获取ftp目录及子目录下的文件并下载下来

/*

利用AI写了一个在fedora系统下的c++程序，
实现功能为，先获取ftp指定目录及子目录下的文件名称，然后把本地目录中没有的文件下载下来。

# 安装依赖
sudo dnf install libcurl-devel

# 编译
g++ -o ftp_sync_curl ftp_sync_curl.cpp -lcurl -lstdc++fs -O2

*/

#include <iostream>
#include <string>
#include <vector>
#include <set>
#include <fstream>
#include <filesystem>
#include <curl/curl.h>
#include <regex>

using namespace std;
namespace fs = std::filesystem;

// 配置信息
const string FTP_SERVER = "ftp://example.com/";
const string FTP_USERPWD = "username:password";
const string REMOTE_BASE = "/remote/path/";
const string LOCAL_BASE = "/local/path/";

// 用于存储目录列表的结构体
struct FtpFileInfo {
    string path;
    bool is_directory;
};

// libcurl写入回调
size_t write_callback(void* ptr, size_t size, size_t nmemb, string* data) {
    data->append((char*)ptr, size * nmemb);
    return size * nmemb;
}

// 解析FTP LIST输出
vector<FtpFileInfo> parse_ftp_list(const string& list) {
    vector<FtpFileInfo> files;
    // 改进的正则表达式，更准确匹配UNIX风格列表
    regex unix_pattern(R"(^([d-])([rwx-]{9})\s+\d+\s+\w+\s+\w+\s+(\d+)\s+(\w+\s+\d+\s+[\d:]+)\s+(.+)$)");
    
    istringstream iss(list);
    string line;
    while (getline(iss, line)) {
        smatch match;
        if (regex_match(line, match, unix_pattern)) {
            if (match.size() >= 6) {
                FtpFileInfo info;
                // 修正：访问第一个捕获组（文件类型标识）
                info.is_directory = (match.str() == "d"); // 正确比较方式
                info.path = match.str();  // 第5个捕获组是文件名
                files.push_back(info);
            }
        }
        else {
            cerr << "Unmatched line: " << line << endl;
        }
    }
    return files;
}

// 递归获取远程文件列表
void get_remote_files(CURL* curl, const string& path, set<string>& result) {
    string full_url = FTP_SERVER + path;
    string list_buffer;

    curl_easy_setopt(curl, CURLOPT_URL, full_url.c_str());
    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
    curl_easy_setopt(curl, CURLOPT_WRITEDATA, &list_buffer);
    
    CURLcode res = curl_easy_perform(curl);
    if (res != CURLE_OK) {
        cerr << "curl failed: " << curl_easy_strerror(res) << endl;
        return;
    }

    vector<FtpFileInfo> files = parse_ftp_list(list_buffer);
    for (const auto& fi : files) {
        if (fi.path == "." || fi.path == "..") continue;
        
        string full_path = path + fi.path;
        if (fi.is_directory) {
            get_remote_files(curl, full_path + "/", result);
        } else {
            result.insert(full_path.substr(REMOTE_BASE.length()));
        }
    }
}

// 获取本地文件列表
set<string> get_local_files() {
    set<string> local_files;
    for (const auto& entry : fs::recursive_directory_iterator(LOCAL_BASE)) {
        if (entry.is_regular_file()) {
            string path = entry.path().string().substr(LOCAL_BASE.length());
            // 统一路径分隔符
            replace(path.begin(), path.end(), '\\', '/');
            local_files.insert(path);
        }
    }
    return local_files;
}

// 下载单个文件
bool download_file(CURL* curl, const string& remote_path, const string& local_path) {
    // 创建本地目录
    fs::path local_dir = fs::path(local_path).parent_path();
    if (!fs::exists(local_dir)) {
        fs::create_directories(local_dir);
    }

    ofstream ofs(local_path, ios::binary);
    if (!ofs) {
        cerr << "Failed to create: " << local_path << endl;
        return false;
    }

    string url = FTP_SERVER + REMOTE_BASE + remote_path;
    curl_easy_setopt(curl, CURLOPT_URL, url.c_str());
    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
    curl_easy_setopt(curl, CURLOPT_WRITEDATA, &ofs);

    CURLcode res = curl_easy_perform(curl);
    if (res != CURLE_OK) {
        cerr << "Download failed: " << curl_easy_strerror(res) << endl;
        return false;
    }
    return true;
}

int main() {
    CURL* curl = curl_easy_init();
    if (!curl) {
        cerr << "Failed to initialize libcurl" << endl;
        return 1;
    }

    // 初始化curl选项
    curl_easy_setopt(curl, CURLOPT_USERPWD, FTP_USERPWD.c_str());
    curl_easy_setopt(curl, CURLOPT_FTP_FILEMETHOD, CURLFTPMETHOD_SINGLECWD);
    curl_easy_setopt(curl, CURLOPT_FTP_USE_EPSV, 0L);  // 禁用EPSV

    try {
        // 获取远程文件列表
        set<string> remote_files;
        get_remote_files(curl, REMOTE_BASE, remote_files);
        cout << "Found " << remote_files.size() << " remote files" << endl;

        // 获取本地文件列表
        set<string> local_files = get_local_files();
        cout << "Found " << local_files.size() << " local files" << endl;

        // 比较并下载缺失文件
        for (const auto& file : remote_files) {
            if (!local_files.count(file)) {
                cout << "Downloading: " << file << "...";
                if (download_file(curl, file, LOCAL_BASE + file)) {
                    cout << "OK" << endl;
                }
            }
        }
    } 
    catch (const exception& e) {
        cerr << "Error: " << e.what() << endl;
    }

    curl_easy_cleanup(curl);
    return 0;
}