RTF转HTML,HTML转TXT(Java版)之威力加强版

增强的功能有：

自由指定要转换的RTF文件和输出文件，输出文件可以不指定

支持HTML与TXT两种格式，默认为TXT

可以指定转换编码，默认为UTF-8

提高HTML字符反转义效率（循环代替递归）

显示执行时间，5M的RTF转换在4s内完成

代码如下，展开查看！

import java.io.File;
import java.io.FileWriter;
import java.util.Date;
import java.util.HashMap;
import pt.tumba.parser.rtf.RTF2HTML;
/**
 * Convert RTF to HTML, RTF to TXT
 * 
 * @author KNIGHTRCOM([e-mail])
 * {@link http://blog.csdn.net/rcom10002}
 */
public class Main {
	private static String sourceFilename;
	private static String outputFilename;
	private static boolean isForced = false; // indicate to overwrite the existing file
	private static boolean isSystemListed = false;
	private static String type = "txt";
	private static String encoding = "UTF-8";
	/**
	 * -i System information (optional)
	 * -s Source RTF file    (mandatory)
	 * -o Output file name   (optional)
	 * --force Overwrite output file if it exists
	 * 
	 * @param args
	 * @throws Exception
	 */
	public static void main(String[] args) {
		try {
			if (args != null && args.length > 0) {
				for (String arg : args) {
					if (arg.equals("-i")) {
						isSystemListed = true;
					} else if (arg.startsWith("-s")) {
						sourceFilename = arg.substring(2);
					} else if (arg.startsWith("-o")) {
						outputFilename = arg.substring(2);
					} else if (arg.equals("--force")) {
						isForced = true;
					} else if (arg.startsWith("-t")) {
						type = arg.substring(2);
						if (!"txt".equals(type) && !"html".equals(type)) {
							isSystemListed = false;
							sourceFilename = null;
							break;
						}
					} else if (arg.startsWith("-e")) {
						encoding = arg.substring(2);
					} else {
						isSystemListed = false;
						sourceFilename = null;
						break;
					}
				}
			}
			if (((sourceFilename == null || !new File(sourceFilename).exists()) && !isSystemListed) ||
					args != null && args.length == 1 && args[0].equals("--help")){
				System.out.println("usage: java Main [-t<html|txt>] [-eEncodingName] [--force] -sSourceFileName [-oOutputFileName]");
				System.out.println("       java Main [-i,--help]");
				return;
			}
			if (isSystemListed) {
				listSystemInfo();
			}
			executeConvertion();
		} catch (Exception e) {
			System.err.print("Errmsg: " + e.getMessage());
			System.exit(1);
		}
	}
	/**
	 * List the system info you may concern
	 */
	public static void listSystemInfo() {
		System.getProperties().list(System.out);
	}
	/**
	 * Convert RTF to required format
	 * 
	 * @throws Exception
	 */
	private static void executeConvertion() throws Exception {
		long duration = new Date().getTime();
		// Convert rtf to HTML
		String result = new RTF2HTML().convertRTFToHTML(new File(sourceFilename));
		// This step is important for rendering the text with proper encoding
		result = new String(result.getBytes(System.getProperty("sun.jnu.encoding")), encoding);
		if ("txt".equals(type)) {
			// Extract plain text from HTML
			result = result.replaceAll("(?i)<br.*?/?>", "<br />/n").replaceAll("<.+?>", "");
			result = StringUtils.unescapeHTML(result, 0);
		}
		// Write the result to the file
		if (outputFilename == null) {
			outputFilename = sourceFilename.concat(".txt");
		}
		File outputFile = new File(outputFilename);
		if (outputFile.exists() && !isForced) {
			System.out.print("Warning: Output file already exists! Try execute with --force.");
			System.exit(-1);
		}
		FileWriter w = new FileWriter(outputFile);
		w.write(StringUtils.trimThroughLines(result));
		w.close();
		duration -= new Date().getTime();
		System.out.print("Complete!(" + (duration / 1000 * -1) + "s)");
