External/common/CurlFtp/CurlFtp.cpp

#include "CurlFtp.h"
#include <iostream>
#include <regex>
#include <iosfwd>

#include "fmtlog.h"
#include "CurlFtpInfo.h"


#if defined(_WIN32)

#endif /* _WIN32 */

/* ==================================================================================
 * *********************************** 全局变量 *************************************
 * ==================================================================================
 */

/* 是否有初始化实例，主要用于静态函数调用curl_global_cleanup()函数前判断，有实例就不调用 */
static bool hasInstace = false;

/* ==================================================================================
 * *********************************** 全局函数 *************************************
 * ==================================================================================
 */

/**
 * @brief 写入回调函数
 * 
 * @param contents curl的数据缓冲区
 * @param size 数据大小，单位是size_t
 * @param nmemb size_t的单位字节数
 * @param userStr 用户提供的字符串，格式可以是任意的
 * @return size_t 总共获取到的数据大小，单位是字节
 */
static size_t writeStringCallback(void *contents, size_t size, size_t nmemb, std::string *userStr) 
{
    size_t newLength = size * nmemb;
    size_t oldLength = userStr->size();
    try
    {
        userStr->resize(oldLength + newLength);
    }
    catch(std::bad_alloc &e)
    {
        //handle memory problem
        return 0;
    }
    std::copy_n((char*)contents, newLength, userStr->begin() + oldLength);
    return size * nmemb;
}


/**
 * @brief 写入回调函数，listFiles需要调用
 * 
 * @param buffer curl下载回来的数据
 * @param size 数据大小
 * @param nmemb 数据单位字节数
 * @param userp 用户传进来的容器
 * @return int 返回拷贝的字节数
 */
static int writeStringListCallback(void* buffer, size_t size, size_t nmemb, void* userp)
{
    std::vector<std::string>* fileList = static_cast<std::vector<std::string>*>(userp);
    std::string line(static_cast<char*>(buffer), size * nmemb);
    // printf("line = %s\n", line.c_str());
    fileList->push_back(line);
    return size * nmemb;
}

/* 使用Windows API进行编码转换 */
#if defined(_WIN32)
static char* GBToUTF8(const char* gb2312)
{
    int len = MultiByteToWideChar(CP_ACP, 0, gb2312, -1, NULL, 0);
    wchar_t* wstr = new wchar_t[len+1];
    memset(wstr, 0, len+1);
    MultiByteToWideChar(CP_ACP, 0, gb2312, -1, wstr, len);
    len = WideCharToMultiByte(CP_UTF8, 0, wstr, -1, NULL, 0, NULL, NULL);
    char* str = new char[len+1];
    memset(str, 0, len+1);
    WideCharToMultiByte(CP_UTF8, 0, wstr, -1, str, len, NULL, NULL);
    if(wstr) delete[] wstr;
    return str;
}
#endif /* _WIN32 */

/**
 * @brief 解析字符串文件信息
 * 
 * @param strSrc 
 * @param fileList 
 * @return true 
 * @return false 
 */
static bool parseFileInfo(const std::string& strSrc, std::vector<CF_FileInfo>& fileInfo)
{
#if defined(_WIN32)
    auto str1 = GBToUTF8(strSrc.c_str());
    std::string str2 = str1;
#else
    std::string str2 = strSrc;
#endif /* _WIN32 */
    
    // FMTLOG_DEBUG("\n{}", str2);
    /* 正则表达式，匹配多个空格 */
    std::regex reg(R"(( )+)");
    /* 匹配以非空格开头的字符，空格隔开，中间任意字符，空格隔开，非空格组成的结尾
     * (文件类型) ... (文件大小,$5) ... (文件名)
     */
    std::regex reg1(R"(^([^ ]+) (\d*) (\w*) (\w*) (\d*) (.*) ([^ ]+)$)");
    // FMTLOG_INFO("\n-------------------\n");
    /* 匹配换行符，分割每一行 */
    std::regex reg2(R"(\n)");
    /* 匹配文件夹 */
    std::regex reg3(R"(^d.*)");
    /* 匹配文件 */
    std::regex reg4(R"(^-.*$)");

