package com.lss.common.doc;
import java.io.File;
import java.io.FileInputStream;
import org.apache.log4j.Logger;
import org.textmining.text.extraction.WordExtractor;
/**
* Word文件工具 *.doc
*
* @Description
* @author SHOUSHEN LUAN
* @Date 2011-12-10
*/
public class WordUtils {
private final static Logger logger = Logger.getLogger(WordUtils.class);
/**
* 读取Word文件
*
* @param filePath
* @throws Exception
* @throws Exception
*/
public String readWordFile(String filePath) throws Exception {
return readWordFile(new File(filePath));
}
public String readWordFile(File file) throws Exception {
FileInputStream in = null;
String docContent = null;
try {
in = new FileInputStream(file);
WordExtractor extractor = new WordExtractor();
docContent = extractor.extractText(in);
} catch (Exception e) {
logger.error(e);
} finally {
if (in != null) {
in.close();
}
}
return docContent;
}
}