src/CurlWikiGrabber.cpp

   1 #include "CurlWikiGrabber.h"
   2
   3 static size_t write_callback(char *ptr, size_t size, size_t nmemb, void *userdata)
   4 {
   5     static_cast<std::string*>(userdata)->append(ptr, size * nmemb);
   6     return size * nmemb;
   7 }
   8
   9 CurlWikiGrabber::CurlWikiGrabber()
  10 {
  11     int error = curl_global_init(CURL_GLOBAL_ALL);
  12     if(error)
  13     {
  14         throw WalkerException("CURL init failed");
  15     }
  16 }
  17
  18 //! \todo change to passing page title?
  19 //! \todo Curl return code checking
  20 std::string CurlWikiGrabber::grabUrl(std::string url) const
  21 {
  22     CURL *handle = curl_easy_init();
  23
  24     if(NULL == handle)
  25     {
  26         throw WalkerException("error initiating curl");
  27     }
  28
  29     curl_easy_setopt(handle, CURLOPT_URL, url.c_str());
  30     curl_easy_setopt(handle, CURLOPT_USERAGENT, "WikiWalker/ test program");
  31     curl_easy_setopt(handle, CURLOPT_WRITEFUNCTION, write_callback);
  32     std::string gotContent;
  33     curl_easy_setopt(handle, CURLOPT_WRITEDATA, &gotContent);
  34
  35
  36     gotContent = "";
  37     curl_easy_perform(handle);
  38
  39     long httpcode = 0;
  40     curl_easy_getinfo(handle, CURLINFO_RESPONSE_CODE, &httpcode);
  41
  42     curl_easy_cleanup(handle);
  43     handle = NULL;
  44
  45     if(httpcode != 200) {
  46       return "";
  47     }
  48
  49     return gotContent;
  50 }
  51
  52 // note to self: API
  53 // https://en.wikipedia.org/w/api.php
  54 // /w/api.php?action=query&format=json&prop=links&plnamespace=0&titles=<title>
  55 // maybe &pllimit=100