util/FilterAnalyzedArticles.cpp

   1 #include <algorithm>
   2 #include <iostream>
   3 #include <fstream>
   4 #include <vector>
   5
   6 #include "Article.h"
   7 #include "ArticleCollection.h"
   8 #include "JsonSerializer.h"
   9 #include "ToGraphvizWriter.h"
  10
  11
  12 int main(int argc, char** argv)
  13 {
  14   if(argc != 3) {
  15     std::cout << argv[0] << " "
  16               << "<cache file>"
  17               << "<dot file>" << std::endl;
  18     return 1;
  19   }
  20
  21   WikiWalker::JsonSerializer jser;
  22   std::ifstream cache(argv[1]);
  23
  24   // assumption: having write-only access to a file is so rare that I don't care
  25   // also, currently the file is used for both read and write, so initially it
  26   // won't exist.
  27   if(!cache.is_open()) {
  28     std::cerr << "Couldn't open cache file" << std::endl;
  29     return 1;
  30   }
  31
  32   WikiWalker::CollectionUtils::ArticleCollection ac;
  33   jser.deserialize(ac, cache);
  34
  35   if(cache.fail()) {
  36     cache.close();
  37     std::cerr << "Error reading from file" << std::endl;
  38     return 1;
  39   }
  40
  41   WikiWalker::CollectionUtils::ArticleCollection filteredColl;
  42   for(auto& anArticle : ac) {
  43     auto art = anArticle.second;
  44     if(art->analyzed()) {
  45       WikiWalker::CollectionUtils::add(filteredColl, art);
  46     }
  47   }
  48
  49   WikiWalker::ToGraphvizWriter g;
  50   std::ofstream filteredGraph(argv[2]);
  51
  52   if(!filteredGraph.is_open()) {
  53     std::cerr << "Couldn't open dot file for writing" << std::endl;
  54     return 1;
  55   }
  56
  57   g.output(filteredColl, filteredGraph);
  58
  59   if(filteredGraph.fail()) {
  60     filteredGraph.close();
  61     std::cerr << "Error writing to file" << std::endl;
  62     return 1;
  63   }
  64
  65   filteredGraph.close();
  66
  67   std::cout << "There are " << filteredColl.size() << " analyzed articles"
  68             << std::endl;
  69
  70   return 0;
  71 }