search.java

lucene 是java 的版的搜索引擎公共模块

package com.bitmechanic.spindle;

import org.apache.lucene.analysis.standard.StandardAnalyzer;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.Hits;
import org.apache.lucene.search.Searcher;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.queryParser.QueryParser;
import java.util.ArrayList;
import java.util.HashMap;
import javax.servlet.jsp.PageContext;
import com.bitmechanic.listlib.*;

/**
 * Use this class to search indexes created by the Spider class.
 *
 * @author James Cooper
 * @version $Id: Search.java,v 1.5 2002/03/30 20:25:46 pixel Exp $
 */
public class Search implements ListCreator { 

    /**
     * Command line usage:
     * 
     * <code>java com.bitmechanic.spindle.Search [index dir] [keyword]</code>
     *
     */
    public static void main(String[] args) throws Exception {
        String indexPath = args[0];
        String queryString = args[1];

        Searcher searcher = new IndexSearcher(indexPath);
        //Query q = searcher.
        QueryParser parser=new QueryParser("body",new StandardAnalyzer());
     	Query q=parser.parse(queryString);
        //parse(queryString, "body", new StandardAnalyzer());
        Hits hits = searcher.search(q);

        for (int i = 0; i < hits.length(); i++) {
            System.out.println(hits.doc(i).get("url") + "; Title: " +
                               hits.doc(i).get("title") + "; Score: " + 
                               hits.score(i));
            System.out.println("Description: " + hits.doc(i).get("desc"));
        };
        searcher.close();
    }

    private static final String attrib[] = { "url", "title", "desc" };

    /**
     * Runs a query against the index in the directory specified.  Returns
     * a list of HashMap objects.  The objects have the following keys:
     * 'url', 'title', 'desc', 'score'. 
     *
     * @param dir Directory containing the index to search
     * @param query Query to run against the index 
     */
    public static ArrayList search(String dir, String query) throws Exception {
        ListDesc desc = search(dir, query, 0, Integer.MAX_VALUE);
        return desc.list;
    }

    /**
     * Runs a query against the index in the directory specified.  Returns
     * a list of HashMap objects.  The objects have the following keys:
     * 'url', 'title', 'desc', 'score'. 
     *
     * @param dir Directory containing the index to search
     * @param query Query to run against the index 
     * @param offset Position in the result set to start returning results
     * @param max Max number of results to return
     */
    public static ListDesc search(String dir, String query, int offset, 
                                  int max) throws Exception {

        Searcher searcher = new IndexSearcher(dir);
       // Query q = QueryParser.parse(query, "body", new StandardAnalyzer());
                                  
        QueryParser parser=new QueryParser("body",new StandardAnalyzer());
     	Query q=parser.parse(query);
     	
        Hits h = searcher.search(q);
        ArrayList list = new ArrayList(h.length());
        for (int i = offset; list.size() < max && i < h.length(); i++) {
            HashMap map = new HashMap();
            for (int x = 0; x < attrib.length; x++) {
                map.put(attrib[x], h.doc(i).get(attrib[x]));
            }
            map.put("score", new Float(h.score(i)));
            list.add(map);
        }
        searcher.close();

        ListDesc desc = new ListDesc();
        desc.list = list;
        desc.count = h.length();
        return desc;
    }

    ////////////////////////////////////////////////////////////////////////
    //
    // ListCreator implementation
    //
    ////////////////////////////////////////////////////////////////////////

    private String _dir, _query;

    public void setDir(String dir) {
        _dir = dir;
    }

    public void setQuery(String query) {
        _query = query;
    }

    /**
     * Runs a query against the index in the directory specified.  Returns
     * a list of HashMap objects.  The objects have the following keys:
     * 'url', 'title', 'desc', 'score'. 
     *
     * @param dir Directory containing the index to search
     * @param query Query to run against the index 
     * @param offset Position in the result set to start returning results
     * @param max Max number of results to return
     */
    public ListContainer execute(PageContext context, int offset, int max) 
        throws Exception {

        ListDesc desc = search(_dir, _query, offset, max);
        return new ListContainer(desc.list.iterator(), desc.count);
    }
}