/// <summary>
/// Constructs a new instance of the KDNode.
/// </summary>
/// <param name="key">A Hyper Point representing the key to use for storing this value</param>
/// <param name="value">A valid object value to use for copying this.</param>
/// <remarks>The constructor is used only by class; other methods are static</remarks>
private KdNode(HPoint key, object value)
{
_k = key;
_v = value;
_left = null;
_right = null;
_isDeleted = false;
}
/// <summary>
/// This method was written by Ted Dunsford by restructuring the nearest neighbor
/// algorithm presented by Andrew and Bjoern
/// </summary>
/// <param name="kd">Since this is recursive, this represents the current node</param>
/// <param name="target">The target is the HPoint that we are trying to calculate the farthest distance from</param>
/// <param name="hr">In this case, the hr is the hyper rectangle bounding the region that must contain the furthest point.</param>
/// <param name="maxDistSqd">The maximum distance that we have calculated thus far, and will therefore be testing against.</param>
/// <param name="lev">The integer based level of that we have recursed to in the tree</param>
/// <param name="k">The dimensionality of the kd tree</param>
/// <param name="fnl">A list to contain the output, prioritized by distance</param>
public static void FarthestNeighbor(KdNode kd, HPoint target, HRect hr,
double maxDistSqd, int lev, int k,
FarthestNeighborList fnl)
{
// 1. if kd is empty then set dist-sqd to infinity and exit.
if (kd == null)
{
return;
}
// 2. s := split field of kd
int s = lev % k;
// 3. pivot := dom-elt field of kd
HPoint pivot = kd._k;
double pivotToTarget = HPoint.SquareDistance(pivot, target);
// 4. Cut hr into to sub-hyperrectangles left-hr and right-hr.
// The cut plane is through pivot and perpendicular to the s
// dimension.
HRect leftHr = hr; // optimize by not cloning
HRect rightHr = hr.Copy();
leftHr._max[s] = pivot[s];
rightHr._min[s] = pivot[s];
// 5. target-in-left := target_s <= pivot_s
bool targetInLeft = target[s] < pivot[s];
KdNode nearerKd;
HRect nearerHr;
KdNode furtherKd;
HRect furtherHr;
// 6. if target-in-left then
// 6.1. nearer-kd := left field of kd and nearer-hr := left-hr
// 6.2. further-kd := right field of kd and further-hr := right-hr
if (targetInLeft)
{
nearerKd = kd._left;
nearerHr = leftHr;
furtherKd = kd._right;
furtherHr = rightHr;
}
//
// 7. if not target-in-left then
// 7.1. nearer-kd := right field of kd and nearer-hr := right-hr
// 7.2. further-kd := left field of kd and further-hr := left-hr
else
{
nearerKd = kd._right;
nearerHr = rightHr;
furtherKd = kd._left;
furtherHr = leftHr;
}
// 8. Recursively call Nearest Neighbor with paramters
// (nearer-kd, target, nearer-hr, max-dist-sqd), storing the
// results in nearest and dist-sqd
//FarthestNeighbor(nearer_kd, target, nearer_hr, max_dist_sqd, lev + 1, K, nnl);
// This line changed by Ted Dunsford to attempt to find the farther point
FarthestNeighbor(furtherKd, target, furtherHr, maxDistSqd, lev + 1, k, fnl);
//KDNode nearest = (KDNode)nnl.Highest;
double distSqd;
if (!fnl.IsCapacityReached)
{
//dist_sqd = 1.79769e+30;// Double.MaxValue;
distSqd = 0;
}
else
{
distSqd = fnl.MinimumPriority;
}
// 9. max-dist-sqd := minimum of max-dist-sqd and dist-sqd
//max_dist_sqd = Math.Min(max_dist_sqd, dist_sqd);
maxDistSqd = Math.Max(maxDistSqd, distSqd);
// 10. A nearer point could only lie in further-kd if there were some
// part of further-hr within distance sqrt(max-dist-sqd) of
// target. If this is the case then
// HPoint closest = further_hr.Closest(target);
HPoint furthest = nearerHr.Farthest(target);
//if (HPoint.EuclideanDistance(closest, target) < Math.Sqrt(max_dist_sqd))
if(HPoint.EuclideanDistance(furthest, target) > Math.Sqrt(maxDistSqd))
{
// 10.1 if (pivot-target)^2 < dist-sqd then
if (pivotToTarget > distSqd)
{
// 10.1.1 nearest := (pivot, range-elt field of kd)
//nearest = kd;
// 10.1.2 dist-sqd = (pivot-target)^2
distSqd = pivotToTarget;
// add to nnl
if (!kd.IsDeleted)
