/**
 * 
 */
package at.gv.egiz.pdfas.framework.input;

import java.io.InputStream;

/**
 * The input document data source.
 * 
 * <p>
 * Usually this is a PdfDataSource, but it may be a TextDataSource as well.
 * </p>
 * 
 * @author wprinz
 * 
 */
public interface DataSource
{
  /**
   * Creates a new InputStream that allows to read out the document's binary
   * data from the beginning.
   * 
   * @return Returns the InputStream with the binary data.
   */
  public InputStream createInputStream();

  /**
   * Returns the length (number of bytes) of the stream.
   * 
   * @return Returns the length (number of bytes) of the stream.
   */
  public int getLength();

  /**
   * Returns the data of this DataSource as a byte array.
   * 
   * <p>
   * Calling this method indicates that you need a byte array for random <strong>read</strong> access.
   * The DataSource implementation should of course cache this byte array to avoid too much memory usage.
   * </p>
   * <p>
   * Performance analysis has shown that the libraries internally convert the streams to byte arrays and
   * that file system access is very slow.
   * </p>
   * 
   * @return
   */
  public byte [] getAsByteArray();
}