org.eclipse.epf.web.search/src/org/eclipse/epf/web/search/IndexSearch.java

package org.eclipse.epf.web.search;

import java.util.StringTokenizer;

import org.apache.lucene.document.Document;
import org.apache.lucene.queryParser.ParseException;
import org.apache.lucene.queryParser.QueryParser;
import org.apache.lucene.search.Hits;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Searcher;
import org.apache.lucene.search.Sort;
import org.apache.lucene.search.SortField;
import org.apache.lucene.store.RAMDirectory;
import org.eclipse.epf.web.search.analysis.ChineseAnalyzer;
import org.eclipse.epf.web.search.analysis.TextAnalyzer;

import com.ibm.icu.text.DecimalFormat;

public class IndexSearch {

	private static Searcher searcher;

	public static Hits search(String indexLocation, String inputQuery,
		String searchField) {
		
		Hits hits = null;
		Query q = null;
		String queryString = null;
		DecimalFormat scoreFormatter = new DecimalFormat("0.######");

		if (indexLocation == null || inputQuery == null || searchField == null
			|| searchField.length() == 0 || inputQuery.length() == 0)
			return null;

		try {
			queryString = detectHyphenated(inputQuery);

			if (queryString == null || queryString.length() == 0)
				return null;

			if (searcher == null) {
				searcher = new IndexSearcher(new RAMDirectory(indexLocation));
//				searcher = new IndexSearcher(FSDirectory.getDirectory(indexLocation,
//					false));
			}		
						
			if ( searcher == null ) {
				return hits;
			}
			
//			Sort sort = new Sort(new SortField[]{					
//					new SortField("title"),
//					new SortField("summary")
//					});
			
			// per user's request, the sorting should be based on the hit rate, 
			// so don't pass in any sorting criteria
			Sort sort = null;
			
			try {
				
				q = QueryParser.parse(queryString, searchField,
						new TextAnalyzer());
				if ( q != null ) {
					hits = searcher.search(q, (Sort)sort);
				}
			} catch (Throwable th ) {
				
			}
			
			// note: icu4j failed with jre 1.5 in server mode.
			// in such a case, try the old one
			q = QueryParser.parse(queryString, searchField,
					new ChineseAnalyzer());
			if ( q != null ) {
				hits = searcher.search(q, (Sort)sort);
			}
			
			return hits;
			

		} catch (ParseException pe) {
			pe.printStackTrace();
		} catch (java.io.IOException io) {
			io.printStackTrace();
		} catch (Throwable p1) {
			p1.printStackTrace();
		}

		return hits;
	}

	/*
	 * detect hyphenated words and if hyphenate, put quotes around so that they are
	 * considered one word for example,"use-case" as opposed to use-case, which Lucene
	 * interprets as "use" NOT "case" which is not what we want. algorithm: first tokenize
	 * by space to isolate words next, if words have hyphens, put quotes at the beginning
	 * and end append to buffer
	 */
	public static String detectHyphenated(String queryString) {		
		StringTokenizer spaceTokenizer = new StringTokenizer(queryString);
		StringBuffer wordBuffer = new StringBuffer();
				
		while (spaceTokenizer.hasMoreTokens()) {
			// isolate into words
			String word = spaceTokenizer.nextToken();
			if ((word.indexOf('-') != -1 ) && !word.startsWith("\"")) //$NON-NLS-1$
			{
				// hyphen found, and word is not already an exact phrase,
				// so add quotes to beginning and end of word
				wordBuffer.append("\""); //$NON-NLS-1$
				wordBuffer.append(word);
				wordBuffer.append("\""); //$NON-NLS-1$
				//System.out.println("if: " + wordBuffer.toString());
			} else {
				//System.out.println("else: " + queryString);
				return queryString;  //RATLC00251031
				//wordBuffer.append(word);				
			}
			if (spaceTokenizer.hasMoreTokens()) {
				wordBuffer.append(" "); //$NON-NLS-1$
			}
		}
		
		
		return wordBuffer.toString();
	}

	/**
	 * @param args
	 */
//	public static void main(String[] args) {
//		if (args.length > 0) {
//			// String rupPath = System.getProperty( "user.home" ) + java.io.File.separator
//			// + "rup";
//			String rupPath = args[0];
//			SortedHits sortedHits = IndexSearch.search(rupPath + "\\index", "Analyst",
//				"contents");
//			System.out
//				.println("There are " + sortedHits.length() + " hits for 'analyst'");
//		}
//
//	}
}