KMP 算法(上篇)

/*
 * KMP 算法
 * 要点如下：
 *      1 如何构建next表
 *      2 如何利用next表进行模式匹配
 *          指针位置的控制问题
 */

#include <cstdio>
#include <cstring>

void get_next(char *str, int *next) {
    size_t len = strlen(str);
    memset(next, 0, len * sizeof(int));

    if (len==0) return;

    next[0] = -1;
    char *p = str+1;
    while (*p != '\0') {
        char *q = str + next[p-1-str];
        while (q >= str) {
            if (*(p-1)==*q) {
                next[p-str] = next[q-str] + 1;
                break;
            }
            q = str + next[q-str];
        }
        if (q < str) {
            next[p-str] = next[0] + 1;
        }
        ++p;
    }

    next[0] = 0;
}

char *KMP_strstr(char *str1, char *str2) {
    int *next = new int[strlen(str2)];
    get_next(str2, next);

    int index = 0;
    char *p = str1;
    char *q = str1;
    while (*q != '\0') {
        char *s = str2 + index;
        while (*s==*q && *s!='\0')
            ++s, ++q;

        if (*s=='\0') {
            break;
        }

        // 控制 s 的位置
        index = next[s-str2];

        // 控制 q 的位置
        if (s==str2) ++q;

        // 控制 p 的位置
        p = q - index;
    }

    delete [] next;
    return p;
}

int main() {
    char *str1 = "ababaaababababcaaaa";
    char str[] = "ababababca";
    char *pch = KMP_strstr(str1, str);
    printf("%d\n", pch-str1+1);
    return 0;
}

/*
 * KMP 算法
 * 要点如下：
 *      1 如何构建next表
 *      2 如何利用next表进行模式匹配
 *          指针位置的控制问题
 */

#include <cstdio>
#include <cstring>

void get_next(char *str, int *next) {
    next[0] = -1;
    char *p = str+1;
    while (*p != '\0') {
        char *q = str + next[p-1-str];
        while (q >= str) {
            if (*(p-1) == *q) {
                next[p-str] = next[q-str] + 1;
                break;
            }
            q = str + next[q-str];
        }
        if (q < str) {
            next[p-str] = next[0] + 1;
        }
        ++p;
    }
}

char *KMP_strstr(char *str1, char *str2) {
    int *next = new int[strlen(str2)];
    get_next(str2, next);

    char *p = str1;
    char *q = str2;
    while (*p != '\0' && *q != '\0') {
        if (q < str2 || *p==*q) ++p, ++q;
        else q = str2 + next[q-str2];
    }

    delete [] next;
    if (*q == '\0') return p - (q - str2);
    return nullptr;
}

int main() {
    char *str1 = "ababaaababababcaaaa";
    char str[] = "ababababca";
    char *pch = KMP_strstr(str1, str);
    printf("%d\n", pch-str1+1);
    return 0;
}

/*
 * KMP 算法
 * 要点如下：
 *      1 如何构建next表
 *      2 如何利用next表进行模式匹配
 *          指针位置的控制问题
 *  这个是精炼的算法，不好理解
 */

#include <cstdio>
#include <cstring>

void get_next(char *str, int *next) {
    next[0] = -1;
    char *p = str;
    char *q = str-1;
    while (*(p+1) != '\0') {
        if (q<str || *p==*q)  {
            ++q, ++p;
            if (*p != *q) next[p-str] = q-str;
            else next[p-str] = next[q-str];
        }
        else q = str + next[q-str];
    }
}

char *KMP_strstr(char *str1, char *str2) {
    int *next = new int[strlen(str2)];
    get_next(str2, next);

    char *p = str1;
    char *q = str2;
    while (*p != '\0' && *q != '\0') {
        if (q < str2 || *p==*q) ++p, ++q;
        else q = str2 + next[q-str2];
    }

    delete [] next;
    if (*q == '\0') return p - (q - str2);
    return nullptr;
}

int main() {
    char *str1 = "ababaaababababcaaaa";
    char str[] = "ababababca";
    char *pch = KMP_strstr(str1, str);
    printf("%d\n", pch-str1+1);
    return 0;
}