博客
关于我
2020编码大赛(4)LZW压缩算法
阅读量:607 次
发布时间:2019-03-12

本文共 3601 字,大约阅读时间需要 12 分钟。

LZW的原理:

初始化字典为255个字符,随着压缩的过程,字典一步步扩充,最终把全文本转换成字典的id序列,再用变长编码的方法,把id序列编码成二进制。

编码方法:0-254用8位表示,255到510用9位表示,......

当第一次需要编码9位的时候,插入255表示升位,此后0-510都用9位表示,......

因为代码是用string来操作的,所以额外加了对于/0的编码,保证算法的通用性。

 

压缩率:大概在1024:420左右

代码:

#include "compress.h"#include #define BUF_SIZE 1048576#define ASCII 256#define ASCIILEN 8using namespace std;char g_buff[BUF_SIZE + 5];char g_curByte;int g_bufSize;int g_bits;int g_byte;bool g_EOF;void writeBit(unsigned int data, int length){    if (!length) {        return;    }    if (!g_bits) {        g_buff[g_byte++] = g_curByte, g_curByte = 0;    }    g_bits = (g_bits + 7) % 8, g_curByte |= (((data >> length - 1) & 1) << g_bits);    writeBit(data, length - 1);}void writeInt(unsigned int data, int &length){    if (data >= unsigned((1 << length) - 1)) {        writeBit((1 << length) - 1, length);        length++;    }    writeBit(data, length);}void readBit(unsigned int &val){    if (!g_bits) {        g_EOF = (g_byte == g_bufSize), g_curByte = g_buff[g_byte++];    }    g_bits = (g_bits + 7) % 8, val = val * 2 + ((g_curByte >> g_bits) & 1);}unsigned int readInt(int length){    unsigned int val = 0;    while (length-- && !g_EOF) {        readBit(val);    }    return val;}void compress(ifstream &fin, ofstream &fout){    map
m; for (int i = 0; i < ASCII; i++) { m[string("") + char(i)] = i; } int bitLen = ASCIILEN; unsigned int index = ASCII; char ch; string str; while (fin.get(ch)) { if (!ch) { if (m[str]) { writeInt(m[str], bitLen); str = ""; } writeInt(0, bitLen); continue; } if (m[str + ch]) { str += ch; } else { writeInt(m[str], bitLen); m[str + ch] = index++; str = ch; } } if (str != "") { writeInt(m[str], bitLen); }}void decompress(ofstream &fout, ifstream &fin){ map
m; for (int i = 0; i < ASCII; i++) { m[i] = string("") + char(i); } int bitLen = ASCIILEN; unsigned int index = ASCII; char zero = 0; string oldstr, newstr; g_curByte = g_buff[0]; int code = g_curByte; if (code == ASCII - 1) { // 以\255开头的文件 bitLen++; code = readInt(bitLen); } if (code == 0) { fout << zero; } else { oldstr = m[code]; fout << oldstr; } while (true) { code = readInt(bitLen); if (g_EOF) { break; } if (code == (1 << bitLen) - 1) { bitLen++; code = readInt(bitLen); } if (code == 0) { fout << zero; oldstr = ""; continue; } if (m[code] != "") { newstr = m[code]; } else { newstr = oldstr + oldstr[0]; } if (oldstr != "") { m[index++] = oldstr + newstr[0]; } oldstr = newstr; fout << oldstr; }}void compress(const string input, const string output){ ifstream fin(input.c_str(), ios::binary); ofstream fout(output.c_str(), ios::binary); g_bits = 8; g_byte = 0; g_curByte = 0; compress(fin, fout); fin.close(); g_buff[g_byte++] = g_curByte; fout.write(g_buff, sizeof(char) * g_byte); fout.close();}void decompress(const string input, const string output){ ifstream fin(input.c_str(), ios::binary); ofstream fout(output.c_str(), ios::binary); g_EOF = false; g_bits = 0; g_byte = 1; fin.read(g_buff, sizeof(char) * BUF_SIZE); g_bufSize = int(fin.gcount()); decompress(fout, fin); fout.close(); fin.close();}

 

转载地址:http://lloxz.baihongyu.com/

你可能感兴趣的文章
mac mysql 进程_Mac平台下启动MySQL到完全终止MySQL----终端八步走
查看>>
Mac OS 12.0.1 如何安装柯美287打印机驱动,刷卡打印
查看>>
MangoDB4.0版本的安装与配置
查看>>
Manjaro 24.1 “Xahea” 发布!具有 KDE Plasma 6.1.5、GNOME 46 和最新的内核增强功能
查看>>
mapping文件目录生成修改
查看>>
MapReduce程序依赖的jar包
查看>>
mariadb multi-source replication(mariadb多主复制)
查看>>
MariaDB的简单使用
查看>>
MaterialForm对tab页进行隐藏
查看>>
Member var and Static var.
查看>>
memcached高速缓存学习笔记001---memcached介绍和安装以及基本使用
查看>>
memcached高速缓存学习笔记003---利用JAVA程序操作memcached crud操作
查看>>
Memcached:Node.js 高性能缓存解决方案
查看>>
memcache、redis原理对比
查看>>
memset初始化高维数组为-1/0
查看>>
Metasploit CGI网关接口渗透测试实战
查看>>
Metasploit Web服务器渗透测试实战
查看>>
MFC模态对话框和非模态对话框
查看>>
Moment.js常见用法总结
查看>>
MongoDB出现Error parsing command line: unrecognised option ‘--fork‘ 的解决方法
查看>>