学步园

全文检索技术在文档管理方面应用非常广泛，lucene是apache里面的一个子项目。

全文检索，主要几个过程，创建索引文件、检索、通过索引查找相关文件。

通过一个简单的入门实例来理解lucene的处理过程。

1、为文件创建索引

package cn.com.raul.lucene;<br /> import java.io.File;<br /> import java.io.FileReader;<br /> import java.io.IOException;<br /> import java.util.Date;<br /> import junit.framework.TestCase;<br /> import org.apache.lucene.analysis.standard.StandardAnalyzer;<br /> import org.apache.lucene.document.Document;<br /> import org.apache.lucene.document.Field;<br /> import org.apache.lucene.index.IndexWriter;<br /> import org.apache.lucene.store.FSDirectory;<br /> import org.apache.lucene.util.Version;<br /> /**<br /> * 创建索引<br /> *<br /> * @author Administrator<br /> *<br /> */<br /> public class IndexWork extends TestCase {<br /> private static String INDEX_DIR = "D://index";// 索引存放目录<br /> private static String DATA_DIR = "D://hibernate//";// 文件存放的目录<br /> public void testLucene() throws Exception {<br /> long start = new Date().getTime();<br /> int numIndexed = index(new File(INDEX_DIR), new File(DATA_DIR));//调用index方法<br /> long end = new Date().getTime();<br /> System.out.println("Indexing " + numIndexed + " files took " + (end - start) + " milliseconds");<br /> }<br /> /**<br /> *<br /> * @param indexFile<br /> * 索引文件<br /> * @param dataFile<br /> * 待索引文件<br /> * @return<br /> */<br /> public static int index(File indexFile, File dataDir) throws Exception {<br /> if (!indexFile.exists()) {<br /> indexFile.mkdir() ;<br /> }<br /> if (!dataDir.exists()) {<br /> throw new IOException("文件不存在");<br /> }<br /> IndexWriter writer = new IndexWriter(FSDirectory.open(indexFile),<br /> new StandardAnalyzer(Version.LUCENE_30), true,<br /> IndexWriter.MaxFieldLength.LIMITED);<br /> indexDirectory(writer, dataDir);<br /> int numIndexed = writer.numDocs();<br /> writer.optimize();<br /> writer.close();<br /> return numIndexed;<br /> }<br /> /**<br /> * 取出文件并循环索引<br /> *<br /> * @param writer<br /> * @param dataDir<br /> * @throws IOException<br /> */<br /> public static void indexDirectory(IndexWriter writer, File dataDir)<br /> throws IOException {<br /> File[] fileList = dataDir.listFiles();<br /> for (int i = 0; i < fileList.length; i++) {<br /> File f = fileList[i];<br /> if (f.isDirectory()) {<br /> indexDirectory(writer, f); // recurse<br /> } else if (f.getName().endsWith(".java")) {<br /> indexFile(writer, f);<br /> }<br /> }<br /> }<br /> /**<br /> * 对单个文件进行索引<br /> *<br /> * @param writer<br /> * @param f<br /> */<br /> public static void indexFile(IndexWriter writer, File f) throws IOException {<br /> if (f.isHidden() || !f.exists() || !f.canRead()) {<br /> return;<br /> }<br /> System.out.println("Indexing " + f.getCanonicalPath());<br /> Document doc = new Document();<br /> doc.add(new Field("contents", new FileReader(f)));// 有变化的地方<br /> doc.add(new Field("filename", f.getCanonicalPath(), Field.Store.YES,<br /> Field.Index.ANALYZED));// 有变化的地方<br /> writer.addDocument(doc);<br /> }<br /> }<br />

2、检索文档

package cn.com.raul.lucene;<br /> import java.io.File;<br /> import java.util.Date;<br /> import junit.framework.TestCase;<br /> import org.apache.lucene.analysis.standard.StandardAnalyzer;<br /> import org.apache.lucene.document.Document;<br /> import org.apache.lucene.queryParser.QueryParser;<br /> import org.apache.lucene.search.IndexSearcher;<br /> import org.apache.lucene.search.Query;<br /> import org.apache.lucene.search.ScoreDoc;<br /> import org.apache.lucene.search.TopScoreDocCollector;<br /> import org.apache.lucene.store.FSDirectory;<br /> import org.apache.lucene.util.Version;<br /> /**<br /> * 查询<br /> *<br /> * @author Administrator<br /> *<br /> */<br /> public class SearcheWork extends TestCase {<br /> private static String INDEX_DIR = "D://index//";// 索引所在的路径<br /> private static String KEYWORD = "insert";// 关键词<br /> private static int TOP_NUM = 10;// 显示前10条结果<br /> public void testSearch() throws Exception {<br /> search(new File(INDEX_DIR), KEYWORD);<br /> }<br /> public static void search(File indexDir, String q) throws Exception {<br /> IndexSearcher is = new IndexSearcher(FSDirectory.open(indexDir), true);<br /> String field = "contents";<br /> QueryParser parser = new QueryParser(Version.LUCENE_30, field,<br /> new StandardAnalyzer(Version.LUCENE_30));// 有变化的地方<br /> Query query = parser.parse(q);<br /> TopScoreDocCollector collector = TopScoreDocCollector.create(TOP_NUM,<br /> false);<br /> long start = new Date().getTime();// start time<br /> is.search(query, collector);<br /> ScoreDoc[] hits = collector.topDocs().scoreDocs;<br /> System.out.println(hits.length);<br /> for (int i = 0; i < hits.length; i++) {<br /> Document doc = is.doc(hits[i].doc);// new method is.doc()<br /> System.out.println(doc.getField("filename") + " "<br /> + hits[i].toString() + " ");<br /> }<br /> long end = new Date().getTime();// end time<br /> System.out.println("Found " + collector.getTotalHits()<br /> + " document(s) (in " + (end - start)<br /> + " milliseconds) that matched query '" + q + "':");<br /> }<br /> }<br />

Indexing D:/hibernate/CommonDao.java
Indexing D:/hibernate/HibernateDao.java
Indexing D:/hibernate/SimpleHibernateDao.java
Indexing 3 files took 467 milliseconds

1
stored,indexed,tokenized<filename:D:/hibernate/SimpleHibernateDao.java>   doc=2 score=0.046585027 
Found 1 document(s) (in 31 milliseconds) that matched query 'insert':