文本压缩

程序有问题,暂时做个记录,以后会修改

#include <stdio.h>
#include <stdlib.h>
#include <assert.h>
#include <string.h>

const int SIZE = 256;

void tong_ji_ping_du(FILE *f, int ping_du[]) {
    int c;
    while ((c=fgetc(f)) != EOF) {
        ++ping_du[c];
    }
}

struct Huffman_Node {
    int key;
    Huffman_Node *parent, *lChild, *rChild;
};

void Select_Node(Huffman_Node *tree, Huffman_Node **s1, Huffman_Node **s2, int n) {
    *s1 = *s2 = NULL;
    for (int i=0; i<n; ++i) {
        if (tree[i].key!=0 && tree[i].parent==NULL && (*s1==NULL || tree[i].key< (*s1)->key)) {
            *s2 = *s1;
            *s1 = &tree[i];
        }
        else if (tree[i].key!=0 && tree[i].parent==NULL && (*s2==NULL || tree[i].key < (*s2)->key)) {
            *s2 = &tree[i];
        }
    }
}

void Huffman_code(int ping_du[], char codes[][SIZE+1]) {
    Huffman_Node *tree;
    Huffman_Node *s1;
    Huffman_Node *s2;
    int m = 2 * SIZE - 1;
    tree = new Huffman_Node[m];
    memset(tree, 0, sizeof(Huffman_Node) * m);
    m = SIZE;
    for (int i=0; i<SIZE; ++i) {
        if (ping_du[i] != 0) ++m;

        tree[i].key = ping_du[i];
    }

    for (int i=SIZE; i<m-1; ++i) {
        Select_Node(tree, &s1, &s2, i);
        assert(s1!=NULL && s2!=NULL);
        tree[i].key = s1->key + s2->key;
        s1->parent = &tree[i];
        s2->parent = &tree[i];
        tree[i].lChild = s1;
        tree[i].rChild = s2;
    }

    char code[SIZE+1];
    int start;
    Huffman_Node *node;
    for (int i=0; i<SIZE; ++i) {
        if (ping_du[i]==0) continue;

        memset(code, 0, SIZE+1);
        start = SIZE;
        node = &tree[i];
        while (node->parent!=NULL) {
            if (node->parent->lChild==node) code[--start] = '0';
            else code[--start] = '1';
            node = node->parent;
        }
        strcpy(&codes[i][0], &code[start]);
    }
}

void ya_suo(FILE *f, FILE *f2, int ping_du[], char codes[][SIZE+1]) {
    size_t f_size = 0;

    // 计算文件大小
    for (int i=0; i<SIZE; ++i) {
        f_size += ping_du[i];
    }

    // 写入HEAD信息
    fprintf(f2, "HEAD BEGIN\n");
    fprintf(f2, "file_size:%lu\n", f_size);
    for (int i=0; i<SIZE; ++i) {
        if (ping_du[i] == 0) continue;
        fprintf(f2, "%d:%s\n", i, codes[i]);
    }
    fprintf(f2, "HEAD END\n");

    int c;
    char cByte;
    int byte_size = 0;
    while ((c=fgetc(f)) != EOF) {
        for (int i=0; i<strlen(&codes[c][0]); ++i) {
            ++byte_size;
            cByte >>= 1;
            cByte &= 0x7F; // 最高位设置为0
            if (codes[c][i] == '1') cByte ^= 0x80;
            if (byte_size == 8) {
                byte_size = 0;
                fputc(cByte, f2);
            }
        }
    }
    for (int i=byte_size; i<8; ++i) {
        cByte >>= 1;
    }
    fputc(cByte, f2);
}

void jie_ya(FILE *f, FILE *f2) {
    char codes[SIZE][SIZE+1] = {0};
    char buf[BUFSIZ];
    int ch;
    char code[SIZE+1] = {0};
    size_t f_size;

    fgets(buf, BUFSIZ, f2);
    buf[strlen(buf)-1] = '\0';
    if (strcmp(buf, "HEAD BEGIN") == 0) {
        fgets(buf, BUFSIZ, f2);
        buf[strlen(buf)-1] = '\0';
        sscanf(buf, "file_size:%ld", &f_size);
        fgets(buf, BUFSIZ, f2);
        buf[strlen(buf)-1] = '\0';
        while (strcmp(buf, "HEAD END")!=0) {
            sscanf(buf, "%d:%s", &ch, code);
            strcpy(&codes[ch][0], code);

            fgets(buf, BUFSIZ, f2);
            buf[strlen(buf)-1] = '\0';
        }
    }

    memset(code, 0, SIZE+1);
    int i_ch = 0;
    while (f_size>0 && (ch=fgetc(f2))!=EOF) {
        for (int i=0; i<8; ++i) {
            code[i_ch++] = (ch & 0x1) + '0';
            ch >>= 1;
            for (int j=0; j<SIZE; ++j) {
                if (strcmp(&codes[j][0], code) == 0) {
                    if (f_size>0)
                        fputc(j, f);
                    --f_size;
                    i_ch = 0;
                    memset(code, 0, SIZE+1);
                    break;
                }
            }
        }
    }
}

int main(int argc, char *argv[]) {
    assert(argc==4);
    int cmd = atoi(argv[1]);
    char *f_name = argv[2];
    char *f_name2 = argv[3];

    if (cmd==1) {
        FILE *f = fopen(f_name, "rb");
        FILE *f2 = fopen(f_name2, "wb");

        int ping_du[SIZE]={0};
        char codes[SIZE][SIZE+1]={0};

        // 统计字节的频率
        tong_ji_ping_du(f, ping_du);

        // 生成huffman编码
        Huffman_code(ping_du, codes);

        rewind(f);
        ya_suo(f, f2, ping_du, codes);

        fclose(f);
        fclose(f2);
    }
    else if (cmd==2) {
        FILE *f = fopen(f_name2, "wb");
        FILE *f2 = fopen(f_name, "rb");

        jie_ya(f, f2);

        fclose(f);
        fclose(f2);
    }

    return 0;
}
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值