java词频统计——web版支持

需求概要：

1.把程序迁移到web平台，通过用户上传TXT的方式接收文件。

2.用户直接输入要统计的文本，服务器返回结果

3.在页面上给出链接 (如果有封皮、作者、字数、页数等信息更佳)或表格，展示经典英文小说词频统计结果；

4.支持用户自定义单词分隔符；

5.词汇范围对比（额外项）。

分析和设计：

1.创建web工程，利用servlet上传文件的技术实现用户向服务器上传文件。页面设置表单类型为enctype="multipart/form-data"，创建文件上传文本框<input type="file" id="upfilename" name="upfilename" value="" />，服务器端使用Part p = request.getPart("upfilename");获取上传的文件，然后写入到指定地址即可。

2.直接分析用户post到服务器的内容，为了使用原有的api，可以将输入内容写到文件中，再进行分析。

3.页面展示统计结果

4.用户可以输入自定义的分隔符和设置显示统计结果前10行（可修改）。需要修改原词频统计的有效字符函数。

5.暂时不考虑

部分代码实现：

表单实现

     <div align="center" id="txtform">
         <form action="upload" method="post" enctype="multipart/form-data">
             <input type="file" id="upfilename" name="upfilename" value="" /> 自定义分隔符<input
                 type="text" id="splitter" name="splitter"> <input
                 type="submit" id="submit" value="上传" />
         </form>
     </div>
     <div align="center" id="txtform">
         <form action="wordcount" method="post">
             <div align="center">待统计内容</div>
             <textarea name="content" id="content"
                 style="width: 700px; height: 200px;"></textarea>
             <br> 统计前<select id="num" name="num">
                 <option value="10">10</option>
                 <option value="20">20</option>
                 <option value="0">所有</option>
             </select>项 <br>自定义分隔符<input type="text" id="splitter" name="splitter"> <input
                 type="submit" value="提交" /> <input type="button"
                 onclick="if(confirm('确认重置？')){reset()}" value="重置">
         </form>
     </div>

文件上传：

         request.setCharacterEncoding("UTF-8");
         response.setCharacterEncoding("UTF-8");
         response.setContentType("text/html");
         PrintWriter out = response.getWriter();
         byte b[] = new byte[2048];
         @SuppressWarnings("unused")
         int len = 0;
         Part p = request.getPart("upfilename");
         if(p==null){
             System.out.println("p == null");
         }
         String splitter = request.getParameter("splitter");
         InputStream in = p.getInputStream();
         String name = ""+System.currentTimeMillis();
         FileWriter fr = new FileWriter("D:\\upload\\" + name+".txt");
         while ((len = in.read(b)) > 0) {
             fr.write(new String(b));
         }
         fr.close();
         out.println("uploaded");
         response.sendRedirect("wordcount?id="+name+"&splitter"+splitter);
         out.flush();
         out.close();

servlet处理：

     protected void doGet(HttpServletRequest request, HttpServletResponse response)
             throws ServletException, IOException {
         request.setCharacterEncoding("UTF-8");
         response.setCharacterEncoding("UTF-8");
         response.setContentType("text/html");
         PrintWriter out = response.getWriter();
         String id = request.getParameter("id");
         int num = 10;
         String filename = "D:\\upload\\" + id + ".txt";
         WordUtil wu = WordUtilFactory.getWordUtil();
         long start = System.currentTimeMillis();
         String splitter = request.getParameter("splitter");
         wu.setSplitter(splitter);
         List<String[]> result = wu.getSortedWordGroupCountBuffered(filename, splitter);
         int size = result.size();
         for (int i = 0; i < (size > num ? num == 0 ? size : num : size); i++) {
             String[] strs = result.get(i);
             out.println(strs[1] + "  : " + strs[0] + "<br>");
         }
         long end = System.currentTimeMillis();
         out.println("execution time :" + (end - start) + "ms");
         out.flush();
         out.close();
     }

     protected void doPost(HttpServletRequest request, HttpServletResponse response)
             throws ServletException, IOException {
         request.setCharacterEncoding("UTF-8");
         response.setCharacterEncoding("UTF-8");
         response.setContentType("text/html");
         PrintWriter out = response.getWriter();
         String content = request.getParameter("content");
         String numStr = request.getParameter("num");
         int num = 10;
         if (numStr != null) {
             num = Integer.parseInt(numStr);
         }
         WordUtil wu = WordUtilFactory.getWordUtil();

         long start = System.currentTimeMillis();
         String filename = "D://tmp.txt";

         FileWriter fr = new FileWriter(filename);
         fr.write(content);
         fr.close();
         String splitter = request.getParameter("splitter");
         wu.setSplitter(splitter);
         List<String[]> result = wu.getSortedWordGroupCountBuffered(filename, splitter);
         int size = result.size();
         for (int i = 0; i < (size > num ? num == 0 ? size : num : size); i++) {
             String[] strs = result.get(i);
             out.println(strs[1] + "  : " + strs[0] + "<br>");
         }
         long end = System.currentTimeMillis();
         out.println("execution time :" + (end - start) + "ms");
         out.flush();
         out.close();
     }

有效字符判定（即自定义分隔符）

     public void setSplitter(String splitter) {
         char[] tmp = splitter.toCharArray();
         ArrayList<Character> deleted = new ArrayList<>();
         for(int i=0;i<tmp.length-1;i++){
             if(tmp[i]=='\\'){
                 char c = tmp[i+1];
                 if(c=='n'){
                     deleted.add('\n');
                 }
                 if(c=='r'){
                     deleted.add('\n');
                 }
                 if(c=='t'){
                     deleted.add('\n');
                 }
                 char[] copy = new char[tmp.length-2];
                 for(int j = 0;j <i;j++){
                     copy[j]=tmp[j];
                 }
                 for(int j=i;j<tmp.length-2;j++){
                     copy[j]=tmp[j+2];
                 }
                 i++;
             }
         }
         split = new char[tmp.length+deleted.size()];
         for(int i = 0;i<tmp.length;i++){
             split[i]=tmp[i];
         }
         for(int i=tmp.length;i<split.length;i++){
             split[i]=deleted.get(split.length-tmp.length-1);
         }
     }

     private int isCharacter(char ch, String splitter) {
         if (split == null) {
             if ((ch >= 'a' && ch <= 'z'))
                 return 1;
             if ((ch >= 'A' && ch <= 'Z'))
                 return 1;
             if (ch >= '0' && ch <= '9')
                 return 2;
             return 0;
         }
         if (split.equals("")) {
             if ((ch >= 'a' && ch <= 'z'))
                 return 1;
             if ((ch >= 'A' && ch <= 'Z'))
                 return 1;
             if (ch >= '0' && ch <= '9')
                 return 2;
             return 0;
         }
         for (int i = 0; i < split.length; i++) {
             if (ch == split[i]) {
                 return 0;
             }
         }
         if ((ch >= 'a' && ch <= 'z'))
             return 1;
         if ((ch >= 'A' && ch <= 'Z'))
             return 1;
         if (ch >= '0' && ch <= '9')
             return 2;
         return 1;
     }

 运行截图：

web版工程地址：https://git.coding.net/jx8zjs/wordcount-web.git

ssh:  git@git.coding.net:jx8zjs/wordcount-web.git

console版工程地址：https://coding.net/u/jx8zjs/p/wordCount/git

ssh:  git@git.coding.net:jx8zjs/wordCount.git