{
fnl.Insert(kd, distSqd);
}
// 10.1.3 max-dist-sqd = dist-sqd
// max_dist_sqd = dist_sqd;
if (fnl.IsCapacityReached)
{
maxDistSqd = fnl.MinimumPriority;
}
else
{
// max_dist_sqd = 1.79769e+30;//Double.MaxValue;
maxDistSqd = 0;
}
}
// 10.2 Recursively call Nearest Neighbor with parameters
// (further-kd, target, further-hr, max-dist_sqd),
// storing results in temp-nearest and temp-dist-sqd
FarthestNeighbor(nearerKd, target, nearerHr, maxDistSqd, lev + 1, k, fnl);
KdNode tempFarthest = (KdNode)fnl.Farthest;
double tempDistSqd = fnl.MinimumPriority;
// 10.3 If tmp-dist-sqd < dist-sqd then
if (tempDistSqd > distSqd)
{
// 10.3.1 nearest := temp_nearest and dist_sqd := temp_dist_sqd
distSqd = tempDistSqd;
}
}
// SDL: otherwise, current point is nearest
else if (pivotToTarget < maxDistSqd)
{
distSqd = pivotToTarget;
}
}
/// <summary>
/// Searches for values in a range
/// </summary>
/// <param name="lowk"></param>
/// <param name="uppk"></param>
/// <param name="t"></param>
/// <param name="lev"></param>
/// <param name="k"></param>
/// <param name="v"></param>
/// <remarks>Method rsearch translated from 352.range.c of Gonnet and Baeza-Yates</remarks>
public static void SearchRange(HPoint lowk, HPoint uppk, KdNode t, int lev,
int k, List<KdNode> v)
{
if (t == null) return;
if (lowk[lev] <= t._k[lev])
{
SearchRange(lowk, uppk, t._left, (lev + 1) % k, k, v);
}
int j;
for (j = 0; j < k && lowk[j] <= t._k[j] &&
uppk[j] >= t._k[j]; j++)
{
}
if (j == k) v.Add(t);
if (uppk[lev] > t._k[lev])
{
SearchRange(lowk, uppk, t._right, (lev + 1) % k, k, v);
}
}
/// <summary>
/// Searches for a specific value
/// </summary>
/// <param name="key"></param>
/// <param name="t"></param>
/// <param name="k"></param>
/// <returns></returns>
/// <remarks>Method srch translated from 352.srch.c of Gonnet and Baeza-Yates</remarks>
public static KdNode Search(HPoint key, KdNode t, int k)
{
for (int lev = 0; t != null; lev = (lev + 1) % k)
{
if (!t.IsDeleted && key.Equals(t._k))
{
return t;
}
t = key[lev] > t._k[lev] ? t._right : t._left;
}
return null;
}
/// <summary>
/// Inserts a
/// </summary>
/// <param name="key"></param>
/// <param name="value"></param>
/// <param name="t"></param>
/// <param name="lev"></param>
/// <param name="k"></param>
/// <returns></returns>
/// <remarks> Method ins translated from 352.ins.c of Gonnet and Baeza-Yates</remarks>
public static KdNode Insert(HPoint key, object value, KdNode t, int lev, int k)
{
if (t == null)
{
t = new KdNode(key, value);
}
else if (key.Equals(t._k))
{
// "re-insert"
if (t.IsDeleted)
{
t.IsDeleted = false;
t._v = value;
}
else
{
throw (new KeyDuplicateException());
}
}
else if (key[lev] > t._k[lev])
{
t._right = Insert(key, value, t._right, (lev + 1) % k, k);
}
else
{
t._left = Insert(key, value, t._left, (lev + 1) % k, k);
}
return t;
}
/// <summary>
/// Creates a new tree with the specified number of dimensions.
/// </summary>
/// <param name="k">An integer value specifying how many ordinates each key should have.</param>
public KDTree(int k)
{
if (k < 0) throw new NegativeArgumentException("k");
_k = k;
_root = null;
}
/// <summary>
/// Insert a node into the KD-tree.
/// </summary>
/// <param name="key">The array of double valued keys marking the position to insert this object into the tree</param>
/// <param name="value">The object value to insert into the tree</param>
/// <exception cref="KeySizeException"> if key.length mismatches the dimension of the tree (K)</exception>
/// <exception cref="KeyDuplicateException"> if the key already exists in the tree</exception>
/// <remarks>
/// Uses algorithm translated from 352.ins.c of
/// @Book{GonnetBaezaYates1991,
/// author = {G.H. Gonnet and R. Baeza-Yates},
/// title = {Handbook of Algorithms and Data Structures},
/// publisher = {Addison-Wesley},
/// year = {1991}
/// </remarks>
public void Insert(double[] key, object value)
{
if (key.Length != _k) throw new KeySizeException();
_root = KdNode.Insert(new HPoint(key), value, _root, 0, _k);
_count++;
}
