词法分析器C++读取外部txt（编译原理）

词法分析器C++读取外部txt（编译原理）

词法分析器的参考文章：
https://blog.csdn.net/qq_32623363/article/details/79911321)
在原文章的基础上，我加入了关键字的识别，并将词法分析器设计为一个独立的程序，从文件中读取源程序，将其变换成相应的符号序列；

c++读取txt文件的参考文章https://blog.csdn.net/m0_38033475/article/details/94394087

#include <iostream>
#include <string>
#include<fstream>
using namespace std;

#define MAX_LEN 200

//将一个字符串内容清空
void clearArray(char *c)
{
    int len = strlen(c);
    for(int i = 0; i < len; i++)
    {
        c[i] = '\0';
    }
}

//判断是否为数字，true为是，false为不是
bool isDigit(char c){
    if(c >= '0' && c <= '9')
    {
        return true;
    }
    else
    {
        return false;
    }
}

//判断是否为字母，ture为是，false为不是
bool isChar(char c)
{
    if( (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') )
    {
        return true;
    }
    else
    {
        return false;
    }
}


/*
* 单词输出函数
* 该单词的意义
* 经过词法分析器分析后得到的单个单词
*/
void print(string label, char *word)
{
    ofstream outfile("output.txt",ios::app);
    outfile << label << ":  " << word << endl;
}



/*
* 词法分析函数
* 输入字符串指针
* 该函数对输入串进行分析，并分割成单词后输出
*/
void analysis(char* inChar)
{

    //用于对词法分析后的单词进行保存
    char tempChar[MAX_LEN] = {0};
    // 作为tempChar的一个索引
    int j = 0;

    //用于遍历inChar字符数组
    int i = 0;
    while(inChar[i] != '\0')
    {
        switch(inChar[i])
        {
            case '-':
                tempChar[j++] = inChar[i];
                print("减号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '=':
                tempChar[j++] = inChar[i];
                print("等号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '+':
                tempChar[j++] = inChar[i];
                print("加号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case ',':
                tempChar[j++] = inChar[i];
                print("逗号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case ';':
                tempChar[j++] = inChar[i];
                print("分号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '/':
                tempChar[j++] = inChar[i];
                print("除号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '(':
                tempChar[j++] = inChar[i];
                print("左括号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case ')':
                tempChar[j++] = inChar[i];
                print("右括号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '<':
                tempChar[j++] = inChar[i];
                print("小于号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '>':
                tempChar[j++] = inChar[i];
                print("大于号",tempChar);
                j = 0;
                clearArray(tempChar);
                break;
            case '*':
                if(inChar[i+1] != '*')
                {
                    tempChar[j++] = inChar[i];
                    print("乘法",tempChar);
                    j = 0;
                    clearArray(tempChar);
                }
                else
                {
                    tempChar[j++] = inChar[i++];
                    tempChar[j++] = inChar[i];
                    print("平方",tempChar);
                    j = 0;
                    clearArray(tempChar);
                }
                break;
            default:
                //判断是否为字母+数字
                if(isChar(inChar[i]))
                {
                    if(inChar[i]=='i' && inChar[i+1]=='f')
                    {
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i];
                        print("关键字",tempChar);
                        j=0;
                        clearArray(tempChar);
                        i++;
                        continue;
                    }
                    else if(inChar[i]=='d' && inChar[i+1]=='o')
                    {
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i];
                        print("关键字",tempChar);
                        j=0;
                        clearArray(tempChar);
                        i++;
                        continue;
                    }
                     else if(inChar[i]=='t' && inChar[i+1]=='h' && inChar[i+2]=='e' && inChar[i+3]=='n')
                    {
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i];
                        print("关键字",tempChar);
                        j=0;
                        clearArray(tempChar);
                        i++;
                        continue;
                    }
                    else if(inChar[i]=='e' && inChar[i+1]=='l' && inChar[i+2]=='s' && inChar[i+3]=='e')
                    {
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i];
                        print("关键字",tempChar);
                        j=0;
                        clearArray(tempChar);
                        i++;
                        continue;
                    }
                    else if(inChar[i]=='w' && inChar[i+1]=='h' && inChar[i+2]=='i' && inChar[i+3]=='l' && inChar[i+4]=='e')
                    {
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i++];
                        tempChar[j++] = inChar[i];
                        print("关键字",tempChar);
                        j=0;
                        clearArray(tempChar);
                        i++;
                        continue;
                    }
                    else
                     {
                        j = 0;
                        tempChar[j++] = inChar[i];
                        while (isChar(inChar[i + 1]) || isDigit(inChar[i + 1]))
                        {
                            i++;
                            tempChar[j++] = inChar[i];
                        }
                        j = 0;
                        print("标识符", tempChar);
                        clearArray(tempChar);
                     }
                }
                else if(isDigit(inChar[i]))
                {
                    j = 0;
                    tempChar[j++] = inChar[i];
                    while(isDigit(inChar[i+1]))
                    {
                        i++;
                        tempChar[j++] = inChar[i];
                    }
                    j = 0;
                    print("数字",tempChar);
                    clearArray(tempChar);
                }
                else
                {
                    tempChar[j++] = inChar[i];
                    print("其他",tempChar);
                    j = 0;
                    clearArray(tempChar);
                }

        }
        i++;

    }
}

int main()
{
    char inChar[MAX_LEN] = {0};
    ifstream myfile("input.txt");
    ofstream outfile("output.txt",ios::app); //ios::app指追加写入

    //获取输入字符串，输入的字符串以空格、回车、换行作为结束符
    myfile>>inChar;

    //对输入的字符串进行分析
    analysis(inChar);

    myfile.close();
    outfile.close();
    
    return 0;
}

输入示例：

输出：