gr.eap.LSHDB.HammingKey.java Source code

Introduction

Here is the source code for gr.eap.LSHDB.HammingKey.java

Source

/*
 * To change this license header, choose License Headers in Project Properties.
 * To change this template file, choose Tools | Templates
 * and open the template in the editor.
 */
package gr.eap.LSHDB;

import gr.eap.LSHDB.embeddables.BloomFilter;
import java.security.InvalidParameterException;
import java.util.Random;
import org.apache.commons.math3.distribution.BinomialDistribution;
import org.apache.log4j.Logger;

/**
 *
 * @author dimkar
 */
public class HammingKey extends Key {
    final static Logger log = Logger.getLogger(HammingKey.class);

    public static final long serialVersionUID = 501L;

    public int t;
    public int[][] samples;
    String[] tokens;

    public HammingKey(String keyFieldName, int k, double delta, int t, boolean tokenized,
            boolean performComparisons, Embeddable emb) {
        this.keyFieldName = keyFieldName;
        this.k = k;
        this.delta = delta;
        this.t = t;
        this.size = emb.getSize();
        optimizeL();
        this.tokenized = tokenized;
        log.info("Number of hash tables generated L=" + this.L + " using k=" + this.k + " and size=" + this.size);
        this.samples = new int[this.L][this.k];
        initSamples();
        this.performComparisons = performComparisons;
        //if (emb==null)
        //  throw new StoreInitException("Embeddable object cannot be null.");
        setEmbeddable(emb);
    }

    public HammingKey(String keyFieldName) {
        this(keyFieldName, 30, .1, 75, true, true, new BloomFilter(700, 15, 2));
    }

    public HammingKey(int k, double delta, int t, int size) {
        this.k = k;
        this.delta = delta;
        this.t = t;
        this.size = size;
        optimizeL();
    }

    @Override
    public int optimizeL() {
        L = (int) Math.ceil(Math.log(delta) / Math.log(1 - Math.pow((1.0 - (t * 1.0 / this.size)), k)));
        return L;
    }

    @Override
    public Key create(double thresholdRatio) {
        if (thresholdRatio == 1.0)
            thresholdRatio = 1.23;
        int t = (int) Math.round(this.t * thresholdRatio);
        return new HammingKey(k, delta, t, emb.getSize());
    }

    public int getLc() {
        double p = 1 - (t * 1.0) / (this.size * 1.0);
        p = Math.pow(p, k);
        double exp = (L * p);
        double std = Math.sqrt(exp * (1 - p));
        int C = (int) Math.round(exp - std);

        double x = (Math.sqrt(Math.log(delta) * Math.log(delta) - 2 * C * Math.log(delta)) - Math.log(delta) + C)
                / p;
        int Lc = (int) Math.ceil(x);
        double b = Lc * p;
        if (C > b)
            System.out.println("does not apply C > np.");
        BinomialDistribution bd1 = new BinomialDistribution(L, p);
        for (int l = L; l < L * 2; l++) {
            bd1 = new BinomialDistribution(l, p);
            double result = bd1.cumulativeProbability(C - 1);
            if (result < delta) {
                Lc = l;
                break;
            }
        }
        System.out.println("Lc reduced to=" + Lc);
        return Lc;
    }

    public void initSamples() {
        Random r = new Random(System.currentTimeMillis());
        for (int j = 0; j < this.L; j++) {
            for (int k = 0; k < this.k; k++) {
                samples[j][k] = r.nextInt(this.size);
            }
        }
    }
}