/**
* @author Evan Fabry
* CS 440
* MP3 - NLP Document Classification/Information Retrieval
*/


import weka.core.converters.ConverterUtils.DataSource;
import weka.core.converters.ConverterUtils.*;
//import weka.core.*;
import weka.core.Instance;
import weka.core.Instances;
import weka.classifiers.functions.*;
import weka.core.converters.AbstractSaver;
import weka.core.converters.AbstractFileSaver;
//import weka.core.converters.ArffSaver;
//import weka.filters.supervised.instance.*;
import weka.filters.supervised.instance.Resample;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.RemoveUseless;
import weka.filters.unsupervised.attribute.RandomProjection;
import weka.core.Utils.*;
import weka.classifiers.Evaluation;
import weka.classifiers.functions.Logistic;
import weka.classifiers.functions.SMO;
import weka.classifiers.functions.MultilayerPerceptron;
//import weka.classifiers.AbstractClassifier;
import weka.classifiers.Classifier;

import java.util.*;
import java.io.*;
import java.lang.Object;
import java.lang.Math;


class LearnerThread extends Thread {
    private Thread t;
    private Instances thread_instances;
    private int learner_type;
    private Classifier cla;

    LearnerThread( int type, Instances inst, Classifier c ){
        this.learner_type = type;
        this.thead_instances = inst;
        this.cla = c;
        File output = new File("out"+Integer.toString(this.learner_type)+".txt");
        System.out.println("Creating " +  Integer.toString(type) );
    }
    public void run() {
        System.out.println("Running " +  Integer.toString(learner_type) );
        try {
            Evaluation eval = new Evaluation(thread_instances);
            eval.crossValidateModel(this.cla, this.thead_instances, 10, new Random(1));
		System.out.println(eval.toSummaryString());
        } catch (InterruptedException e) {
            System.out.println("Thread " +  Integer.toString(this.learner_type) + " interrupted.");
        }
        System.out.println("Thread " +  Integer.toString(this.learner_type) + " exiting.");
    }

    public void start ()
    {
        System.out.println("Starting " +  Integer.toString(this.learner_type));
        if (t == null)
        {
            t = new Thread (this, Integer.toString(this.learner_type));
            t.start ();
        }
    }

}

public class efabry2 {
    public static void main(String [] args) throws Exception
    {
		DataSource tf1_train_source = new DataSource("MP3Part1Solution/data/train.tf1.arff");
    	//DataSource tf2_train_source = new DataSource("MP3Part1Solution/data/train.tf2.arff");
    	//DataSource tf_idf_train_source = new DataSource("MP3Part1Solution/data/train.tfidf.arff");
    	//DataSource idf_train_source = new DataSource("MP3Part1Solution/data/train.idf.arff");

        Instances tf1_train_all = tf1_train_source.getDataSet();
        //Instances tf2_train_all = tf2_train_source.getDataSet();
        //Instances tf_idf_train_all = tf_idf_train_source.getDataSet();
        //Instances idf_train_all = idf_train_source.getDataSet();

        tf1_train_all.setClassIndex(tf1_train_all.numAttributes() - 1);
        //tf2_train_all.setClassIndex(tf2_train_all.numAttributes() - 1);
        //tf_idf_train_all.setClassIndex(tf_idf_train_all.numAttributes() - 1);
        //idf_train_all.setClassIndex(idf_train_all.numAttributes() - 1);
    	for (int k = 0; k < 3; k++) {

    		int percent = 100;
    		if (k == 0) { percent = 20; }
    		else if (k == 1) { percent = 60; }
            System.out.print("starting loop with percent = ");
            System.out.println(percent);

            RemoveUseless ru = new RemoveUseless();
            RandomProjection rp = new RandomProjection();
            Resample rs = new Resample();
            rp.setOptions( weka.core.Utils.splitOptions( "-N 1000" ));
            rs.setOptions( weka.core.Utils.splitOptions( "-Z "+Integer.toString(percent) ));

            rp.setInputFormat(tf1_train_all);
            ru.setInputFormat(tf1_train_all);
			Instances tf1_train = Filter.useFilter(tf1_train_all, ru);
            tf1_train = Filter.useFilter(tf1_train, rp);
            rs.setInputFormat(tf1_train);
            tf1_train = Filter.useFilter(tf1_train, rs);
            //.resample().setOptions( resample_options ).removeUseless().randomProjection().setOptions("-N 1000");
			//tf2_train[k].resample().setOptions( "-Z "+Integer.toString(percent) ).removeUseless().randomProjection().setOptions("-N 1000");
			//tf_idf_train[k].resample().setOptions( "-Z "+Integer.toString(percent) ).removeUseless().randomProjection().setOptions("-N 1000");
			//idf_train[k].resample().setOptions( "-Z "+Integer.toString(percent) ).removeUseless().randomProjection().setOptions("-N 1000");
            System.out.println("evaluating");

            MultilayerPerceptron slp = new MultilayerPerceptron();
            slp.setHiddenLayers("0");

            MultilayerPerceptron mlp = new MultilayerPerceptron();
            mlp.setHiddenLayers("5");

            SMO smo = new SMO();

            Logistic log= new Logisitic();

            LearnerThread T1 = new LearnerThread(0, tf1_train, slp);
            LearnerThread T2 = new LearnerThread(1, tf1_train, mlp);
            LearnerThread T3 = new LearnerThread(2, tf1_train, smo);
            LearnerThread T4 = new LearnerThread(2, tf1_train, log);
            T1.start();
            T2.start();
            T3.start();
            T4.start();

            T1.join(10000);
            T2.join(10000);
            T3.join(10000);
            T4.join(10000);

            //Evaluation eval = new Evaluation(tf1_train);
            //eval.crossValidateModel(slp, tf1_train, 10, new Random(1));
            //System.out.println(eval.toSummaryString());
            //System.out.println(eval.toClassDetailsString());
            //System.out.println(eval.toMatrixString());
            //System.out.println(out.getAttributes());
            System.out.print("ending loop with percent = ");
            System.out.println(percent);
            //tf1_train[k] = NULL;
            //tf2_train[k] = NULL;
            //tf_idf_train[k] = NULL;
            //idf_train[k] = NULL
		}

    	//PreprocessPanel instances = new PreprocessPanel();
    }
}