HDU 2222 Keywords Search【AC自动机|字典树】

AC机做法

202MS

32548K

#include <cstdio>
#include <cmath>
#include <cstring>
#include <string>
#include <set>
#include <map>
#include <stack>
#include <queue>
#include <vector>
#include <iostream>
#include <algorithm>
using namespace std;
#define ll long long
#define eps 10^(-6)
#define Q_CIN ios::sync_with_stdio(false);
#define REP( i , n ) for ( int i = 0 ; i < n ; ++ i )
#define FOR( i , a , b ) for ( int i = a ; i <= b ; ++ i )
#define CLR( a , x ) memset ( a , x , sizeof (a) );
#define RE freopen("1.in","r",stdin);
#define WE freopen("1.out","w",stdout);
#define MOD 10009
#define debug(x) cout<<#x<<":"<<(x)<<endl;
const int NODE_MAX=27;
const int NMAX=50*10000+5;
struct node
{
    node *next[NODE_MAX];
    int cnt;
    node *fail;
    node(){
        CLR(next,NULL);
        cnt=0;
        fail=NULL;
    }
};
node *q[NMAX];

void insert_tree(node *p,char s1[])
{
    int i=0;
    while(s1[i])
    {
        int k=s1[i]-'a';
        if(!p->next[k])
            p->next[k]=new node();
        p=p->next[k];
        i++;
    }
    p->cnt++;
}
void buildFail(node *rt)
{
    int head=0,tail=0;
    node *p=rt;
    p->fail=NULL;       //根结点fail为空
    q[tail++]=rt;
    while(head<tail){
        node *tmp=q[head++];
        for(int i=0;i<NODE_MAX;i++){
            if(tmp->next[i]){       //存在该结点
                if(tmp==rt)
                    tmp->next[i]->fail=rt;
                else{
                    p=tmp->fail;
                    while(p)
                    {       //其他相同字母的结点
                        if(p->next[i]){
                            tmp->next[i]->fail=p->next[i];
                            break;
                        }
                        p=p->fail;
                    }
                    if(!p){     //未找到
                        tmp->next[i]->fail=rt;
                    }
                }
                q[tail++]=tmp->next[i];
            }
        }
    }
}
int search_tree(node *rt,char s1[])
{
    int i=0,ans=0;
    node *p=rt;
    while(s1[i])
    {
        int k=s1[i]-'a';
        while(!p->next[k]&&p!=rt)   //找到同字母的父结点fail
            p=p->fail;
        p=p->next[k];
        if(!p)  p=rt;       //如果父结点下没有该结点就指回根

        node *tmp=p;        //对应结点
        while(tmp!=rt&&tmp->cnt!=-1)    //把所有该字母的fail都遍历一次
        {                               //放这里就相当于把每个以这个字母为后缀的串都查询了
            ans+=tmp->cnt;
            tmp->cnt=-1;        //防止模式串重复出现时的重复计数
            tmp=tmp->fail;
        }
        i++;
    }
    return ans;
}
char word[52],s1[1000005];
int main()
{
//    RE
    int ncase,n;
    scanf("%d",&ncase);
    while(ncase--){
        node *rt = new node();
        scanf("%d%*c",&n);
        for(int i=0;i<n;i++){
            gets(word);
            insert_tree(rt,word);
        }
        gets(s1);
        buildFail(rt);
        printf("%d\n",search_tree(rt,s1));
    }
    return 0;
}

字典树解法：字典树查询必须是前缀出现，如样例的ya根本没出现，所以不能直接查询整个串，要对主串的每一个子串都查询一次树，当某个前缀子串不行的时候，以它为前缀的子串也不用找了，所以复杂度没有LEN的阶乘那么恐怖，不过也快TLE了..

936MS

29856K

#include <cstdio>
#include <cmath>
#include <cstring>
#include <string>
#include <set>
#include <map>
#include <stack>
#include <queue>
#include <vector>
#include <iostream>
#include <algorithm>
using namespace std;
#define ll long long
#define eps 10^(-6)
#define Q_CIN ios::sync_with_stdio(false);
#define REP( i , n ) for ( int i = 0 ; i < n ; ++ i )
#define FOR( i , a , b ) for ( int i = a ; i <= b ; ++ i )
#define CLR( a , x ) memset ( a , x , sizeof (a) );
#define RE freopen("1.in","r",stdin);
#define WE freopen("1.out","w",stdout);
#define MOD 10009
#define debug(x) cout<<#x<<":"<<(x)<<endl;
const int NODE_MAX=27;

struct node
{
    node *next[NODE_MAX];
    int cnt;
    node(){
        CLR(next,NULL);
        cnt=0;
    }
};

void insert_tree(node *p,char s1[])
{
    int i=0;
    while(s1[i])
    {
        int k=s1[i]-'a';
        if(!p->next[k])
            p->next[k]=new node();
        p=p->next[k];
        i++;
    }
    p->cnt++;
}

int search_tree(node *rt,char s1[])
{
    int i=0,ans=0;
    node *p=rt;
    while(s1[i])
    {
        node *p=rt;
        for(int j=i;s1[j];j++)
        {
            int k=s1[j]-'a';
            if(p->next[k])
            {
                p=p->next[k];
                ans+=p->cnt;
                p->cnt=0; //防止模式串重复的情况下多计数，如"he"有2个那只能算1次
            }
            else    break;  //该子串没有
        }
        i++;
    }
    return ans;
}
char word[52],s1[1000005];
int main()
{
    RE
    int ncase,n;
    scanf("%d",&ncase);
    while(ncase--){
        node *rt = new node();
        scanf("%d%*c",&n);
        for(int i=0;i<n;i++){
            gets(word);
            insert_tree(rt,word);
        }
        gets(s1);
        printf("%d\n",search_tree(rt,s1));
    }
    return 0;
}