    /* -1 表示对正则表达式之前的子序列感兴趣
     * 0 表示对正则表达式本身感兴趣，输出的就是换行符 */
    std::sregex_token_iterator it2(str2.begin(), str2.end(), reg2, -1);
    /* 这里取出每一行 */
    for(; it2 != std::sregex_token_iterator(); ++it2)
    {
        /* 去掉多余的空格 */
        auto line = std::regex_replace(it2->str(), reg, " ");
        // FMTLOG_INFO("{}", line);
        CF_FileInfo fi;
        /* 取出文件类型 */
        std::string strFileType = std::regex_replace(line, reg1, "$1");
        if(std::regex_match(strFileType, reg3))
        {
            fi.type = CF_FileType::DIR;
        }else if (std::regex_match(strFileType, reg4))
        {
            fi.type = CF_FileType::FILE;
        }
        /* 取出文件大小 */
        std::string strFileSize = std::regex_replace(line, reg1, "$5");
        fi.size = std::stoull(strFileSize);
        /* 取出文件名 */
        std::string strFileName = std::regex_replace(line, reg1, "$7");
        fi.name = strFileName;
        /* 加入队列 */
        fileInfo.push_back(fi);
    }

    return true;
}


/* ==================================================================================
 * *********************************** 成员函数 *************************************
 * ================================================================================== */


CurlFtp::CurlFtp()
{
    /* 调用初始化函数，这个函数可以重复调用 */
    curl_global_init(CURL_GLOBAL_DEFAULT);
    /* 初始化curl */
    m_curl = curl_easy_init();
    hasInstace = true;
}


CurlFtp::~CurlFtp()
{
    /* 清理curl */
    curl_easy_cleanup(m_curl);
    curl_global_cleanup();
    hasInstace = false;
}

/* 检查文件夹是否存在，不确定好不好用 */
// bool CurlFtp::isDirExists(const std::string &ftpUrl, const std::string &username, const std::string &password)
// {
//     CURL *curl;
//     CURLcode res;
//     bool result = false;
//     std::string retList;

//     curl_global_init(CURL_GLOBAL_DEFAULT);
//     curl = curl_easy_init();
//     if(curl) 
//     {
//         curl_easy_setopt(curl, CURLOPT_URL, ftpUrl.c_str());
//         curl_easy_setopt(curl, CURLOPT_USERNAME, username.c_str());
//         curl_easy_setopt(curl, CURLOPT_PASSWORD, password.c_str());
//         // curl_easy_setopt(curl, CURLOPT_DIRLISTONLY, 1L); // We only want the directory listing

        
//         curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writeCallback);
//         curl_easy_setopt(curl, CURLOPT_WRITEDATA, &retList);

//         res = curl_easy_perform(curl);
//         // printf("res = %d\n", res);
//         if(res != CURLE_OK) 
//         {
//             fprintf(stderr, "getDirList() failed, error code %d, :%s\n",res, curl_easy_strerror(res));
//         }
//         else 
//         {
//             // std::cout << "Directory list: \n" << retList << std::endl;
//         }
//         curl_easy_cleanup(curl);
//     }
    
//     curl_global_cleanup();
    
//     return result;
// }

/* 使用正则表达式提取出文件夹 */
// bool CurlFtp::extractDirectories(const std::string& responseString)
// {
//     std::regex directoryRegex(R"(^d.*\s(\S+)$)");
//     std::smatch match;

//     std::istringstream retStream(responseString);
//     std::string line;

//     while (std::getline(retStream, line)) 
//     {
//         if (std::regex_match(line, match, directoryRegex)) 
//         {
//             std::cout << "Directory: " << match[1] << std::endl;
//         }
//     }
//     return true;
// }

