org.jmol.bspt
Class Bspt

java.lang.Object
  org.jmol.bspt.Bspt

public final class Bspt
extends java.lang.Object

a Binary Space Partitioning Tree

The tree partitions n-dimensional space (in our case 3) into little boxes, facilitating searches for things which are *nearby*.

For some useful background info, search the web for "bsp tree faq". Our application is somewhat simpler because we are storing points instead of polygons.

We are working with three dimensions. For the purposes of the Bspt code these dimensions are stored as 0, 1, or 2. Each node of the tree splits along the next dimension, wrapping around to 0.

    mySplitDimension = (parentSplitDimension + 1) % 3;
  

A split value is stored in the node. Values which are <= splitValue are stored down the left branch. Values which are >= splitValue are stored down the right branch. If searchValue == splitValue then the search must proceed down both branches.

Planar and crystaline substructures can generate values which are == along one dimension.

To get a good picture in your head, first think about it in one dimension, points on a number line. The tree just partitions the points. Now think about 2 dimensions. The first node of the tree splits the plane into two rectangles along the x dimension. The second level of the tree splits the subplanes (independently) along the y dimension into smaller rectangles. The third level splits along the x dimension. In three dimensions, we are doing the same thing, only working with 3-d boxes.

Author:

Miguel, miguel@jmol.org

Constructor Summary

Bspt(int dimMax)
Create a bspt with the specified number of dimensions.

Method Summary

void	addTuple(Tuple tuple) Iterate through all of your data points, calling addTuple
SphereIterator	allocateSphereIterator()
void	stats() prints some simple stats to stdout

Methods inherited from class java.lang.Object

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

Constructor Detail

Bspt

public Bspt(int dimMax)

Create a bspt with the specified number of dimensions. For a 3-dimensional tree (x,y,z) call new Bspt(3).

Parameters:

dimMax -

Method Detail

addTuple

public void addTuple(Tuple tuple)

Iterate through all of your data points, calling addTuple

Parameters:

tuple -

stats

public void stats()

prints some simple stats to stdout

allocateSphereIterator

public SphereIterator allocateSphereIterator()