数据结构---串的朴素模式匹配算法（暴力算法）和kmp ，next代码

2022.2.15

子串匹配 

#include <stdio.h>
#include <stdbool.h>
#include <stdlib.h>
#include <malloc.h>

typedef struct Str
{
    char *ch;
    int length;
} Str;

//比较操作
int StrCompare(Str *s1, Str *s2)
{
    int i = s1->length, j = s2->length;

    for (int a = 0; a < i && a < j; a++)
    {
        if (s1->ch[a] != s2->ch[a])
        {
            return (s1->ch[a]) - (s2->ch[a]);
        }
    }
    // for循环结束 则说明其中一个串已经完了且两个串前面全部相同 此时比串的长度就行
    return (s1->length) - (s2->length);
}

//求长度
int length(Str *s)
{
    if (s->ch[0] == '\0' || s->ch == NULL)
        return 0;
    int i = 0;
    while (s->ch[i] != '\0')
    {
        i++;
    }
    return i;
}

//求子串操作   pos为起始位置   len为子串的总长度
bool Substring(Str *s, Str *r, int pos, int len)
{
    if (pos < 0 || pos > r->length || len < 0 || pos + len > r->length)
        return false;
    s->ch = (char *)malloc(sizeof(char) * (len + 1));
    int j = 0;
    if (len == 0)
    {
        s->ch = '\0';
        s->length = 0;
        return true;
    }
    else
    {
        for (int i = pos; i < pos + len; i++, j++)
        {
            s->ch[j] = r->ch[i];
        }
        s->ch[j] = '\0';
        s->length = len;
        return true;
    }
}

//暴力算法1 用已知函数+循环
int Index(Str *str, Str *s)
{

    if (str->ch[0] == '\0' || s->ch[0] == '\0')
        return 0;

    Str *c = (Str *)malloc(sizeof(Str));
    int i = 0, j = (str->length) - (s->length) + 1;
    for (i; i < j; i++)
    {
        c->ch = (char *)malloc(sizeof(char) * (s->length + 1));
        Substring(c, str, i, s->length);
        if (!StrCompare(c, s))
        {
            return i + 1;
        }
    }
    return 0;
}

//暴力算法2 回溯数组下标算法
int Index2(Str *str, Str *s)
{
    if (s->ch[0] == '\0')
        return 0;
    int i = 0, j = 0; // i对应str j对应s
    int a = length(str), b = length(s);

    for (i, j; i < a && j < b;)
    {
        if (str->ch[i] == s->ch[j])
            i++, j++;
        else
        {
            i = i - j + 1;
            j = 0;
        }
    }

    if (j == b)
        return i - j + 1;
    else
        return 0;
}

//最大相同前后缀数表整体右移，0位置赋0，全部-1得next数组
// next数组0位置赋值为-1

//递推求next数组
void GetNext(char *p, int next[])
{
    int Length = 7;
    next[0] = -1;
    int k = -1;
    int j = 0;

    while (j < Length - 1) // next最后的下标是j+1 所以j最多到len-2
    {
        if (k == -1 || p[j] == p[k])
        {
            next[j + 1] = k + 1;
            j++;
            k++;
        }
        else
        {
            k = next[k];
        }
    }
}


// kmp算法
int KMP(char *s, char *p, int next[])
{
    int i = 0, j = 0;
    int a = 7, b = 2;

    while (i < a && j < b)
    {
        if (s[i] == p[j] || j == -1)
        {
            i++, j++;
        }
        else
        {
            j = next[j];
        }
    }
    if (j == b)
        return i - j + 1;
    else
        return 0;
}



int main()
{

    Str *s, *s1;
    s = (Str *)malloc(sizeof(Str));
    s1 = (Str *)malloc(sizeof(Str));

    s->ch = "Hua Wei P30-------.-";
    s1->ch = "--.-";
    s->length = length(s);
    s1->length = length(s1);

    int b = Index2(s, s1);
    printf("%d\n", b);

    int a = Index(s, s1);
    printf("%d\n", a);

    printf("------------------------------\n");

    char *p = "ABCDABD";
    int arr[7] = {0};
    GetNext(p, arr);
    int c = KMP(p, "BD", arr);
    printf("%d\n", c);

    return 0;
}