source: trunk/greenstone3-extensions/vishnu/src/vishnu/builder/LuceneWrapper.java@ 8290

package vishnu.builder;

import java.io.*;
import java.util.*;
import org.apache.lucene.analysis.standard.StandardAnalyzer;
import org.apache.lucene.index.IndexWriter;


public class LuceneWrapper extends Indexer
{
    
    private static void usage()
    {
    System.out.println("java LucenceWrapper -i InputDirectory \n -o OutputDirectory \n -c CollectionName");
    }
    
    
    public static void main(String[] args)
    {
    String in = null;
    String out = null;
    String coll = null;

    if( args.length == 0 ){
        usage();
        System.exit(0);
    }
    
    int a = 0;
    while( a < args.length ){
        if( args[a].equals("-i") )
        in = args[++a];
        if( args[a].equals("-o") )
        out = args[++a];
        if( args[a].equals("-c") )
        coll = args[++a];
        a++;
    }
    if (out==null || in==null|| coll==null) {
        usage();
        System.exit(0);
    }
    if (!out.endsWith(File.separator)) {
        out += File.separator;
    }
    if (!in.endsWith(File.separator)) {
        in += File.separator;
    }

    LuceneWrapper luc = new LuceneWrapper();
    
    out = out + "luc_index";
    luc.setCollectionName(coll);
        luc.setOutputDirectory(out);
    luc.setInputDirectory(in);
    luc.startIndexing();
    }
    
    
    public LuceneWrapper()
    {
    super("Lucene Indexer");
    }
    
    public void startIndexing() {
    initialize();
    try {
        IndexWriter writer = new IndexWriter(outputDir, new StandardAnalyzer(), true);
        // this is a bit of a hack putting import here
        indexFiles(writer, new File(inputDir), "import/");
        writer.optimize();
        writer.close();
    } catch (Exception e) {
        System.err.println("Error creating the Lucene index: "+e);
        System.exit(0);
    }

    }

    private void indexFiles(IndexWriter writer, File input_dir, String current_path) {
    File [] doc_list = input_dir.listFiles();
    for (int i=0; i<doc_list.length; i++) {
        File f = doc_list[i];
        if (f.isDirectory()) {
        indexFiles(writer, f, current_path+f.getName()+File.separator);
        } else {
        try {
            writer.addDocument(FileDocument.Document(f, current_path+f.getName())); 
        } catch (Exception e) {
            System.err.println("Error indexing document "+current_path+f.getName()+": "+e);
        }
        }
    }

    }

    
    public void startIndexingOld()
    {
    
    if( inputDir == null || outputDir == null ){
        usage();
        System.exit(0);
    }
    
    
    /**** check if output directory exists, if not create it ****/
    
    
        File f = new File(outputDir);
        if( !f.exists() ){
            System.out.println("Creating output directory..");
        
            StringTokenizer strt = new StringTokenizer(outputDir,"/");
            String d = "";
            while(strt.hasMoreTokens()){
                String str = strt.nextToken();
                if( str.indexOf(".") == -1 ){
            d += "/" + str;
            if( !(new File(d)).exists() ){
            try{
                Runtime rt = Runtime.getRuntime();
                Process proc = rt.exec("mkdir " + d);
                int exitVal = proc.waitFor();
            }catch(Exception ex){ex.printStackTrace();}
            }
        }
        }
        }   
    
    File docList = new File(inputDir + "/docs.lst");
    
    if (!docList.exists()){
        System.out.println("No documents found");
        System.exit(0);
    }
    
    Vector docs = new Vector();
    
    try{
            BufferedReader dis = new BufferedReader(new InputStreamReader(new FileInputStream(docList)));
        String line = "";
            while( (line=dis.readLine())!=null ){
        docs.addElement(new String(inputDir+"/" + line));
        }
            dis.close();
    } catch(IOException ex){ex.printStackTrace();}
    
    try{
        
        IndexWriter writer = new IndexWriter(outputDir, new StandardAnalyzer(), true);
        
        for( int i = 0; i < docs.size(); i++ ){
        File file = new File((String)docs.elementAt(i));        
        //writer.addDocument(FileDocument.Document(file));
        }
        writer.optimize();
        writer.close();
        }catch(Exception ex){ex.printStackTrace();}
    }
}