Separate the dendrogram printing from clustering.

Author: mdimura

src/averagecluster.cpp

@@ -257,7 +257,7 @@ bool AverageCluster::getCandidate(float &minInexact, float &minExact, AverageVer
 	return (minExact < 1.0f) && (minExact < minInexact || fabsf(minExact - minInexact) < EPSILON);
 }
 
-void AverageCluster::clusterMatrix(InMatrix * mat) {
+Dendrogram AverageCluster::clusterMatrix(InMatrix * mat) {
 
 	uint maxEdges = mat->numPoints; // maxEdges = num of outEdges can be stored in RAM
 
@@ -315,5 +315,6 @@ void AverageCluster::clusterMatrix(InMatrix * mat) {
 
 	fprintf(stderr, "Finished! number of edges = %u. linear ratio: %.2f. quadratic ratio: %zu\n", maxEdges,
 			(double) maxEdges / mat->numPoints, mat->numElements / maxEdges);
+	return dendrogram;
 }
 

src/averagecluster.h

@@ -68,16 +68,16 @@ struct AverageCluster: public Cluster {
 	float tempMax;
 	uint totalEdges;
 
-	AverageCluster(uint n, char * treeName) :
-			Cluster(n, treeName) {
+	AverageCluster(uint n) :
+			Cluster(n) {
 		totalEdges = 0;
 		tempMax = 0;
 	}
 
 	// overrider super methods
 	Vertex* createVertex(uint id);
 
-	void clusterMatrix(InMatrix * mat);
+	Dendrogram clusterMatrix(InMatrix * mat);
 
 	// current class methods
 	void merge(AverageVertex * v1, AverageVertex * v2, AverageVertex* v, float max);

src/cluster.cpp

@@ -19,13 +19,9 @@ void Vertex::updateAncestor(Vertex* ancestor) {
 
 //----------------------------------------------------------------------------------------------------
 
-Cluster::Cluster(uint n, char * treeName) :
-		numLeaves(n) {
+Cluster::Cluster(uint n) :
+		numLeaves(n),dendrogram(n){
 	// add the leaves into the final graph
-	phytree.open(treeName, ofstream::binary);
-	if (!phytree)
-		EXIT_MSG(treeName);
-	phytree.precision(5);
 }
 
 void Cluster::createLeaves() {
@@ -38,15 +34,10 @@ Cluster::~Cluster() {
 
 	for (uint i = 0; i < vertices.size(); ++i)
 		delete vertices[i];
-	phytree.close();
 
 }
 
 void Cluster::updateGraph(uint id1, uint id2, float dist) {
-
-	if (id1 < id2)
-		phytree << id1 + 1 << " " << id2 + 1 << " " << dist << endl;
-	else
-		phytree << id2 + 1 << " " << id1 + 1 << " " << dist << endl;
+		dendrogram.add(id1 ,id2 ,dist);
 }
 

src/cluster.h

@@ -12,6 +12,7 @@
 #include <list>
 #include <fstream>
 #include "inmatrix.h"
+#include "dendrogram.h"
 using namespace std;
 
 struct Vertex {
@@ -37,11 +38,12 @@ struct Cluster {
 protected:
 	vector<Vertex *> vertices;
 	uint numLeaves;
-
+	Dendrogram dendrogram;
 private:
-	ofstream phytree;
+	//ofstream phytree;
+
 public:
-	Cluster(uint n, char * treeName);
+	Cluster(uint n);
 
 	virtual ~Cluster();
 
@@ -52,7 +54,7 @@ struct Cluster {
 	// overriden methods
 	virtual Vertex* createVertex(uint id) = 0;
 
-	virtual void clusterMatrix(InMatrix* mat) = 0;
+	virtual Dendrogram clusterMatrix(InMatrix* mat) = 0;
 
 	// non-overriden methods
 	void createLeaves();

src/completecluster.cpp

@@ -112,7 +112,7 @@ void CompleteCluster::printAllEdges() {
 	}
 }
 
-void CompleteCluster::clusterMatrix(InMatrix * mat) {
+Dendrogram CompleteCluster::clusterMatrix(InMatrix * mat) {
 
 	uint maxEdges = mat->numPoints; // maxNumEdges = num of links can be stored in RAM
 	uint newId = numLeaves;
@@ -127,7 +127,7 @@ void CompleteCluster::clusterMatrix(InMatrix * mat) {
 		v2 = reinterpret_cast<CompleteVertex*>(vertices[col]->ancestor);
 
 		if (v1 == v2)
-			return;
+			return dendrogram;
 
 		if (v1->id < v2->id) {
 			tempV = v1;
@@ -158,5 +158,6 @@ void CompleteCluster::clusterMatrix(InMatrix * mat) {
 
 	fprintf(stderr, "Finished! number of edges = %u. linear ratio: %.2f. quadratic ratio: %zu\n", maxEdges,
 			(double) maxEdges / mat->numPoints, mat->numElements / maxEdges);
+	return dendrogram;
 }
 

src/completecluster.h

@@ -42,15 +42,15 @@ struct CompleteCluster: public Cluster {
 
 	uint totalEdges;
 
-	CompleteCluster(uint matDim, char * treeName) :
-			Cluster(matDim, treeName) {
+	CompleteCluster(uint matDim) :
+			Cluster(matDim) {
 		totalEdges = 0;
 	}
 
 	// overrider super methods
 	Vertex* createVertex(uint id);
 
-	void clusterMatrix(InMatrix * mat);
+	Dendrogram clusterMatrix(InMatrix * mat);
 
