package iitb.MaxentClassifier;import iitb.CRF.*;/** * * @author Sunita Sarawagi * */ public class FeatureGenRecord implements FeatureGenerator {    int numCols;    int numLabels;    DataRecord dataRecord;    class FeatureColumn implements Feature {	int colId;	int _y;	float val;	void copy(FeatureColumn f) {	    colId = f.colId;	    _y = f._y;	    val = f.val;	}	public int index() {return colId + _y*numCols;}	public int y() {return _y;}	public int yprev() {return -1;}	public float value() {return val;}	public int[] yprevArray() {return null;}    };    FeatureColumn feature, featureToReturn;    FeatureGenRecord(int numColumns, int numYs) {	numCols =  numColumns;	numLabels = numYs;	feature = new FeatureColumn();	featureToReturn = new FeatureColumn();     }    public int numFeatures() {return numCols*numLabels;}    public void startScanFeaturesAt(DataSequence data, int pos) {	dataRecord = (DataRecord)data;	assert (pos == 0);	feature.colId = 0;	feature._y = 0;    }    public boolean hasNext() {	return (feature.y() < numLabels);    }    public Feature next() {	featureToReturn.copy(feature);	feature.colId++;	if (feature.colId >= numCols) {	    feature.colId = 0;	    feature._y++;	}	featureToReturn.val = dataRecord.getColumn(featureToReturn.colId);	return featureToReturn;    }    /* (non-Javadoc)     * @see iitb.CRF.FeatureGenerator#featureName(int)     */    public String featureName(int featureIndex) {        return "ColumnId=" + (featureIndex % numCols) + " label="+(featureIndex/numCols);    }};