C++ 代码优化：获取所有网页的 NCBI BioProject 结果

#include #include #include #include <curl/curl.h>

// 回调函数，用于处理libcurl接收到的数据 size_t WriteCallback(void* contents, size_t size, size_t nmemb, std::string* output) { size_t total_size = size * nmemb; output->append((char*)contents, total_size); return total_size; }

int main() { // 输入基因名字 std::string gene_name; std::cout << "请输入基因名字："; std::cin >> gene_name;

// 构建搜索URL
std::string search_url = "https://www.ncbi.nlm.nih.gov/bioproject/?term=" + gene_name;

// 初始化libcurl
curl_global_init(CURL_GLOBAL_ALL);
CURL* curl = curl_easy_init();
if (curl) {
    // 设置libcurl选项
    curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L);

    // 创建回调函数的输出字符串
    std::string response;

    // 设置回调函数
    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, WriteCallback);
    curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response);

    // 声明当前页和总页数的变量
    size_t current_page = 1;
    size_t total_pages = 1;

    // 执行HTTP请求直到获取所有页的结果
    while (current_page <= total_pages) {
        // 发送请求并检查是否成功
        curl_easy_setopt(curl, CURLOPT_URL, (search_url + "&page=" + std::to_string(current_page)).c_str());
        CURLcode res = curl_easy_perform(curl);

        if (res != CURLE_OK) {
            std::cerr << "请求失败：" << curl_easy_strerror(res) << std::endl;
            return 1;
        }

        // 在response中查找总页数
        if (current_page == 1) {
            size_t page_start_pos = response.find("Page 1 of ");
            if (page_start_pos != std::string::npos) {
                size_t page_end_pos = response.find("</div>", page_start_pos);
                std::string page_info = response.substr(page_start_pos, page_end_pos - page_start_pos);
                total_pages = std::stoi(page_info.substr(10));
            }
        }

        // 在response中查找Accession号
        size_t start_pos = 0;
        while ((start_pos = response.find("Accession:", start_pos)) != std::string::npos) {
            size_t accession_start = start_pos + 10;
            size_t accession_end = response.find("</dd>", accession_start);
            std::string accession_number = response.substr(accession_start, accession_end - accession_start);
            std::ofstream output_file("accession_numbers.txt", std::ios::app);
            if (output_file.is_open()) {
                output_file << accession_number << std::endl;
                output_file.close();
            } else {
                std::cerr << "无法打开文件" << std::endl;
                return 1;
            }
            start_pos = accession_end;
        }

        // 当前页数加1
        current_page++;
        response.clear(); // 清空response以便获取下一页的结果
    }

    // 关闭libcurl
    curl_easy_cleanup(curl);

    std::cout << "Accession号已写入accession_numbers.txt" << std::endl;
} else {
    std::cerr << "无法初始化libcurl" << std::endl;
    return 1;
}

// 清理libcurl
curl_global_cleanup();

return 0;

}