#include <iostream>
#include <curl/curl.h>
#include "pugixml.hpp"
#include <regex>
 
using namespace std;
 
// Callback function to write data received from cURL request
size_t WriteCallback(void* contents, size_t size, size_t nmemb, string* output) {
    size_t totalSize = size * nmemb;
    output->append((char*)contents, totalSize);
    return totalSize;
}
 
// Function to fetch HTML content using cURL
string fetchHTML(const string& url) {
    CURL* curl;
    CURLcode res;
    string response;
 
    curl = curl_easy_init();
    if (curl) {
        curl_easy_setopt(curl, CURLOPT_URL, url.c_str());
        curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, WriteCallback);
        curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response);
        curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1L);
        res = curl_easy_perform(curl);
        curl_easy_cleanup(curl);
    }
    return response;
}
 
// Function to extract pronunciation URLs using regex
void extractPronunciationLinks(const string& html) {
    regex audioRegex(R"((https://d...content-available-to-author-only...e.org/.*?\.mp3))");
    smatch matches;
    string::const_iterator searchStart(html.cbegin());
 
    while (regex_search(searchStart, html.cend(), matches, audioRegex)) {
        cout << "Pronunciation link: " << matches[1] << endl;
        searchStart = matches.suffix().first;
    }
}
 
int main() {
    string words[] = {"personal", "natural", "nationality", "atmosphere"};
 
    for (const string& word : words) {
        string url = "https://d...content-available-to-author-only...e.org/dictionary/english/" + word;
        cout << "Fetching pronunciation for: " << word << endl;
        string htmlContent = fetchHTML(url);
        extractPronunciationLinks(htmlContent);
        cout << "----------------------------------" << endl;
    }
    return 0;
}
 

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