Class DocumentContentImpl

src/main/java/gate/corpora/DocumentContentImpl.java:34–210

Represents the commonalities between all sorts of document contents.

/** Represents the commonalities between all sorts of document contents.
  */
public class DocumentContentImpl implements DocumentContent
{
  /** Buffer size for reading
   *  16k is 4 times the block size on most filesystems
   *  so it should be efficient for most cases
   *  */
  private static final int INTERNAL_BUFFER_SIZE  = 16*1024;

  /** Default construction */
  public DocumentContentImpl() {
    content = "";
  } // default construction

  /** Contruction from URL and offsets. */
  public DocumentContentImpl(URL u, String encoding, Long start, Long end)
  throws IOException {

    int readLength = 0;
    char[] readBuffer = new char[INTERNAL_BUFFER_SIZE];

    BufferedReader uReader = null;
    InputStream uStream = null;
    StringBuffer buf = new StringBuffer();

    long s = 0, e = Long.MAX_VALUE;
    if(start != null && end != null) {
      s = start.longValue();
      e = end.longValue();
    }

    try {
      URLConnection conn = u.openConnection();
      uStream = conn.getInputStream();
            
      if ("gzip".equals(conn.getContentEncoding())) {
        uStream = new GZIPInputStream(uStream);
      }
      
      if(encoding != null && !encoding.equalsIgnoreCase("")) {
        uReader = new BomStrippingInputStreamReader(uStream, encoding, INTERNAL_BUFFER_SIZE);
      } else {
        uReader = new BomStrippingInputStreamReader(uStream, INTERNAL_BUFFER_SIZE);
      };
  
      // 1. skip S characters
      uReader.skip(s);
  
      // 2. how many character shall I read?
      long toRead = e - s;
  
      // 3. read gtom source into buffer
      while (
        toRead > 0 &&
        (readLength = uReader.read(readBuffer, 0, INTERNAL_BUFFER_SIZE)) != -1
      ) {
        if (toRead <  readLength) {
          //well, if toRead(long) is less than readLenght(int)
          //then there can be no overflow, so the cast is safe