/**
 * 
 */
package edu.odu.cs.extract.control;

import org.jdom.Document;

import edu.odu.cs.extract.dataflow.Dataflow;
import edu.odu.cs.extract.dataflow.QuickTransformer;
import edu.odu.cs.extract.dataflow.TransformationResult;
import edu.odu.cs.extract.inputprocessing.segmentation.Segmentation;
import edu.odu.cs.extract.utils.Properties;

/**
 * Transforms a PDF file dataflow into Raw IDM by attempting a direct translation
 * of text PDF, but passing pages thought to be scanned on for OCR and thenby trimming to a selected number of pages
 * OCR-to-rawIDM conversion.
 * 
 * @author zeil
 *
 */
public class SegmentationTransformer extends QuickTransformer {


    /**
     * 
     */
    public SegmentationTransformer() {
       super();
    }

    /* (non-Javadoc)
     * @see edu.odu.cs.extract.dataflow.ThreadedTransformer#doTransform(edu.odu.cs.extract.dataflow.Dataflow[])
     */
    @Override
    public TransformationResult doTransform(Dataflow[] in) throws Exception {
       String status = "success";
       String message = "OK";

       IDMDataflow inputDF = (IDMDataflow) in[0];
       Document unsegmentedIDM = inputDF.getDocument();
       String mergeFailed = unsegmentedIDM.getRootElement().getAttributeValue("OCRmerge");

       if (mergeFailed != null && "failed".equals(mergeFailed)) {
         status = "warning";
         message = "unable to merge pages from OCR";
       }

       // Segment document
       Document segmentedIDM = new Segmentation(unsegmentedIDM).reSegment();

       IDMDataflow outputDF = new IDMDataflow (in[0].getTrace(), segmentedIDM);

       return new TransformationResult(outputDF,status, message, null);
    }

    @Override
    public String getOutputExtension() {
       Properties p = Properties.getProperties();
       return p.getProperty(Properties.Names.SEGMENTATION_OUT_EXT);
    }



}

unformatted source