package RobustHMM;
/*
* Copyright (C) 2019 Evan Tarbell and Tao Liu
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.
*/
import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.FileReader;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Scanner;
import net.sf.javaml.core.DenseInstance;
import org.apache.commons.math3.stat.descriptive.moment.Mean;
import org.apache.commons.math3.stat.descriptive.moment.Variance;
import be.ac.ulg.montefiore.run.jahmm.Hmm;
import be.ac.ulg.montefiore.run.jahmm.ObservationReal;
import be.ac.ulg.montefiore.run.jahmm.ObservationVector;
import be.ac.ulg.montefiore.run.jahmm.OpdfGaussian;
import be.ac.ulg.montefiore.run.jahmm.OpdfMultiGaussian;
import be.ac.ulg.montefiore.run.jahmm.io.HmmBinaryWriter;
public class RandomInitHMM {
//class variables
private String input;
private int numStates;
private Hmm<?> hmm;
//Main method variables
private static int states = 0;
private static String file = null;
private static String output = null;
public RandomInitHMM(String i,int n) throws FileNotFoundException{
input = i;
numStates = n;
read();
}
public Hmm<?> getHMM(){return hmm;}
private void read() throws FileNotFoundException{
ObservationReader reader = new ObservationReader(input);
String method = reader.getMethod();
if (method.equals("Vector")){
hmm = readVector(reader);
}
else if(method.equals("Real") ){
hmm = readReal(reader);
}
}
private void read2() throws FileNotFoundException{
Scanner inFile =new Scanner ((Readable) new FileReader(input));
ArrayList<double[]> data = new ArrayList<double[]>();
int numFeat = 0;
while (inFile.hasNext()){
String line = inFile.nextLine();
String[] features = line.split(",");
double[] values = new double[features.length];
numFeat = features.length;
for (int i = 0;i < features.length;i++){
values[i] = Double.parseDouble(features[i]);
}
data.add(values);
}
double[][] values = new double[data.size()][numFeat];
for(int i = 0;i < data.size();i++){
double[] temp = data.get(i);
for (int a = 0;a < temp.length;a++){
values[i][a] = temp[a];
}
}
data = null;
double[] mu = new double[numFeat];
double[] var = new double[numFeat];
Mean mean = new Mean();
Variance variance = new Variance();
for (int i = 0; i < numFeat;i++){
mu[i] = mean.evaluate(values[i]);
var[i] = variance.evaluate(values[i]);
}
List<OpdfMultiGaussian> opdf = new ArrayList<OpdfMultiGaussian>();
for (int i = 0;i < numStates;i++){
}
}
@SuppressWarnings({"unchecked"})
private Hmm<ObservationVector> readVector(ObservationReader reader){
ArrayList<ObservationVector> obs = (ArrayList<ObservationVector>) reader.getObs();
double[] initial = setInitial();
double[][] trans = setTrans();
List<OpdfMultiGaussian> opdf = setVectorPDF(obs);
Hmm<ObservationVector> h = new Hmm<ObservationVector>(initial, trans, opdf);
return h;
}
@SuppressWarnings({ "unused", "unchecked" })
private Hmm<ObservationReal> readReal(ObservationReader reader){
ArrayList<ObservationReal> obs = (ArrayList<ObservationReal>) reader.getObs();
double[] initial = setInitial();
double[][] trans = setTrans();
//TODO: write opdf maker for monovariate gaussian
return null;
}
private List<OpdfGaussian> setRealPDF(ArrayList<ObservationReal> obs){
//TODO: write this method
return null;
}
private List<OpdfMultiGaussian> setVectorPDF(ArrayList<ObservationVector> obs){
List<OpdfMultiGaussian> opdf = new ArrayList<OpdfMultiGaussian>();
ObservationVector o = obs.get(0);
int dim = o.dimension();
double[] means = new double[dim];
Mean mu = new Mean();
Variance v = new Variance();
double[] vars = new double[dim];
double[] vals = new double[obs.size()];
for (int i = 0;i < dim;i++){
for (int a = 0;a < obs.size();a++){
vals[a] = obs.get(a).value(i);
}
means[i] = mu.evaluate(vals, 0, vals.length);
vars[i] = v.evaluate(vals);
}
double[][] Means = new double[numStates][dim];
double[][] Vars = new double[numStates][dim];
for (int i = 0; i < means.length;i++){
double mean = means[i];
double var = vars[i];
double sd = Math.sqrt(var);
double meanLower = mean - (3.0*sd);
double meanUpper = mean + (3.0*sd);
double varLower = 0.5 * var;
double varUpper = 3 * var;
double meanStep = (meanUpper - meanLower) / ((double)numStates - 1);
double varStep = (varUpper - varLower) / ((double) numStates - 1);
//System.out.println((meanUpper-meanLower)+"\t"+meanStep);
for(int a = 0;a < numStates;a++){
Means[a][i] = meanLower + (a * meanStep);
Vars[a][i] = varLower + (a * varStep);
//System.out.println(Means[a][i]);
//System.out.println(Vars[a][i]);
}
}
for (int i = 0;i < Means.length;i++){
//System.out.println(Means.length);
double[][] cov = new double[dim][dim];
for (int a = 0;a < Means[i].length;a++){
//System.out.println(Means[i].length);
cov[a][a] = Vars[i][a];
//System.out.println(cov[a][a]);
}
OpdfMultiGaussian pdf = new OpdfMultiGaussian(Means[i],cov);
opdf.add(pdf);
}
return opdf;
}
private double[][] setTrans(){
double[][] trans = new double[numStates][numStates];
for (int i = 0;i < numStates;i++){
for(int a = 0;a < numStates;a++){
trans[i][a] = (double) 1/numStates;
}
}
return trans;
}
private double[] setInitial(){
double[] initial = new double[numStates];
for(int i = 0;i < numStates;i++){
initial[i] = (double) 1/numStates;
}
return initial;
}
public static void main(String[] args) throws IOException{
for (int i = 0; i < args.length; i++) {
switch (args[i].charAt((1))) {
case'i':
file = (args[i+1]);
i++;
break;
case 'n':
states = Integer.parseInt(args[i+1]);
i++;
break;
case'o':
output = args[i+1];
i++;
break;
}
}
if (file == null || states == 0 || output == null){
printUsage();
System.exit(1);
}
RandomInitHMM init = new RandomInitHMM(file,states);
Hmm<?> hmm = init.getHMM();
System.out.println(hmm.toString());
FileOutputStream out = new FileOutputStream(output);
HmmBinaryWriter writer = new HmmBinaryWriter();
writer.write(out, hmm);
}
private static void printUsage(){
System.out.println("Usage: java -jar RandomInitHMM.jar");
System.out.println("Required Parameters:");
System.out.println("-i <File> Observation File in proper format");
System.out.println("-n <int> Number of States");
System.out.println("-o <File> Output file to write binary HMM");
}
}
