import org.apache.poi.hwpf.converter.WordToHtmlConverter; //导入依赖的package包/类
public static String parseDoc2Html(InputStream input, String charset) throws Exception {
  HWPFDocument wordDocument = new HWPFDocument(input);
  Document doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument();
  WordToHtmlConverter converter = new WordToHtmlConverter(doc);
  converter.processDocument(wordDocument);

  ByteArrayOutputStream output = new ByteArrayOutputStream();
  try {
    DOMSource domSource = new DOMSource(converter.getDocument());
    StreamResult streamResult = new StreamResult(output);
    Transformer serializer = TransformerFactory.newInstance().newTransformer();
    // TODO 有乱码
    serializer.setOutputProperty(OutputKeys.ENCODING, charset);
    serializer.setOutputProperty(OutputKeys.INDENT, "yes");
    serializer.setOutputProperty(OutputKeys.METHOD, "html");
    serializer.transform(domSource, streamResult);
  } finally {
    input.close();
    output.close();
  }

  return new String(output.toByteArray());
}
 

import org.apache.poi.hwpf.converter.WordToHtmlConverter; //导入依赖的package包/类
public static void main(String[] args)throws Exception
{
	HWPFDocumentCore wordDocument = WordToHtmlUtils.loadDoc(new FileInputStream("data/document.doc"));

    WordToHtmlConverter wordToHtmlConverter = new WordToHtmlConverter(
            DocumentBuilderFactory.newInstance().newDocumentBuilder()
                    .newDocument());
    wordToHtmlConverter.processDocument(wordDocument);
    Document htmlDocument = wordToHtmlConverter.getDocument();
    ByteArrayOutputStream out = new ByteArrayOutputStream();
    DOMSource domSource = new DOMSource(htmlDocument);
    StreamResult streamResult = new StreamResult(out);

    TransformerFactory tf = TransformerFactory.newInstance();
    Transformer serializer = tf.newTransformer();
    serializer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
    serializer.setOutputProperty(OutputKeys.INDENT, "yes");
    serializer.setOutputProperty(OutputKeys.METHOD, "html");
    serializer.transform(domSource, streamResult);
    out.close();
    
	FileOutputStream outputStream = new FileOutputStream("data/Apache_DocToHTML.html");
	outputStream.write(out.toByteArray());
	outputStream.close();
	
	System.out.println("Apache - Doc file converted in specified formats");
}
 

import org.apache.poi.hwpf.converter.WordToHtmlConverter; //导入依赖的package包/类
public static void main(String[] args)throws Exception
{
	String dataPath = "src/featurescomparison/workingwithdocuments/converttoformats/data/";
	
	HWPFDocumentCore wordDocument = WordToHtmlUtils.loadDoc(new FileInputStream(dataPath + "document.doc"));

    WordToHtmlConverter wordToHtmlConverter = new WordToHtmlConverter(
            DocumentBuilderFactory.newInstance().newDocumentBuilder()
                    .newDocument());
    wordToHtmlConverter.processDocument(wordDocument);
    Document htmlDocument = wordToHtmlConverter.getDocument();
    ByteArrayOutputStream out = new ByteArrayOutputStream();
    DOMSource domSource = new DOMSource(htmlDocument);
    StreamResult streamResult = new StreamResult(out);

    TransformerFactory tf = TransformerFactory.newInstance();
    Transformer serializer = tf.newTransformer();
    serializer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
    serializer.setOutputProperty(OutputKeys.INDENT, "yes");
    serializer.setOutputProperty(OutputKeys.METHOD, "html");
    serializer.transform(domSource, streamResult);
    out.close();
    
	FileOutputStream outputStream = new FileOutputStream(dataPath + "Apache_DocToHTML_Out.html");
	outputStream.write(out.toByteArray());
	outputStream.close();
	
	System.out.println("Apache - Doc file converted in specified formats");
}