对于文章的字母、单词、短语，（无用词表）的检索Java代码实现

日期：2019.5.9

博客期：073

星期四

 　　今天软件工程课上，又做了测试，老师说我们的速度太慢了，实际上我也觉得自己很慢。老师说了这是我们的上一届的大二上半学期学习中的速度，所以呢？意思就是说我们和上一届的学长学姐们相比差的是天与地的距离啊！emmmm......唉~我也承认了！以下是我提交的源码，请各位欣赏，有疑问的话，评论区里见！

 源码：

package basic;

public class Chara {
    public char str;
    public int times;
}

package basic;

import java.io.File;
import java.util.ArrayList;
import java.util.List;
import java.util.Scanner;

public class Justice {
    public List <Chara> list = new ArrayList<Chara>();
    public int length = 0;
    public boolean isAccess(char c){
        return ((c>='a'&&c<='z')||(c>='A'&&c<='Z'));
    }
    public double p(char c){
        int seat = (changeToSmall(c)-'a');
        int time = list.get(seat).times;
        return (double)((double)time/(double)length);
    }
    public char changeToSmall(char c){
        if(c>='a'&&c<='z')
            return c;
        else
            return (char)(c-'A'+'a');
    }
    public void dealFile(String filename) throws Exception{
        File f = new File(filename);
        if(!f.exists())
        {
            f.createNewFile();
        }
        Scanner sc = new Scanner(f);
        while(sc.hasNext())
        {
            String s = sc.next();
            for(int t=0;t<s.length();++t)
            {
                char c = s.charAt(t);
                if(this.isAccess(c))
                {
                    int seat = (changeToSmall(c)-'a');
                    Chara ch = list.get(seat);
                    ch.times++;
                    list.set(seat, ch);
                    ++length;
                }
            }
        }
        
        sc.close();
    }
    public void changeChannel(){
        int size = list.size();
        for(int i=0;i<size;++i)
        {
            for(int j=0;j<size-1;++j)
            {
                if(list.get(j).times<list.get(j+1).times)
                {
                    Chara temp = list.get(j);
                    list.set(j,list.get(j+1));
                    list.set(j+1, temp);
                }
            }
        }
    }
    public void display(){
        int leng = list.size();
        for(int i=0;i<leng;++i)
        {
            char c = (char)(i+'a');
            System.out.println((char)(i+'a')+"出现了"+(list.get(i).times)+"次，它的频率是"+p(c)+"	");
        }
    }
    public Justice(){
        for(char i='a';i<='z';++i)
        {
            Chara c = new Chara();
            c.str = i;
            c.times = 0;
            list.add(c);
        }
    }
}

package basic;

public class Main {
    public static void main(String[] args) throws Exception {
        Justice justice = new Justice();
        justice.dealFile("txt/piao.txt");
        justice.changeChannel();
        justice.display();
    }
}

package more;

public class Chara {
    public String str = "";
    public int times = 0;
    public Chara(){
        
    }
    public Chara(String str,int times){
        this.str = str;
        this.times = times;
    }
}

package more;

import java.io.File;
import java.io.FileNotFoundException;
import java.util.ArrayList;
import java.util.List;
import java.util.Scanner;

public class EditTable{

    public List<Chara> list = new ArrayList<Chara>();
    public List<String> listString = new ArrayList<String>();
    public int length = 0;
    public boolean isAddmitted = false;
    //单词处理
    public String changeWord(String str){
        String newStr = str.replace("“","");
        newStr = newStr.replace(",","");
        newStr = newStr.replace("”","");
        newStr = newStr.replace(".","");
        newStr = newStr.replace("?","");
        newStr = newStr.replace("!","");
        newStr = newStr.replace(":","");
        return newStr.toLowerCase();
    }
    public boolean isInUnUseTable(String str){
        int leng = listString.size();
        for(int i=0;i<leng;++i)
        {
            if(str.toLowerCase().compareTo(listString.get(i))==0)
                return true;
        }
        return false;
    }
    //判断是否可以通过
    public boolean isAccess(String str){
        int leng = str.length();
        str = this.changeWord(str);
        
        if(isAddmitted)
        {
            if(isInUnUseTable(str))
                return false;
        }
        

