java 处理人民日报98年语料库处理复合标注词

来源：互联网发布：dnf 登陆器源码编辑：程序博客网时间：2024/05/16 03:14

主要处理nz i l

public static void main(String[] args) {

       // TODO Auto-generated method stub

       StringBuffer sb = new StringBuffer();
       String tempstr = null;
       try {
           FileWriter fp = new FileWriter("E:/text/destext.txt");
           String path = "E:/text/srctext.txt";
           File file = new File(path);
           if (!file.exists())
               throw new FileNotFoundException();

           Pattern pattern1 = Pattern.compile("\\[.*?]n");
           Pattern pattern0 = Pattern.compile("\\[.*?]i");
           Pattern pattern2 = Pattern.compile("\\[.*?]l");

           Pattern pattern3 = Pattern.compile("\\[.*?]nz");
           FileInputStream fis = new FileInputStream(file);
           BufferedReader br = new BufferedReader(new InputStreamReader(fis));
           while ((tempstr = br.readLine()) != null) {
               String line = "";
               String line1 = "";
               // sb.append(tempstr);

               Matcher matcher = pattern1.matcher(tempstr);

               while (matcher.find()) {


                   String src = "";
                   String tmp = "";
                   for (int j = 0; j <= matcher.groupCount(); j++) {
                       // System.out.print( matcher.group(j) +" ");
                       src = matcher.group(j);
                       tmp = matcher.group(j);
                       tmp = tmp.replaceAll("/\\w+", "");
                       tmp = tmp.replaceAll(" ", "");
                       tmp = tmp.replaceAll("\\[", "");
                       tmp = tmp.replaceAll("\\]", "/");
                       System.out.println(tmp);
                   }
                   tempstr = tempstr.replace(src, tmp);
                   System.out.println();
               }

               Matcher matcher1 = pattern0.matcher(tempstr);

               while (matcher1.find()) {


                   String src = "";
                   String tmp = "";
                   for (int j = 0; j <= matcher1.groupCount(); j++) {

                       src = matcher1.group(j);
                       tmp = matcher1.group(j);
                       tmp = tmp.replaceAll("/\\w+", "");
                       tmp = tmp.replaceAll(" ", "");
                       tmp = tmp.replaceAll("\\[", "");
                       tmp = tmp.replaceAll("\\]", "/");
                       System.out.println(tmp);
                   }
                   tempstr =tempstr.replace(src, tmp);
                   System.out.println();
               }

Matcher matcher2 = pattern2.matcher(tempstr);

               while (matcher2.find()) {

                   String src="";
                   String tmp ="";
                   for (int j = 0; j <= matcher2.groupCount(); j++)
                   {
                   //   System.out.print( matcher.group(j) +" ");
                       src = matcher2.group(j);
                       tmp= matcher2.group(j);
                       tmp = tmp.replaceAll("/\\w+", "");
                       tmp = tmp.replaceAll(" ", "");
                       tmp = tmp.replaceAll("\\[", "");
                       tmp = tmp.replaceAll("\\]", "/");
                       System.out.println(tmp);
                   }
                   tempstr =tempstr.replace(src, tmp);
                   System.out.println();
               }



               fp.write(tempstr + "\n");


           }

           fp.close();

       } catch (IOException ex) {
           System.out.println(ex.getStackTrace());
       }

   }

0 0