gistfile1.java

package org.gbif.hadoop.compress;

import java.io.IOException;
import java.io.OutputStream;
import java.util.zip.CRC32;
import java.util.zip.Checksum;
import java.util.zip.Deflater;
import java.util.zip.DeflaterOutputStream;

/**
 * A {@link java.util.zip.DeflaterOutputStream} which will run a {@Deflater} in <i>no wrap</i> mode so no header or
 * footer are written to the stream, provides the length of the compressed data and CRC32 and length of the uncompressed
 * data.
 */
public class SplittableDeflaterOutputStream extends DeflaterOutputStream {

  private final Checksum checksum = new CRC32(); // trace uncompressed data

  public SplittableDeflaterOutputStream(OutputStream out) {
    this(out, 1024);
  }

  public SplittableDeflaterOutputStream(OutputStream out, int bufferSize) {
    // deflater set to nowrap, and using SYNC_MODE
    super(out, new Deflater(Deflater.BEST_COMPRESSION, true), bufferSize, true);
  }

  @Override
  public void write(int b) throws IOException {
    super.write(b);
    checksum.update(b);
  }

  @Override
  public void write(byte[] b, int off, int len) throws IOException {
    super.write(b, off, len);
    checksum.update(b, off, len);
  }

  /**
   * Does nothing, thus no footers are written to the stream unlike the parent.
   */
  @Override
  public void finish() throws IOException {
  }

  /**
   * Ends the delater without calling finish() so no footers are written, and then flushes and closes the underlying
   * stream.
   */
  @Override
  public void close() throws IOException {
    def.end(); // super will not do this
    flush();
    super.close(); // super will call finish, but we overload that to do nothing
  }

  /**
   * @return The CRC of the uncompressed data that was written
   */
  public long getCRC32() {
    return checksum.getValue();
  }

  /**
   * @return The length of the uncompressed data
   */
  public long getUncompressedLength() {
    return def.getBytesRead();
  }

  /**
   * @return The length of the compressed data
   */
  public long getCompressedLength() {
    return def.getBytesWritten();
  }
}