C++（SOCKET）简单爬虫

  1 #include <iostream>
  2 #include <stdio.h>
  3 #include <string>
  4 #include <cstdlib>
  5 #include <fstream>
  6 #include <WinSock2.h>
  7 
  8 using namespace std;
  9 
 10 #pragma warning(disable:4996)
 11 //忽略VS特有警告
 12 #pragma comment(lib, "ws2_32.lib")
 13 //加载ws2_32.dll
 14 #define BUFF_SIZE 1024
 15 
 16 int ncount = 0;
 17 string host, pos;
 18 
 19 SOCKET ConnectFunc(string host, string pos) {
 20     WSADATA wsaData;
 21     SOCKET serv;
 22     //创建套接字
 23     if (WSAStartup(MAKEWORD(2, 2), &wsaData) != 0) {//初始化DLL
 24         cout << "WSAStartup() Failed:" << WSAGetLastError() << endl;
 25         system("PAUSE");
 26         return -1;
 27     }
 28 
 29     serv = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
 30     //初始化套接字
 31     if (serv == INVALID_SOCKET) {
 32         cout << "socket() Failed:" << WSAGetLastError() << endl;
 33         system("PAUSE");
 34         return -1;
 35     }
 36 
 37     struct hostent *pt = gethostbyname(host.c_str());//解析域名IP
 38     if (!pt) {
 39         cerr << "Get IP Error!!!" << endl;
 40         system("PAUSE");
 41         return -1;
 42     }
 43     struct sockaddr_in serv_addr;
 44     //创建结构体sockaddr_in结构体变量，绑定套接字
 45     memcpy(&serv_addr.sin_addr, pt->h_addr, 4);
 46     //自动响应服务器IP
 47     serv_addr.sin_family = AF_INET;
 48     //IPv4
 49     serv_addr.sin_port = htons(80);
 50     //端口
 51 
 52     /*输出服务器IP
 53     for (int i = 0; pt->h_addr_list[i]; i++) {
 54         printf("IP addr %d: %s\n", i + 1, inet_ntoa(*(struct in_addr*)pt->h_addr_list[i]));
 55     }
 56     */
 57 
 58     if (connect(serv, (LPSOCKADDR)&serv_addr, sizeof(SOCKADDR)) == SOCKET_ERROR) {//连接服务器
 59         cout << "connect() Failed:" << WSAGetLastError() << endl;
 60         system("PAUSE");
 61         return -1;
 62     }//与服务器建立连接
 63 
 64     string request = "GET " + pos + " HTTP/1.1\r\nHost:" + host + "\r\nConnection:Close\r\n\r\n";
 65     //向服务器请求图片资源(发送到服务器的命令)
 66     if (send(serv, request.c_str(), request.size(), 0) == SOCKET_ERROR) {
 67         cout << "send() Failed:" << WSAGetLastError() << endl;
 68         closesocket(serv);
 69         system("PAUSE");
 70         return -1;
 71     }//发送指令消息
 72 
 73     return serv;
 74     //返回套接字
 75 }
 76 
 77 void DownloadPicture() {
 78     SOCKET serv_in = ConnectFunc(host, pos);
 79     //连接服务器
 80     
 81     char buffer[BUFF_SIZE] = { 0 };
 82     //数据缓存文件
 83 
 84     string a = "G:\\Pictures\\", name;
 85     cout << "picture name:";
 86     cin >> name;
 87     a = a + name + ".png";
 88     //文件命名
 89 
 90     FILE *fp = fopen(a.c_str(), "wb+");
 91     //创建文件
 92 
 93     if (NULL == fp) {
 94         cerr << "Open File" << endl;
 95         system("PAUSE");
 96         exit(-1);
 97     }
 98 
 99     ncount = recv(serv_in, buffer, BUFF_SIZE, 0);
100     //跳过不需要信息(状态行和消息报头）
101 
102     char *infor = strstr(buffer, "\r\n\r\n");
103     //区分条件
104     fwrite(infor + strlen("\r\n\r\n"), sizeof(char), ncount - (infor - buffer) - strlen("\r\n\r\n"), fp);
105     //丢弃不需要数据
106     for (; (ncount = recv(serv_in, buffer, BUFF_SIZE, 0)) > 0;) {
107         fwrite(buffer, sizeof(char), BUFF_SIZE, fp);
108         Sleep(2);
109     }//循环写入数据
110 
111     fclose(fp);
112     //关闭文件流指针
113     closesocket(serv_in);
114     //断开连接，清除套接字
115 }
116 
117 int main()
118 {
119     CreateDirectory(L"G:\Pictures", NULL);
120     //创建文件夹
121     
122     host = "images.cnblogs.com";
123     //cin >> host;
124     //输入要爬的网站地址
125     pos = "/cnblogs_com/Mayfly-nymph/1233628/o_images.png";
126     //cin >> pos;
127     //图片在服务器中的位置
128     DownloadPicture();
129     //下载图片
130     system("PAUSE");
131     return 0;
132 }