com.github.jparkie.deduplicator.impl.BSBFSDDeDuplicatorSerializer Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of DeDuplicator Show documentation

Advanced Bloom Filter Based Algorithms for Efficient Approximate Data De-Duplication in Streams

The newest version!

package com.github.jparkie.deduplicator.impl;

import com.github.jparkie.deduplicator.BitArray;
import com.github.jparkie.deduplicator.ProbabilisticDeDuplicatorSerializer;

import java.io.*;

public class BSBFSDDeDuplicatorSerializer implements ProbabilisticDeDuplicatorSerializer {
    @Override
    public byte version() {
        return 1;
    }

    @Override
    public void writeTo(BSBFSDDeDuplicator probabilisticDeDuplicator, OutputStream out) throws IOException {
        final DataOutputStream dos = new DataOutputStream(out);
        dos.writeByte(version());
        dos.writeLong(probabilisticDeDuplicator.numBits);
        dos.writeInt(probabilisticDeDuplicator.numHashFunctions);
        for (BitArray bloomFilter : probabilisticDeDuplicator.bloomFilters) {
            bloomFilter.writeTo(dos);
        }
    }

    @Override
    public BSBFSDDeDuplicator readFrom(InputStream in) throws IOException {
        final DataInputStream dis = new DataInputStream(in);
        final byte serializedVersion = dis.readByte();
        if (serializedVersion != version()) {
            final String error = String.format("Unexpected ProbabilisticDeDuplicator version number (%d), expected %d", serializedVersion, version());
            throw new IllegalArgumentException(error);
        }
        final long numBits = dis.readLong();
        final int numHashFunctions = dis.readInt();
        final BitArray[] bloomFilters = new BitArray[numHashFunctions];
        for (int index = 0; index < numHashFunctions; index++) {
            bloomFilters[index] = BitArray.readFrom(dis);
        }
        return new BSBFSDDeDuplicator(numBits, numHashFunctions, bloomFilters);
    }
}