ch.ethz.globis.tinspin.wrappers

Class Candidate

java.lang.Object

ch.ethz.globis.tinspin.wrappers.Candidate

public abstract class Candidate
extends Object

Constructor Summary

Constructors

Constructor and Description
Candidate()

Method Summary

Modifier and Type	Method and Description
Object	getNative()
String	getNativeStats()
void	getStats(TestStats S)
double	knnQuery(int k, double[] center) Called when the tree should execute the k-nearest-neighbor queries defined by the arguments.
abstract void	load(double[] data, int idxDim)
abstract int	pointQuery(Object qA) Called when the tree should execute the point queries defined by the arguments.
abstract Object	preparePointQuery(double[][] q) This method is called before the actual point query, providing the underlying tree with the possibility to pre-process the query data
abstract int	query(double[] min, double[] max) Called when tree should execute a range query
List<?>	queryToList(double[] min, double[] max)
abstract void	release() This method is invoked at the end of the test run to allow freeing of allocated resources
protected void	run(Candidate x, int N, int DIM)
protected void	runSmokeTest(Candidate x, int dims)
boolean	supportsKNN()
boolean	supportsPointQuery()
boolean	supportsUnload()
boolean	supportsUpdate()
boolean	supportsWindowQuery()
String	toString()
abstract int	unload() Called when the tree should delete all the points it contains.
abstract int	update(double[][] updateTable)

Methods inherited from class java.lang.Object

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

Constructor Detail

Candidate

public Candidate()

Method Detail

release

public abstract void release()

This method is invoked at the end of the test run to allow freeing of allocated resources

load

public abstract void load(double[] data,
                          int idxDim)

preparePointQuery

public abstract Object preparePointQuery(double[][] q)

This method is called before the actual point query, providing the underlying tree with the possibility to pre-process the query data

Parameters:

q - raw query data

Returns:

pre-processed query data

pointQuery

public abstract int pointQuery(Object qA)

Called when the tree should execute the point queries defined by the arguments.

Parameters:

qA - pre-processed query data

Returns:

number of point queries that matched a point

knnQuery

public double knnQuery(int k,
                       double[] center)

Called when the tree should execute the k-nearest-neighbor queries defined by the arguments.

Parameters:

k - number of desired neighbors

center - of query

Returns:

the squared distance to the returned points

unload

public abstract int unload()

Called when the tree should delete all the points it contains. The deletion should happen in N/2 steps where the first and last element (in order of insertion) that are still in the tree should be deleted

Returns:

number of effectively deleted points

query

public abstract int query(double[] min,
                          double[] max)

Called when tree should execute a range query

Parameters:

min - lower corner

max - upper corner

Returns:

number of matched points

update

public abstract int update(double[][] updateTable)

queryToList

public List<?> queryToList(double[] min,
                           double[] max)

run

protected void run(Candidate x,
                   int N,
                   int DIM)

runSmokeTest

protected void runSmokeTest(Candidate x,
                            int dims)

getStats

public void getStats(TestStats S)

supportsWindowQuery

public boolean supportsWindowQuery()

supportsPointQuery

public boolean supportsPointQuery()

supportsKNN

public boolean supportsKNN()

supportsUpdate

public boolean supportsUpdate()

supportsUnload

public boolean supportsUnload()

getNative

public Object getNative()

getNativeStats

public String getNativeStats()

toString

public String toString()

Overrides:

toString in class Object