

public byte[] convertToHilbert(Double[] values,
                               com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve,
                               SFCDimensionDefinition[] dimensionDefinitions)

Description copied from interface: HilbertSFCOperations

Convert the raw values (ordered per dimension) to a single SFC value

Specified by:

convertToHilbert in interface HilbertSFCOperations

Parameters:

values - a raw value per dimension in order

compactHilbertCurve - the compact Hilbert curve to use for the conversion

dimensionDefinitions - a set of dimension definitions to use to normalize the raw values

Returns:

the Hilbert SFC value

indicesFromHilbert

public long[] indicesFromHilbert(byte[] hilbertValue,
                                 com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve,
                                 SFCDimensionDefinition[] dimensionDefinitions)

Description copied from interface: HilbertSFCOperations

Convert the single SFC value to the per dimension SFC coordinates that it represents

Specified by:

indicesFromHilbert in interface HilbertSFCOperations

Parameters:

hilbertValue - the computed hilbert value to invert back to integer coordinates per dimension

compactHilbertCurve - the compact Hilbert curve to use for the conversion

dimensionDefinitions - a set of dimension definitions to use to determine the bits of precision per dimension that is expected in the compact hilbert curve

Returns:

the integer coordinate value per dimension that the given hilbert value represents

convertFromHilbert

public MultiDimensionalNumericData convertFromHilbert(byte[] hilbertValue,
                                                      com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve,
                                                      SFCDimensionDefinition[] dimensionDefinitions)

Description copied from interface: HilbertSFCOperations

Convert the single SFC value to the ranges of raw values that it represents

Specified by:

convertFromHilbert in interface HilbertSFCOperations

Parameters:

hilbertValue - the computed hilbert value to invert back to native coordinates

compactHilbertCurve - the compact Hilbert curve to use for the conversion

dimensionDefinitions - a set of dimension definitions to use to normalize the raw values

Returns:

the ranges of values that the hilbert represents, inclusive on start and exclusive on end for each range

decomposeRange

public RangeDecomposition decomposeRange(NumericData[] rangePerDimension,
                                         com.google.uzaygezen.core.CompactHilbertCurve compactHilbertCurve,
                                         SFCDimensionDefinition[] dimensionDefinitions,
                                         int totalPrecision,
                                         int maxFilteredIndexedRanges,
                                         boolean removeVacuum,
                                         boolean overInclusiveOnEdge)

Description copied from interface: HilbertSFCOperations

Decompose the raw range per dimension values into an optimal set of compact Hilbert SFC ranges

Specified by:

decomposeRange in interface HilbertSFCOperations

Parameters:

rangePerDimension - the raw range per dimension

compactHilbertCurve - the compact Hilbert curve to use for the conversion

dimensionDefinitions - a set of dimension definitions to use to normalize the raw values

totalPrecision - the total precision of the dimension definitions, for convenience

maxFilteredIndexedRanges - the maximum number of ranges, if < 0 it will be unlimited

removeVacuum - a flag to pass to the compact hilbert curve range decomposition

Returns:

the optimal SFC range decomposition for the raw-valued ranges

getEstimatedIdCount

public BigInteger getEstimatedIdCount(MultiDimensionalNumericData data,
                                      SFCDimensionDefinition[] dimensionDefinitions)

Description copied from interface: HilbertSFCOperations

Get a quick (minimal complexity calculation) estimate of the total row IDs a particular data would require to fully cover with SFC values

Specified by:

getEstimatedIdCount in interface HilbertSFCOperations

Parameters:

data - the dataset

dimensionDefinitions - a set of dimension definitions to use to normalize the raw values

Returns:

the total estimated row IDs the data would require to fully cover with SFC values

getInsertionIdRangePerDimension

public double[] getInsertionIdRangePerDimension(SFCDimensionDefinition[] dimensionDefinitions)

Description copied from interface: HilbertSFCOperations

* Get the range/size of a single insertion ID for each dimension

Specified by:

getInsertionIdRangePerDimension in interface HilbertSFCOperations

Parameters:

dimensionDefinitions - a set of dimension definitions to use to calculate the range of each insertion ID

Returns:

the range of a single insertion ID for each dimension

normalizeRange

public long[] normalizeRange(double minValue,
                             double maxValue,
                             int dimension,
                             SFCDimensionDefinition boundedDimensionDefinition)
                      throws IllegalArgumentException

Description copied from interface: HilbertSFCOperations

Determines the coordinates a given range will span within this space filling curve.

Specified by:

normalizeRange in interface HilbertSFCOperations

Parameters:

minValue - the minimum value

maxValue - the maximum value

dimension - the ordinal of the dimension on this space filling curve

boundedDimensionDefinition - the dimension definition to use to normalize the raw values

Returns:

the range of coordinates (ie. [0] would be the min coordinate and [1] would be the max coordinate)

Throws:

IllegalArgumentException