取消特定的转义HTML标记

Question

背景故事

我有一种情况，我收到一个字节数组，可能是一种word文档（有三种可能性）：

1. Word 97二进制文件
2. HTML mascarading as Word with Custom MS Word Tags
3. Word 2013二进制文件

目前我正在处理1和2。 如果解析1）失败，我目前假设该文档是HTML。 遗憾的是，该文档包含某些转义的HTML标记（作为文本内容），我需要将其转移到处理其他人开发的代码（CMS导入等）。

当前实施

下面请参阅我使用Apache POI进行的当前实现1），并切换到字符串操作，以满足我对Poetr和Articl标签的需求。

如果使用正则表达式更有效地执行此操作，请提供建议。

 private void handleWordDocument(Byte[] bytes) {
    //Parse DOC file
    WordExtractor extractor = null;

    try
    {
      InputStream fis = new ByteArrayInputStream(bytes);
      HWPFDocument document = new HWPFDocument(fis);
      extractor = new WordExtractor(document);
      StringBuilder poetryIntroAndContent = new StringBuilder();

      String text = extractor.getText();  //Escaped HTML looks liks HTML as text
      attributions.setAttributionsFile(text.getBytes("UTF-8"));          
    }
    catch (Exception exep)
    {
      LOG.info("Parsing word file failed using Apache POI for word97:" + exep.toString()+". Will attempt to parse file with HTML reader after unescaping html tags.");
      try {
        String content = new String(attributions.getAttributionsFile(),"UTF-8");
        //content = content.replaceAll("&lt;[/]?Poetry(.*?)&gt;","<Poetry$0>").replaceAll("Poetry&lt;","").replaceAll("");
        String[] segments = content.split("&lt;");
        StringBuilder stringBuilder = new StringBuilder();
        String extraDebug = "";

        for(String segment: segments){
          int endTag = -1;

//          //TODO: MBAK: Remove this
//          if(System.currentTimeMillis()/1000 < 1481328000){ //December 10, 2016
//            extraDebug = "-Testing Translation" + System.currentTimeMillis();
//          }

          if(segment.length() < 5){ stringBuilder.append(segment+extraDebug); continue;}

          String segmentSubstring = segment.substring(0,6);

          if(!(
              segmentSubstring.toLowerCase().equals("poetry") ||
                  segmentSubstring.toLowerCase().equals("/poetr") ||
                  segmentSubstring.toLowerCase().equals("articl") ||
                  segmentSubstring.toLowerCase().equals("/artic")
              )
              ){
            stringBuilder.append(segment); continue;
          }

          String[] segments2 = segment.split("&gt;");

          if(segments2.length < 2){  stringBuilder.append(segment); continue; }

          segments2[0] = "<" + segments2[0].replaceAll("&quot;","\"") + ">";

          int counter=0;
          for(String subsegment: segments2){
            stringBuilder.append(subsegment);          }
          }

        attributions.setAttributionsFile(stringBuilder.toString().getBytes("UTF-8"));
      } catch (UnsupportedEncodingException e) {
        e.printStackTrace();
      }
    }
  }