	}
}
/**
 * http://www.rgagnon.com/javadetails/java-0307.html
 * 
 */
class StringUtils {
	private StringUtils() {
	}
	private static HashMap<String, String> htmlEntities;
	static {
		htmlEntities = new HashMap<String, String>();
		htmlEntities.put("<", "<");
		htmlEntities.put(">", ">");
		htmlEntities.put("&", "&");
		htmlEntities.put(""", "/"");
		htmlEntities.put("à", "à");
		htmlEntities.put("À", "À");
		htmlEntities.put("â", "â");
		htmlEntities.put("ä", "ä");
		htmlEntities.put("Ä", "Ä");
		htmlEntities.put("Â", "Â");
		htmlEntities.put("å", "å");
		htmlEntities.put("Å", "Å");
		htmlEntities.put("æ", "æ");
		htmlEntities.put("Æ", "Æ");
		htmlEntities.put("ç", "ç");
		htmlEntities.put("Ç", "Ç");
		htmlEntities.put("é", "é");
		htmlEntities.put("É", "É");
		htmlEntities.put("è", "è");
		htmlEntities.put("È", "È");
		htmlEntities.put("ê", "ê");
		htmlEntities.put("Ê", "Ê");
		htmlEntities.put("ë", "ë");
		htmlEntities.put("Ë", "Ë");
		htmlEntities.put("ï", "ï");
		htmlEntities.put("Ï", "Ï");
		htmlEntities.put("ô", "ô");
		htmlEntities.put("Ô", "Ô");
		htmlEntities.put("ö", "ö");
		htmlEntities.put("Ö", "Ö");
		htmlEntities.put("ø", "ø");
		htmlEntities.put("Ø", "Ø");
		htmlEntities.put("ß", "ß");
		htmlEntities.put("ù", "ù");
		htmlEntities.put("Ù", "Ù");
		htmlEntities.put("û", "û");
		htmlEntities.put("Û", "Û");
		htmlEntities.put("ü", "ü");
		htmlEntities.put("Ü", "Ü");
		htmlEntities.put("", " ");
		htmlEntities.put("©", "/u00a9");
		htmlEntities.put("®", "/u00ae");
		htmlEntities.put("€", "/u20a0");
	}
	public static final String unescapeHTML(String source, int start) {
		int i, j;
		// 将递归算法转换成循环可以防止处理大数据时内存不足的缺陷
//		i = source.indexOf("&", start);
//		if (i > -1) {
//			j = source.indexOf(";", i);
//			if (j > i) {
//				String entityToLookFor = source.substring(i, j + 1);
//				String value = (String) htmlEntities.get(entityToLookFor);
//				if (value != null) {
//					source = new StringBuilder().append(
//							source.substring(0, i)).append(value).append(
//							source.substring(j + 1)).toString();
//					return unescapeHTML(source, i + 1); // recursive call
//				}
//			}
//		}
		while (true) {
			i = source.indexOf("&", start);
			if (i > -1) {
				j = source.indexOf(";", i);
				if (j > i) {
					String entityToLookFor = source.substring(i, j + 1);
					String value = (String) htmlEntities.get(entityToLookFor);
					if (value != null) {
						source = new StringBuilder().append(
								source.substring(0, i)).append(value).append(
								source.substring(j + 1)).toString();
						start = i + 1;
					}
				} else {
					break;
				}
			} else {
				break;
			}
		}
		return source;
	}
	/**
	 * Remove heading and tailing space or tab of all lines
	 * 
	 * @param source
	 * @return
	 */
	public static final String trimThroughLines(String source) {
		return source.replaceAll("(?ms)^//s+|//s+$", "");
	}
	public static void main(String args[]) throws Exception {
		// to see accented character to the console
		java.io.PrintStream ps = new java.io.PrintStream(System.out, true,
				"Cp850");
		String test = "© 2007  Réal Gagnon <www.rgagnon.com>";
		ps.println(test + "/n-->/n" + unescapeHTML(test, 0));
		/*
		 * output : © 2007 Réal Gagnon <www.rgagnon.com>
		 * --> © 2007 Réal Gagnon <www.rgagnon.com>
		 */
	}
}

java -Dfile.encoding=UTF-8 -cp "C:/Documents and Settings/Administrator/My Documents/Workspace/eclipse/RTF/Document Parser;" Main -s"C:/Documents and Settings/Administrator/My Documents/Workspace/php eclipse/QAR Tool/questions/sample.rtf" -o"C:/my.txt" --force -eGB2312 -thtml