 	// current class methods
 	void merge(CompleteVertex * v1, CompleteVertex * v2, CompleteVertex *v);

src/dendrogram.cpp

@@ -0,0 +1,15 @@
+#include "dendrogram.h"
+
+void Dendrogram::add(unsigned p1, unsigned p2, float distance)
+{
+	Link l;
+	l.p1=p1;
+	l.p2=p2;
+	l.distance=distance;
+	l.nump=0;
+	l.nump+=p1>=numPoints?pointsCount.at(p1-numPoints):1;
+	l.nump+=p2>=numPoints?pointsCount.at(p2-numPoints):1;
+	pointsCount.push_back(l.nump);
+	links.push_back(l);
+}
+

src/dendrogram.h

@@ -0,0 +1,42 @@
+#ifndef DENDROGRAM_H_
+#define DENDROGRAM_H_
+/*
+ * A 4 by (n−1) matrix Z is returned.
+ * At the i-th iteration, clusters with indices Z[i, 0] and Z[i, 1] are combined
+ * to form cluster n+i. A cluster with an index less than n corresponds to one
+ * of the n original observations. The distance between clusters Z[i, 0] and Z[i, 1]
+ * is given by Z[i, 2]. The fourth value Z[i, 3] represents the number of original
+ * observations in the newly formed cluster.
+ *
+ */
+#include <vector>
+#include <unordered_map>
+#include <fstream>
+class Dendrogram {
+private:
+	struct Link {
+		unsigned p1,p2;
+		float distance;
+		unsigned nump;
+	};
+	const unsigned numPoints;
+	std::vector<Link> links;
+	std::vector<unsigned> pointsCount; //number of points in the cluster
+public:
+	Dendrogram(unsigned numPoints):numPoints(numPoints)
+	{
+		links.reserve(numPoints-1);
+		pointsCount.reserve(numPoints-1);
+	}
+	void add(unsigned p1, unsigned p2, float distance);
+	void print(const std::string& fname) const
+	{
+		std::ofstream myfile;
+		myfile.open (fname);
+		for(const Link& l:links) {
+			myfile<<l.p1<<' '<<l.p2<<' '<<l.distance<<' '<<l.nump<<'\n';
+		}
+	}
+};
+
+#endif //DENDROGRAM_H_

src/inmatrix.h

@@ -46,7 +46,7 @@ struct InMatrix: public Matrix {
 	void push(uint row, uint col, float value) {
 		Element e(row,col,value);
 		sorter.push(e);
-		threshold=threshold<value?value:threshold;
+		threshold=threshold<value?nextafterf(value,value+100.0f):threshold;
 	}
 	void sort() {
 		sorter.sort();

src/main.cpp

@@ -17,9 +17,12 @@ struct Point {
 		for(int i=0; i<3; i++) {
 			r[i]=unif(re);
 		}
-		for(int i=3; i<n; i++) {
+		for(int i=3; i<n-3; i++) {
 			r[i]=r[i-3]+0.1;
 		}
+		for(int i=n-3; i<n; i++) {
+			r[i]=unif(re);
+		}
 	}
 	static const unsigned n=3*150;
 	const float invn=3.0f/n;
@@ -102,16 +105,16 @@ int main(int argc, char **argv) {
 	Cluster * cluster = NULL;
 
 	if (linkage == 0)
-		cluster = new AverageCluster(inMat.getNumPints(), treeFileName);
+		cluster = new AverageCluster(inMat.getNumPints());
 	else if (linkage == 1)
-		cluster = new CompleteCluster(inMat.getNumPints(), treeFileName);
+		cluster = new CompleteCluster(inMat.getNumPints());
 	else if (linkage == 2)
-		cluster = new SingleCluster(inMat.getNumPints(), treeFileName);
+		cluster = new SingleCluster(inMat.getNumPints());
 	else if (linkage == 3)
-		cluster = new AverageCluster(inMat.getNumPints(), treeFileName);
+		cluster = new AverageCluster(inMat.getNumPints());
 
 	cluster->createLeaves();
-	PROFILE("cluster", cluster->clusterMatrix(&inMat));
+	PROFILE("cluster", cluster->clusterMatrix(&inMat).print(treeFileName));
 
 	delete cluster;
 

src/singlecluster.cpp

@@ -14,7 +14,7 @@ Vertex* SingleCluster::createVertex(uint id) {
 
 //----------------------------------------------------------------------------------------------------
 
-void SingleCluster::clusterMatrix(InMatrix * mat) {
+Dendrogram SingleCluster::clusterMatrix(InMatrix * mat) {
 
 	uint newId = numLeaves, maxVertices = numLeaves * 2 - 1;
 	uint row, col;
@@ -47,4 +47,5 @@ void SingleCluster::clusterMatrix(InMatrix * mat) {
 	}
 
 	fprintf(stderr, "Finished! index = %u\n", mat->pos);
+	return dendrogram;
 }

src/singlecluster.h

@@ -13,14 +13,14 @@ using namespace std;
 
 struct SingleCluster: public Cluster {
 
-	SingleCluster(uint matDim, char * treeName) :
-			Cluster(matDim, treeName) {
+	SingleCluster(uint matDim) :
+			Cluster(matDim) {
 	}
 
 	Vertex* createVertex(uint id);
 
 	// overrider super methods
-	void clusterMatrix(InMatrix * mat);
+	Dendrogram clusterMatrix(InMatrix * mat);
 };
 
 #endif /* MINCLUSTER_H_ */