get_pages.cpp

std::vector<Page> get_pages(const std::vector<std::string>& urls) {
  std::vector<Page> pages;
  std::map<std::string, int> url_conv;
  for(int i = 0; i < static_cast<int>(urls.size()); ++i) {
    pages.emplace_back(i, urls[i]);
    url_conv[urls[i]] = i;
  }

  for(int i = 0; i < static_cast<int>(pages.size()); ++i) {
    std::string dirpath = [](const std::string& url) {
      auto pos = url.find_last_of('/');
      if(pos == std::string::npos) return std::string("./");
      return url.substr(0, pos+1);
    }(pages[i].url());
    std::vector<std::string> links = filter_with_regex(get_words_in(pages[i].url()), ".*\\.html$");
    for(const std::string& to_url : links) {
      if(url_conv.count(dirpath+to_url)) pages[i].add_link(url_conv[dirpath+to_url]);
    }
  }

  return pages;
}