Huffman Coding 哈夫曼编码

作者：jostree 转载请注明出处 http://www.cnblogs.com/jostree/p/4096079.html

使用优先队列实现，需要注意以下几点：

1.在使用priority_queue时，内部需要存储哈夫曼树节点的指针，而不能是节点。因为构建哈夫曼树时，需要把其左右指针指向孩子，而如果储存的是节点，那么孩子的地址是会改变的。同理节点应当使用new在内存中开辟，而不能使用vector，原因是vector在数组大小为2整数次幂时，大小会倍增，开辟新数组并把老数组的数字copy过去，从而也会导致地址变化。

2.优先队列对指针的排列，需要额外写一个比较函数来比较指针指向的节点的大小。bool operator () (wcnode * node1, wcnode * node2) return node1->lessthan(node2);并在定义优先队列时使用这种方法：    priority_queue <wcnode*, vector<wcnode*>, compare> 第一个参数是节点类型，第二个参数是优先队列的储存结构，第三个参数是比较函数。

3.C++在写入文件时，由于只能按字节写入，因此需要把8个bit位转化为一个字节，最后不足8位用0补齐，并记录文件总bit数，便于解码。然后写入文件。另写入二进制文件可以使用ofstream out(“output.txt”,std::ofstream::binary);

4.哈夫曼编码信息包括每种字符的映射，和该文件的总bit数。

其代码如下：

 #include <cstdio>
 #include <cstdlib>
 #include <iostream>
 #include <cstring>
 #include  <fstream>
 #include  <queue>
 #include  <map>
 #include  <vector>
 using namespace std;
 class compare; class wcnode
 {
     public:
         friend class compare;
         char word;
         int count;
         wcnode* left;
         wcnode* right;
         bool lessthan (const wcnode *w)const
         {
             return count > w->count;
         }
         wcnode(char w='\0', int c=, wcnode* l=NULL, wcnode * r=NULL)
         {
             word = w; count = c; left = l; right = r;
         }
 }; class compare
 {
     public:
         bool operator () (wcnode * node1, wcnode * node2)
         {
             return node1->lessthan(node2);
         }
 }; void preorder(wcnode *head, vector<bool> rec, map<char, vector<bool> > & res)
 {
     if( head->left == NULL && head->right == NULL )
     {
         res[head->word] = rec;
         return;
     }
     vector<bool> l = rec;
     l.push_back();
     vector<bool> r = rec;
     r.push_back();
     if(head->left != NULL) preorder(head->left, l, res);
     if(head->right != NULL) preorder(head->right, r, res);
 }
 map<char, vector<bool> > encode(map<char, int> &wordcount)
 {
     map<char, vector<bool> > res;
     priority_queue <wcnode*, vector<wcnode*>, compare> pq;
     map<char, int>::iterator t;
     wcnode *tmp;
     wcnode *t1, *t2, *t3;     for( t = wordcount.begin() ; t != wordcount.end() ; t++ )
     {
         tmp = new wcnode();
         tmp->word = t->first;
         tmp->count = t->second;
         pq.push(tmp);
     }
     while( pq.size() >  )
     {
         t1 = pq.top();
         pq.pop();
         t2 = pq.top();
         pq.pop();
         t3 = new wcnode();
         t3->count = t1->count + t2->count;
         t3->left = t1;
         t3->right = t2;
         pq.push(t3);
     }
     wcnode *huffmanhead = pq.top();
     vector<bool> rec;
     preorder(huffmanhead, rec, res);
     map<char, vector<bool>  >::iterator it;
     for( it = res.begin() ; it != res.end() ; it++ )
     {
         cout<<it->first<<":";
         for( int i = ; i < it->second.size() ; i++ )
         {
             cout<<it->second[i];
         }
         cout<<", ";
     }
     return res;
 } void output(string s, string passage, map<char, vector<bool> > res)
 {
     ofstream out(s.c_str());
     vector<bool> bit;
     for( int i =  ; i < passage.size() ; i++ )
     {
         vector<bool> tmp = res[passage[i]];
         for( int i =  ; i < tmp.size(); i++ )
         {
             bit.push_back(tmp[i]);
         }
     }
     char outputchar = ;
     for( int i =  ; i < bit.size() ; i++ )
     {
         if( i %  ==  )
         {
            out.write(&outputchar, sizeof(outputchar));
            outputchar = ;
         }
         outputchar = outputchar + bit[i];
         outputchar = outputchar * ;
     }
     if( outputchar !=  )
     {
         out.write(&outputchar, sizeof(outputchar));
     }
     out.close();
 }
 int main(int argc, char *argv[])
 {
     char tmp;
     ifstream in("Aesop_Fables.txt");
     map <char, int> wordcount;
     map <char, vector<bool> > res;
     string passage;
     while( in.get(tmp) )
     {
         passage += tmp;
         if( wordcount.count(tmp) ==   )
         {
             wordcount[tmp] = ;
         }
         else
         {
             wordcount[tmp]++;
         }
     }
     res = encode(wordcount);
     output("outAesop.txt", passage, res);
     in.close();
 }