        for(int i=0;i<leng;++i)
        {
            char c = str.charAt(i);
            if(!((c>='a'&&c<='z')||(c>='A'&&c<='Z')))
                return false;
        }
        return true;
    }
    public boolean isAddmitted() {
        return isAddmitted;
    }
    public void setAddmitted(boolean isAddmitted) {
        this.isAddmitted = isAddmitted;
    }
    //判断位置
    public int seatAt(String c){
        int leng = list.size();
        for(int i=0;i<leng;++i)
        {
            String trp = list.get(i).str;
            if(trp.compareTo(c)==0)
                return i;
        }
        return -1;
    }
    //判断并添加
    public void AddString(String str){
        if(isAccess(changeWord(str)))
        {
            int seat = seatAt(changeWord(str));
            if(seat==-1)
            {
                Chara chara = new Chara(changeWord(str),1);
                list.add(chara);
            }
            else
            {
                Chara chara = list.get(seat);
                chara.times++;
                list.set(seat, chara);
            }
            ++length;
        }
    }
    //排序
    public void changeChannel(){
        int size = list.size();
        for(int i=0;i<size;++i)
        {
            for(int j=0;j<size-1;++j)
            {
                if(list.get(j).times<list.get(j+1).times)
                {
                    Chara temp = list.get(j);
                    list.set(j,list.get(j+1));
                    list.set(j+1, temp);
                }
            }
        }
    }
    //概率
    public double p(String str){
        int seat = seatAt(changeWord(str));
        if(seat==-1)
            return 0;
        int time = list.get(seat).times;
        return ((double)time/(double)length);
    }
    //展示
    public void display(int seat){
        Chara cha = list.get(seat);
        System.out.println("单词："+cha.str+(cha.str.length()<=2?"		":"	")+"出现次数："+cha.times+"	"+"概率为："+((double) Math.round(p(cha.str) * 10000) / 100)+"%	");
    }
    public void Display(int N){
        for(int i=0;i<N&&i<list.size();++i)
            display(i);
    }
    public EditTable() throws FileNotFoundException{

        Scanner sc = new Scanner(new File("txt/stopword.txt"));
        while(sc.hasNext())
        {
            String str = sc.next();
            listString.add(str);
        }
    }
    public EditTable(boolean isAd) throws FileNotFoundException{
        this.isAddmitted = isAd;
        Scanner sc = new Scanner(new File("txt/stopword.txt"));
        while(sc.hasNext())
        {
            String str = sc.next();
            listString.add(str);
        }
    }
    //处理文件
    public void DealFile(String fileName) throws Exception{
        File f = new File(fileName);
        Scanner sc = new Scanner(f);
        while(sc.hasNext())
        {
            String str = sc.next();
            AddString(str);
        }
        sc.close();
    }

}

package more;

import java.io.File;
import java.io.FileNotFoundException;
import java.util.ArrayList;
import java.util.List;
import java.util.Scanner;

import more.Chara;

public class Table {
    public List<Chara> list = new ArrayList<Chara>();
    public List<String> listString = new ArrayList<String>();
    public int length = 0;
    public boolean isAddmitted = false;
    //单词处理
    public String changeWord(String str){
        String newStr = str.replace("“","");
        newStr = newStr.replace(",","");
        newStr = newStr.replace("”","");
        newStr = newStr.replace(".","");
        newStr = newStr.replace("?","");
        newStr = newStr.replace("!","");
        newStr = newStr.replace(":","");
        return newStr.toLowerCase();
    }
    public static boolean isContainSpecifical(String str){
        if(str.contains("“"))
            return false;
        if(str.contains("”"))
            return false;
        if(str.contains(","))
            return false;
        if(str.contains("."))
            return false;
        if(str.contains("?"))
            return false;
        if(str.contains("!"))
            return false;
        if(str.contains(":"))
            return false;
        return true;
    }
    public boolean isInUnUseTable(String str){
        int leng = listString.size();
        for(int i=0;i<leng;++i)
        {
            if(str.toLowerCase().compareTo(listString.get(i))==0)
                return true;
        }
        return false;
    }
    //判断是否可以通过
    public boolean isAccess(String str){
        int leng = str.length();
        str = this.changeWord(str);
        
        if(isAddmitted)
        {
            if(isInUnUseTable(str))
                return false;
        }
        

