#include <iostream>
#include <clustering/API.h>
#include <clustering/Parameters.h>

int main(int argc, char **argv)
{
  // First of all, create a Parameters object with a similarity
  // measure and a similarity threshold
  Clustering::Parameters params(Clustering::Jaccard, 0.3);

  // Set path used to store temporary files (for distributed sort,
  // pairs computation, ...)
  params.setTemporaryPath("/tmp");
  // Choose what part of index to put in memory or on disk
  params.setMainIndexInMemory();
  params.setRevertIndexInMemory();

  // Number of threads when process is parallel (distributed sort,
  // pairs computation, ...)
  params.setNumberOfThread(3);

  // Now we can construct the API object using Parameters
  Clustering::API api(params);
  
  // Creation of a new document that contains 2 words
  api.addDocument("doc1");
  api.addDescriptorInCurrentDocument("word1");
  api.addDescriptorInCurrentDocument("word2");

  // Creation of a new document that contains 2 words
  api.addDocument("doc2");
  api.addDescriptorInCurrentDocument("word1");
  api.addDescriptorInCurrentDocument("word2");

  // Creation of a new document that contains 3 words
  api.addDocument("doc3");
  api.addDescriptorInCurrentDocument("word3");
  api.addDescriptorInCurrentDocument("word4");
  api.addDescriptorInCurrentDocument("word5");

  // Creation of a new document that contains 3 words
  api.addDocument("doc4");
  api.addDescriptorInCurrentDocument("word3");
  api.addDescriptorInCurrentDocument("word4");
  api.addDescriptorInCurrentDocument("word5");

  // Creation of a new document that contains 3 words
  api.addDocument("doc5");
  api.addDescriptorInCurrentDocument("word5");
  api.addDescriptorInCurrentDocument("word6");
  api.addDescriptorInCurrentDocument("word7");

  // Creation of a new document that contains 4 words
  api.addDocument("doc6");
  api.addDescriptorInCurrentDocument("word4");
  api.addDescriptorInCurrentDocument("word5");
  api.addDescriptorInCurrentDocument("word6");
  api.addDescriptorInCurrentDocument("word7");

  // launch algorithm
  const Clustering::Results &res = api.startAlgorithm();
  
  // iterates over clusters
  Clustering::DocClusterIterator it = res.begin(), eit = res.end();
  unsigned clusterCnt = 0;
  while (it != eit)
    {
      std::cout << "Cluster[" << ++clusterCnt << "]" << std::endl;

      // Get Documents of this cluster
      const std::vector<Clustering::DocId> &documents = it->getDocs();
      // Get decriptors of this cluster
      const std::vector<Clustering::DescId> &descriptors = it->getDescs();

      unsigned int i = 0;
      // We must use api to convert DocId and DescId in string
      // Display documents in cluster
      for (i = 0; i < documents.size(); ++i)
	std::cout << "  Document[" << i + 1 << "] : " 
		  << api.getDocument(documents[i]) << std::endl;
      // We can display all descriptors present in this cluster
      for (i = 0; i < descriptors.size(); ++i)
	std::cout << "  Descriptor[" << i + 1 << "] Freq[" << descriptors[i].getFrequency()
		  << "]: " << api.getDescriptor(descriptors[i]) << std::endl;
      // We can also choose to display only the N best descriptors
      std::vector<Clustering::DescId> bestDescriptors = 
	api.getBestDesc(*it, 2, 0.1);
      for (i = 0; i < bestDescriptors.size(); ++i)
	std::cout << "  Best Descriptor[" << i + 1 << "] Freq[" << bestDescriptors[i].getFrequency()
		  << "]: " << api.getDescriptor(bestDescriptors[i]) << std::endl;
      ++it;
    }
}