/**
 * @brief 列出FTP文件夹
 * 
 * @param ftpUrl 需要列出文件夹的FTP地址
 * @param username 
 * @param password 
 * @return std::string 
 */
std::string CurlFtp::listDir(const std::string &ftpUrl, const std::string &username, const std::string &password)
{
    CURL *curl;
    CURLcode res;
    bool result = false;
    std::string retList;
    /* 1. 初始化curl，这个函数需要在调用任何curl函数之前 */
    curl_global_init(CURL_GLOBAL_DEFAULT);
    /* 2. 获取一个curl句柄 */
    curl = curl_easy_init();
    if(curl) 
    {
        /* 3. 设置curl选项 */
        curl_easy_setopt(curl, CURLOPT_URL, ftpUrl.c_str());
        curl_easy_setopt(curl, CURLOPT_USERNAME, username.c_str());
        curl_easy_setopt(curl, CURLOPT_PASSWORD, password.c_str());
        // curl_easy_setopt(curl, CURLOPT_DIRLISTONLY, 1L); // We only want the directory listing

        curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writeStringCallback);
        curl_easy_setopt(curl, CURLOPT_WRITEDATA, &retList);

        /* 4. 发送信息，一直等待知道返回 */
        res = curl_easy_perform(curl);
        // printf("res = %d\n", res);
        if(res != CURLE_OK) 
        {
            fprintf(stderr, "getDirList() failed, error code %d, :%s\n",res, curl_easy_strerror(res));
        } else 
        {
            // std::cout << "Directory list: \n" << retList << std::endl;
        }
        /* 5. 清理curl */
        curl_easy_cleanup(curl);
    }
    if(hasInstace == false)
    {
        curl_global_cleanup();
    }
    return retList;
}





/* 列出文件夹中的所有文件 */
bool CurlFtp::listFiles(const std::string &ftpUrl, const std::string &username, const std::string &password, std::vector<std::string>& fileList)
{
    CURL *curl;
    CURLcode res;
    bool result = false;

    curl_global_init(CURL_GLOBAL_DEFAULT);
    curl = curl_easy_init();
    if(curl)
    {
        curl_easy_setopt(curl, CURLOPT_URL, (ftpUrl).c_str());
        curl_easy_setopt(curl, CURLOPT_USERNAME, username.c_str());
        curl_easy_setopt(curl, CURLOPT_PASSWORD, password.c_str());
        curl_easy_setopt(curl, CURLOPT_DIRLISTONLY, 1L);
        curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, writeStringListCallback);
        curl_easy_setopt(curl, CURLOPT_WRITEDATA, &fileList);

        res = curl_easy_perform(curl);
        if(res != CURLE_OK) 
        {
            fprintf(stderr, "Failed to get file list, error code :%d ,%s\n", res, curl_easy_strerror(res));
        }
        else
        {
            // for(const std::string& filename : fileList)
            // {
            //     printf("%s\n", filename.c_str());
            // }
            result = true;
        }

        curl_easy_cleanup(curl);
    }

    if(hasInstace == false)
    {
        curl_global_cleanup();
    }
    return result;
}



/* 创建文件夹 */
bool CurlFtp::createDir(const std::string &ftpUrl, const std::string &username, const std::string &password, const std::string &dirName)
{
    CURL *curl;
    CURLcode res;
    bool result = false;

    curl_global_init(CURL_GLOBAL_DEFAULT);
    curl = curl_easy_init();
    if(curl)
    {
        curl_easy_setopt(curl, CURLOPT_URL, ftpUrl.c_str());
        curl_easy_setopt(curl, CURLOPT_USERNAME, username.c_str());
        curl_easy_setopt(curl, CURLOPT_PASSWORD, password.c_str());