        for(int i=0;i<leng;++i)
        {
            char c = str.charAt(i);
            if(!((c>='a'&&c<='z')||(c>='A'&&c<='Z')))
                return false;
        }
        return true;
    }
    public boolean isAddmitted() {
        return isAddmitted;
    }
    public void setAddmitted(boolean isAddmitted) {
        this.isAddmitted = isAddmitted;
    }
    //判断位置
    public int seatAt(String c){
        int leng = list.size();
        for(int i=0;i<leng;++i)
        {
            String trp = list.get(i).str;
            if(trp.compareTo(c)==0)
                return i;
        }
        return -1;
    }
    //判断并添加
    public void AddString(String str){
        if(isAccess(changeWord(str)))
        {
            int seat = seatAt(changeWord(str));
            if(seat==-1)
            {
                Chara chara = new Chara(changeWord(str),1);
                list.add(chara);
            }
            else
            {
                Chara chara = list.get(seat);
                chara.times++;
                list.set(seat, chara);
            }
            ++length;
        }
    }
    //排序
    public void changeChannel(){
        int size = list.size();
        for(int i=0;i<size;++i)
        {
            for(int j=0;j<size-1;++j)
            {
                if(list.get(j).times<list.get(j+1).times)
                {
                    Chara temp = list.get(j);
                    list.set(j,list.get(j+1));
                    list.set(j+1, temp);
                }
            }
        }
    }
    //概率
    public double p(String str){
        int seat = seatAt(changeWord(str));
        if(seat==-1)
            return 0;
        int time = list.get(seat).times;
        return ((double)time/(double)length);
    }
    //展示
    public void display(int seat){
        Chara cha = list.get(seat);
        System.out.println("单词："+cha.str+(cha.str.length()<=2?"		":"	")+"出现次数："+cha.times+"	"+"概率为："+((double) Math.round(p(cha.str) * 10000) / 100)+"%	");
    }
    public void Display(int N){
        for(int i=0;i<N&&i<list.size();++i)
            display(i);
    }
    public Table() throws FileNotFoundException{

        Scanner sc = new Scanner(new File("txt/stopword.txt"));
        while(sc.hasNext())
        {
            String str = sc.next();
            listString.add(str);
        }
    }
    public Table(boolean isAd) throws FileNotFoundException{
        this.isAddmitted = isAd;
        Scanner sc = new Scanner(new File("txt/stopword.txt"));
        while(sc.hasNext())
        {
            String str = sc.next();
            listString.add(str);
        }
    }
    //读取
    public void AddMenu(List <String> sqlList){
        int length = sqlList.size();
        String ss = "";
        if(length!=0)
            ss = sqlList.get(length-1);
        for(int i=length-2;i>=0;--i)
        {
            ss = sqlList.get(i) + " " + ss;
            AddString(ss);
        }
    }
    //处理文件
    public void DealFile(String fileName) throws Exception{
        File f = new File(fileName);
        Scanner sc = new Scanner(f);
        boolean isT = true;
        while(sc.hasNext())
        {
            while(isT&&sc.hasNext())
            {
                List <String> strSql = new ArrayList<String>();
                String str = sc.next();
                if(isInUnUseTable(str))
                    break;
                strSql.add(str);
                AddMenu(strSql);
                if(Table.isContainSpecifical(str))
                    break;
            }
        }
        sc.close();
    }
}

package running;

import java.io.File;
import java.io.FileWriter;
import java.io.PrintWriter;
import java.util.Scanner;

import more.EditTable;
import more.Table;

import basic.Justice;

public class TestMain {
    //第0步
    public static void process0(String file) throws Exception{
        Justice justice = new Justice();
        justice.dealFile(file);
        justice.changeChannel();
        justice.display();
    }
    //第1步
    public static void process1(String file) throws Exception{
        EditTable table = new EditTable();
        table.DealFile(file);
        table.changeChannel();
        table.Display(99999999);
    }
    //第2步
    public static void process2(String file,int num) throws Exception{
        EditTable table = new EditTable();
        table.DealFile(file);
        table.changeChannel();
        table.Display(num);
    }
    //第3步
    public static void process3(String file,int num) throws Exception{
        EditTable table = new EditTable(true);
        table.DealFile(file);
        table.changeChannel();
        table.Display(num);
    }
    //第4步
    public static void process4(String file,int num) throws Exception{
        Table table = new Table();
        table.DealFile(file);
        table.changeChannel();
        table.Display(999999);
    }
    //停用词
    public static void stopWord(String file) throws Exception{
        File f = new File("txt/stopword.txt");
        PrintWriter pw = new PrintWriter(new FileWriter(f,true));
        pw.println();
        pw.println(file);
        pw.close();
    }
    //主
    public static void main(String[] args) throws Exception {
        Scanner sc = new Scanner(System.in);
        boolean isAccess = true;
        while(isAccess)
        {
            System.out.println("===============================================");
            System.out.println(" p0 + 文件路径                    ----------->      分析字母");
            System.out.println(" p1 + 文件路径                    ----------->      分析单词");
            System.out.println(" p2 + 文件路径  + 前几项     -----------> 分析频率最高的单词");
            System.out.println(" p3 + 文件路径  + 前几项     -----------> 分析频率最高的单词(过滤)");
            System.out.println(" sw + 单词                           ----------->      添加停用词");
            System.out.println(" p4 + 文件路径  + 前几项    ----------->      分析短语");
            System.out.println(" q                     ----------->        退出");
            System.out.println();
            String order = sc.next();
            String file = sc.next();
            if(order.compareTo("p0")==0)
            {
                TestMain.process0(file);
            }
            else if(order.compareTo("p1")==0)
            {
                TestMain.process1(file);
            }
            else if(order.compareTo("p2")==0)
            {
                int num = sc.nextInt();
                TestMain.process2(file,num);
            }
            else if(order.compareTo("p3")==0)
            {
                int num = sc.nextInt();
                TestMain.process3(file,num);
            }
            else if(order.compareTo("sw")==0)
            {
                TestMain.stopWord(file);
            }
            else if(order.compareTo("p4")==0)
            {
                int num = sc.nextInt();
                TestMain.process4(file,num);
            }
            else if(order.compareTo("q")==0)
            {
                System.out.println("欢迎下次使用！");
                break;
            }
            System.out.println("===============================================");
        }
        sc.close();
    }
}

