package oboannotator.annotator;

import java.util.regex.Matcher;
import java.util.regex.Pattern;

import oboannotator.typesys.MedlineDocument;

import org.apache.uima.analysis_component.JCasAnnotator_ImplBase;
import org.apache.uima.jcas.JCas;


/**
 * Example annotator that detects room numbers using Java 1.4 regular expressions.
 */
public class MedlineDocumentAnnotator extends JCasAnnotator_ImplBase {
  //(^UI|PMID)\\s*\\-(.*?)^([A-Z][A-Z]+\\s*\\-)	
  //private Pattern docPattern = Pattern.compile("(^UI|PMID)[\\s]*\\-[\\s]+(.*?)([\n]{2})", Pattern.DOTALL|Pattern.MULTILINE);
  private Pattern docPattern = Pattern.compile("(^UI|PMID)[\\s]*\\-[\\s]+(.*?)([\n]{2})", Pattern.DOTALL|Pattern.MULTILINE);
  
  /**
   * @see JCasAnnotator_ImplBase#process(JCas)
   */
  public void process(JCas aJCas) {
    // get document text
    String docText = aJCas.getDocumentText();
    // search for Yorktown room numbers
    Matcher matcher = docPattern.matcher(docText);
    while (matcher.find()) {
      // found one - create annotation
      MedlineDocument annotation = new MedlineDocument(aJCas);
      annotation.setBegin(matcher.start());
      annotation.setEnd(matcher.end());
      annotation.setDocID(matcher.group(2));
      annotation.setText(matcher.group(1) + "  -  " +  matcher.group(2));
      annotation.addToIndexes();
    }        
  }
}