Master

src/refactor/base/AlphaBeta.java

@@ -0,0 +1,34 @@
+/**
+ * Created锛�May 10, 2013 2:18:41 PM  
+ * Project锛�ThulacJava  
+ * @author cxx
+ * @since JDK 1.6.0_13  
+ * filename锛�AlphaBeta.java  
+ * description锛�  
+ */
+package base;
+
+
+//a structure for alphas and betas
+public class AlphaBeta {
+	public int value;
+	public int nodeId;
+	public int labelId;
+
+	public AlphaBeta() {
+		super();
+		value = 0;
+		nodeId = -2;
+		labelId = 0;
+	}
+
+	public AlphaBeta(int value, int nodeId, int labelId) {
+		super();
+		this.value = value;
+		this.nodeId = nodeId;
+		this.labelId = labelId;
+	}
+
+
+
+}

src/refactor/base/Counter.java

@@ -0,0 +1,26 @@
+/**
+ * Created锛�May 9, 2013 2:32:10 PM  
+ * Project锛�ThulacJava  
+ * @author cxx
+ * @since JDK 1.6.0_13  
+ * filename锛�Counter.java  
+ * description锛�  
+ */
+package base;
+
+import java.util.HashMap;
+
+public class Counter<KeyType> extends HashMap<KeyType, Integer> {
+	/**
+	 * 
+	 */
+	private static final long serialVersionUID = 1L;
+
+	public void update(KeyType key){
+		Integer value = get(key);
+		if(value == null){
+			value = 0;
+		}
+		put(key, value + 1);
+	}
+}

src/refactor/base/Dat.java

@@ -0,0 +1,140 @@
+package base;
+
+
+import java.io.*;
+import java.util.Vector;
+
+public class Dat {
+
+	public Vector<Entry> dat;
+	public int datSize;
+
+	public Dat(){
+		dat = new Vector<Entry>();
+		datSize = 0;
+	}
+	public Dat(int datSize, Vector<Entry> olddat){
+		this.datSize = datSize;
+        dat= new Vector<Entry>();
+		for(int i = 0; i < datSize; i ++){
+			dat.add(new Entry());
+			dat.get(i).base = olddat.get(i).base;
+			dat.get(i).check = olddat.get(i).check;
+		}
+	}
+	public Dat(String filename) throws IOException {
+		filename = Dat.class.getClassLoader().getResource(filename).getFile();
+		File file = new File(filename);
+		datSize = (int)(file.length() / 8);
+		//System.out.println(datSize);
+
+		FileInputStream in = new FileInputStream(file);
+
+		byte[] tempbytes = new byte[8 * datSize];
+		dat = new Vector<Entry>();
+		in.read(tempbytes);
+		for(int i = 0; i < datSize; i ++){
+			Entry entry = new Entry();
+			entry.base = bytesToInt(tempbytes, 8 * i);
+
+			dat.add(entry);
+			dat.get(i).check = bytesToInt(tempbytes, 8 * i + 4);
+		}
+
+		in.close();
+	}
+
+	public static int bytesToInt(byte[] bb, int index) {
+		return (int) (((((int)bb[index + 3] & 0xff) << 24)
+				| (((int)bb[index + 2] & 0xff) << 16)
+				| (((int)bb[index + 1] & 0xff) << 8) | (((int)bb[index + 0] & 0xff) << 0)));
+	}
+
+	public static byte[] intToBytes(int n){
+	    byte[] b = new byte[4];
+	    for(int i = 0;i < 4;i++){
+	        b[i] = (byte)(n >> (8 * i));
+	    }
+	    return b;
+	}
+
+	public void save(String filename) throws IOException{
+		FileOutputStream out = new FileOutputStream(filename);
+		for(Entry e : dat){
+			out.write(intToBytes(e.base));
+		}
+		out.flush();
+		for(Entry e : dat){
+			out.write(intToBytes(e.check));
+		}
+		out.flush();
+		out.close();
+	}
+
+	public boolean search(String sentence, Vector<Integer> bs, Vector<Integer> es){
+		bs.clear();
+		es.clear();
+		boolean empty = true;
+		for(int offset = 0; offset < sentence.length(); offset ++){
+			int preBase = 0;
+			int preInd = 0;
+			int ind = 0;
+			for(int i = offset; i < sentence.length(); i ++){
+				ind = preBase + sentence.charAt(i);
+				if(ind < 0 || ind >= datSize || dat.get(ind).check != preInd)break;
+				preInd = ind;
+				preBase = dat.get(ind).base;
+				ind = preBase;
+				if(!(ind < 0 || ind >= datSize || dat.get(ind).check != preInd)){
+					bs.add(offset);
+					es.add(i + 1);
+					if(empty){
+						empty = false;
+					}
+				}
+			}
+		}
+		return !empty;
+	}
+
+	public int match(String word){
+		int ind = 0;
+		int base = 0;
+		for(int i = 0; i < word.length(); i ++){
+			ind = dat.get(ind).base + word.charAt(i);
+			if((ind >= datSize) || (dat.get(ind).check != base)) return -1;
+			base = ind;
+		}
+		ind = dat.get(base).base;
+		if((ind < datSize) && (dat.get(ind).check == base)){
+			return ind;
+		}
+		return -1;
+	}
+
+	public void update(String word, int value){
+		int base = match(word);
+		if(base >= 0){
+			dat.get(base).base = value;
+		}
+	}
+
+	public int getInfo(String prefix){
+		int ind = 0;
+		int base = 0;
+		for(int i = 0; i < prefix.length(); i ++){
+			ind = dat.get(ind).base + prefix.charAt(i);
+			if((ind >= datSize) || dat.get(ind).check != base) return i;
+			base = ind;
+		}
+		return -base;
+	}
+
+	public int getDatSize(){
+		return datSize;
+	}
+
+	public Vector<Entry> getDat(){
+		return dat;
+	}
+}