package more;

import java.io.File;

public class Test {
    public static void TestForPath(String path) throws Exception{
        File file = new File(path);
        if(file.isDirectory())
        {
            File []fl = file.listFiles();
            int leng = fl.length;
            for(int i=0;i<leng;++i)
            {
                String fileName = path+"/"+fl[i].getName();
                EditTable table = new EditTable();
                System.out.println("====================================================");
                System.out.println("文件名称："+fl[i].getName());
                table.DealFile(fileName);
                table.changeChannel();
                table.Display(40);
            }
        }
    }
    public static void main(String[] args) throws Exception {
        TestForPath("txt");
    }
}

package running;

import java.io.File;
import java.io.FileWriter;
import java.io.PrintWriter;
import java.util.Scanner;

import more.EditTable;
import more.Table;

import basic.Justice;

public class TestMain {
    //第0步
    public static void process0(String file) throws Exception{
        Justice justice = new Justice();
        justice.dealFile(file);
        justice.changeChannel();
        justice.display();
    }
    //第1步
    public static void process1(String file) throws Exception{
        EditTable table = new EditTable();
        table.DealFile(file);
        table.changeChannel();
        table.Display(99999999);
    }
    //第2步
    public static void process2(String file,int num) throws Exception{
        EditTable table = new EditTable();
        table.DealFile(file);
        table.changeChannel();
        table.Display(num);
    }
    //第3步
    public static void process3(String file,int num) throws Exception{
        EditTable table = new EditTable(true);
        table.DealFile(file);
        table.changeChannel();
        table.Display(num);
    }
    //第4步
    public static void process4(String file,int num) throws Exception{
        Table table = new Table();
        table.DealFile(file);
        table.changeChannel();
        table.Display(999999);
    }
    //停用词
    public static void stopWord(String file) throws Exception{
        File f = new File("txt/stopword.txt");
        PrintWriter pw = new PrintWriter(new FileWriter(f,true));
        pw.println();
        pw.println(file);
        pw.close();
    }
    //主
    public static void main(String[] args) throws Exception {
        Scanner sc = new Scanner(System.in);
        boolean isAccess = true;
        while(isAccess)
        {
            System.out.println("===============================================");
            System.out.println(" p0 + 文件路径                    ----------->      分析字母");
            System.out.println(" p1 + 文件路径                    ----------->      分析单词");
            System.out.println(" p2 + 文件路径  + 前几项     -----------> 分析频率最高的单词");
            System.out.println(" p3 + 文件路径  + 前几项     -----------> 分析频率最高的单词(过滤)");
            System.out.println(" sw + 单词                           ----------->      添加停用词");
            System.out.println(" p4 + 文件路径  + 前几项    ----------->      分析短语");
            System.out.println(" q                     ----------->        退出");
            System.out.println();
            String order = sc.next();
            String file = sc.next();
            if(order.compareTo("p0")==0)
            {
                TestMain.process0(file);
            }
            else if(order.compareTo("p1")==0)
            {
                TestMain.process1(file);
            }
            else if(order.compareTo("p2")==0)
            {
                int num = sc.nextInt();
                TestMain.process2(file,num);
            }
            else if(order.compareTo("p3")==0)
            {
                int num = sc.nextInt();
                TestMain.process3(file,num);
            }
            else if(order.compareTo("sw")==0)
            {
                TestMain.stopWord(file);
            }
            else if(order.compareTo("p4")==0)
            {
                int num = sc.nextInt();
                TestMain.process4(file,num);
            }
            else if(order.compareTo("q")==0)
            {
                System.out.println("欢迎下次使用！");
                break;
            }
            System.out.println("===============================================");
        }
        sc.close();
    }
}

 附上截图：