frist step of project

你好，世界，

今天是美好的一天

你好吗世界，世界很美好

 No newline at end of file

package frty;

import java.io.*;

import java.util.*;

import com.hankcs.hanlp.seg.common.Term;

import com.hankcs.hanlp.tokenizer.StandardTokenizer;

public class onr {

	public static void main(String[] args) throws IOException {

		File file = new File("../123.txt");

		//the address of file which you need read

		System.out.println(onr.devideword(onr.read(file)));

		System.out.println(onr.fre(onr.devideword(onr.read(file))));

		//just two simple test

	}

	public static String read(File file) throws IOException{

		BufferedReader br=new BufferedReader(new InputStreamReader(new FileInputStream(file),"UTF-8"));

		String line = null;

		String lines = "";

		while ((line = br.readLine()) != null) {   

			lines+=line;  

		}

		return lines.replaceAll

				("[`~!@#$%^&*()+=|{}':;',\\[\\].<>/?~！@#￥%……& amp;*（）——+|{}【】‘；：”“’。，、？|-]", ""); 

		//lines is a string of your flie without space and punctuation.

	}

	public static ArrayList<String>  devideword(String read) {

		List<Term> outer = StandardTokenizer.segment(read);

		ArrayList<Integer> k=new ArrayList<>();

		ArrayList<String> s=new ArrayList<>();

		for(int j=0;j<outer.size();j++) {

			for(int i=0;i<outer.get(j).toString().length();i++) {

				if(outer.get(j).toString().charAt(i)=='/') {k.add(i);}

			}

			s.add(outer.get(j).toString().substring(0, k.get(j)));

		}

		return s;

		// s is a string array include the words of sentence

	}

	public static ArrayList<Integer> fre(ArrayList<String> s) {

		ArrayList<Integer> n=new ArrayList<>();

		for(int i=0;i<s.size();i++) {

			int count=1;

			int j=i+1;

			while(j<s.size()) {

				if(s.get(i).equals(s.get(j))) {

					s.remove(j);

					count++;

				}

				else {j++;}

			}

			n.add(count);

		}

		//System.out.println(s);

		//s is an arraylist which just has one of each word

		return n;

		//n is the frequency of s

	}

}