        // Create a list of FTP commands to be executed on the server
        struct curl_slist *headerlist = NULL;
        std::string mkdir = "MKD " + dirName;
        headerlist = curl_slist_append(headerlist, mkdir.c_str());

        // Set the list of FTP commands to be executed on the server before the transfer
        curl_easy_setopt(curl, CURLOPT_QUOTE, headerlist);
        // Tell libcurl to not include the body in the output
        curl_easy_setopt(curl, CURLOPT_NOBODY, 1L);

        res = curl_easy_perform(curl);

        if(res != CURLE_OK) 
        {
            fprintf(stderr, "createDir() failed, error code :%d ,%s\n", res, curl_easy_strerror(res));
            result = false;
        }else
        {
            result = true;
        }

        curl_easy_cleanup(curl);
        curl_slist_free_all(headerlist);
    }

    if(hasInstace == false)
    {
        curl_global_cleanup();
    }
    return result;
}


/* 设置IP和端口 */
bool CurlFtp::setFtpIPAndPort(const std::string& IP, const std::string& port)
{
    /*先判断是否有ftp器前缀，通过正则表达式，取出IP地址和端口号，去掉前缀和后面的'/ */
    m_IP = IP;
    m_port = port;
    m_ftpUrl = "ftp://" + m_IP + ":" + m_port;
    printf("ftpUrl = %s\n", m_ftpUrl.c_str());

    return true;
}


/* 设置用户名和密码 */
bool CurlFtp::setFtpUsernameAndPassword(const std::string& username, const std::string& password)
{
    m_username = username;
    m_password = password;

    return true;
}


/* 列出文件列表 */
bool CurlFtp::getFileList(std::string dir, std::vector<std::string>& fileList)
{
    if(m_IP.empty() || m_port.empty())
    {
        FMTLOG_WARN("IP or port is empty");
        return false;
    }
    if(m_curl == nullptr)
    {
        FMTLOG_WARN("m_curl is nullptr");
        return false;
    }
    std::vector<CF_FileInfo> listInfo;
    listAll(dir, listInfo);
    // for(const CF_FileInfo& fi : listInfo)
    // {
    //     FMTLOG_INFO("type = {}, size = {}, name = {}", (int)fi.type, fi.size, fi.name);
    // }
    
    return true;
}



/* 列出文件列表 */
bool CurlFtp::listAll(std::string dir, std::vector<CF_FileInfo>& fileList)
{
    if(m_IP.empty() || m_port.empty())
    {
        printf("IP or port is empty\n");
        return false;
    }
    if(m_curl == nullptr)
    {
        printf("m_curl is nullptr\n");
        return false;
    }
    bool result = false;
    std::string strSrc;
    /* 先设置FTP地址 */
    std::string ftpUrl = m_ftpUrl + dir;
    curl_easy_setopt(m_curl, CURLOPT_URL, ftpUrl.c_str());
    /* 设置用户名和密码 */
    curl_easy_setopt(m_curl, CURLOPT_USERNAME, m_username.c_str());
    curl_easy_setopt(m_curl, CURLOPT_PASSWORD, m_password.c_str());

    /* 设置列出文件命令，只列出文件名称，不携带信息 */
    // curl_easy_setopt(m_curl, CURLOPT_DIRLISTONLY, 1L);
    /* 设置回调函数 */
    curl_easy_setopt(m_curl, CURLOPT_WRITEFUNCTION, writeStringCallback);
    /* 设置需要写入的容器，回调函数的第四个参数 */
    curl_easy_setopt(m_curl, CURLOPT_WRITEDATA, &strSrc);

    /* 发送请求 */
    m_res = curl_easy_perform(m_curl);
    if(m_res != CURLE_OK) 
    {
        FMTLOG_ERROR("Failed to get file list, error code: {} ,{}", (int)m_res, curl_easy_strerror(m_res));
        return false;
    }
    /* 解析字符串 */
    parseFileInfo(strSrc, fileList);

    return true;
}