@Override
protected void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
String line = value.toString();
StringTokenizer tokenizer = new StringTokenizer(line, " ");
int dim = tokenizer.countTokens();
double[] coords = new double[dim];
for (int i = 0; i < dim; i++) {
coords[i] = Double.valueOf(tokenizer.nextToken());
}
Point point = new Point(coords);
Centroid nearest = null;
double nearestDistance = Double.MAX_VALUE;
for (Centroid c : centers) {
double dist = point.euclideanDistance(c);
if (nearest == null) {
nearest = c;
nearestDistance = dist;
} else {
if (dist < nearestDistance) {
nearest = c;
nearestDistance = dist;
}
}
}
context.write(nearest, point);
}
public Point deepCopy() {
Point out = new Point();
out.n = this.n;
out.coord = new double[out.n];
for (int i = 0; i < out.n; i++) {
out.coord[i] = this.coord[i];
}
return out;
}
public void map(
LongWritable key, Point value, OutputCollector<LongWritable, Point> output, Reporter reporter)
throws IOException {
double min = value.sumOfSquares(centers.get(0));
int best = 0;
for (int index = 1; index < numberOfCenters; ++index) {
double current = value.sumOfSquares(centers.get(index));
if (current < min) {
min = current;
best = index;
}
}
reporter.incrCounter("NUMBER", "NODES", 1);
reporter.incrCounter("CENTER", "" + best, 1);
output.collect(new LongWritable(best), value);
}
@Override
protected void map(Centroid key, Point value, Context context)
throws IOException, InterruptedException {
Centroid nearest = null;
double nearestDistance = Double.MAX_VALUE;
for (Centroid c : centers) {
double dist = value.euclideanDistance(c);
if (nearest == null) {
nearest = c;
nearestDistance = dist;
} else {
if (dist < nearestDistance) {
nearest = c;
nearestDistance = dist;
}
}
}
context.write(nearest, value);
}
@Override
protected void reduce(Centroid key, Iterable<Point> values, Context context)
throws IOException, InterruptedException {
ArrayList<Point> points = new ArrayList<Point>();
points.clear();
int clusterId = key.id;
Point newCenter = null;
for (Point p : values) {
Point copy = p.deepCopy();
points.add(copy);
if (newCenter == null) {
newCenter = new Point(copy.deepCopy());
} else {
newCenter = newCenter.add(copy);
}
}
newCenter = newCenter.div(points.size());
Centroid center = new Centroid(clusterId, newCenter);
centers.add(center);
for (Point p : points) {
context.write(center, p);
}
}
@Override
protected void map(Centroid key, Point value, Context context)
throws IOException, InterruptedException {
String out = key.toString() + " " + value.toString();
context.write(new Text(out), new Text(""));
}
@Override
public void readFields(DataInput dataInput) throws IOException {
super.readFields(dataInput);
this.id = dataInput.readInt();
}
@Override
public void write(DataOutput dataOutput) throws IOException {
super.write(dataOutput);
dataOutput.writeInt(id);
}
