/*=========================================================================
 *
 *  Copyright Insight Software Consortium
 *
 *  Licensed under the Apache License, Version 2.0 (the "License");
 *  you may not use this file except in compliance with the License.
 *  You may obtain a copy of the License at
 *
 *         http://www.apache.org/licenses/LICENSE-2.0.txt
 *
 *  Unless required by applicable law or agreed to in writing, software
 *  distributed under the License is distributed on an "AS IS" BASIS,
 *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 *  See the License for the specific language governing permissions and
 *  limitations under the License.
 *
 *=========================================================================*/

#include "itkMath.h"
#include "itkMersenneTwisterRandomVariateGenerator.h"
#include "itkListSample.h"
#include "itkKdTreeGenerator.h"
#include <fstream>

int itkKdTreeTest1(int argc , char * argv [] )
{
  if( argc < 4 )
    {
    std::cerr << "Missing parameters" << std::endl;
    std::cerr << "Usage: " << std::endl;
    std::cerr << argv[0] << " numberOfDataPoints numberOfTestPoints bucketSize [graphvizDotOutputFile]" << std::endl;
    return EXIT_FAILURE;
    }

  // Random number generator
  typedef itk::Statistics::MersenneTwisterRandomVariateGenerator NumberGeneratorType;

  NumberGeneratorType::Pointer randomNumberGenerator = NumberGeneratorType::GetInstance();
  randomNumberGenerator->Initialize();

  typedef itk::Array< double > MeasurementVectorType;
  typedef itk::Statistics::ListSample< MeasurementVectorType > SampleType;

  const SampleType::MeasurementVectorSizeType measurementVectorSize = 2;

  SampleType::Pointer sample = SampleType::New();
  sample->SetMeasurementVectorSize( measurementVectorSize );

  //
  // Generate a sample of random points
  //
  const unsigned int numberOfDataPoints = atoi( argv[1] );
  MeasurementVectorType mv( measurementVectorSize );
  for (unsigned int i = 0; i < numberOfDataPoints; ++i )
    {
    mv[0] = randomNumberGenerator->GetNormalVariate( 0.0, 1.0 );
    mv[1] = randomNumberGenerator->GetNormalVariate( 0.0, 1.0 );
    sample->PushBack( mv );
    std::cout << "Add measurement vector: " << mv << std::endl;
    }

  typedef itk::Statistics::KdTreeGenerator< SampleType > TreeGeneratorType;
  TreeGeneratorType::Pointer treeGenerator = TreeGeneratorType::New();

  const unsigned int bucketSize = atoi( argv[3] );

  treeGenerator->SetSample( sample );
  treeGenerator->SetBucketSize( bucketSize );
  treeGenerator->Update();

  typedef TreeGeneratorType::KdTreeType TreeType;

  TreeType::Pointer tree = treeGenerator->GetOutput();

  MeasurementVectorType queryPoint( measurementVectorSize );

  unsigned int numberOfNeighbors = 1;
  TreeType::InstanceIdentifierVectorType neighbors;

  MeasurementVectorType result( measurementVectorSize );
  MeasurementVectorType test_point( measurementVectorSize );
  MeasurementVectorType min_point( measurementVectorSize );

  unsigned int numberOfFailedPoints1 = 0;

  const unsigned int numberOfTestPoints = atoi( argv[2] );

  //
  //  Check that for every point in the sample, its closest point is itself.
  //
  typedef itk::Statistics::EuclideanDistanceMetric< MeasurementVectorType > DistanceMetricType;

  typedef DistanceMetricType::OriginType OriginType;

  DistanceMetricType::Pointer distanceMetric = DistanceMetricType::New();

  OriginType origin( measurementVectorSize );
  for( unsigned int k = 0; k < sample->Size(); k++ )
    {

    queryPoint = sample->GetMeasurementVector(k);

    for ( unsigned int i = 0; i < sample->GetMeasurementVectorSize(); ++i )
      {
      origin[i] = queryPoint[i];
      }

    distanceMetric->SetOrigin( origin );

    // First, get distance from the search API
    std::vector<double> searchDistance;
    tree->Search( queryPoint, numberOfNeighbors, neighbors, searchDistance );

    // We can also get distance from the distance metric
    tree->Search( queryPoint, numberOfNeighbors, neighbors );

    for ( unsigned int i = 0; i < numberOfNeighbors; ++i )
      {
      const double distanceFromMetric =
        distanceMetric->Evaluate( tree->GetMeasurementVector( neighbors[i] ));

      if( distanceFromMetric > itk::Math::eps ||
          searchDistance[i] > itk::Math::eps ||
          itk::Math::NotAlmostEquals( distanceFromMetric, searchDistance[i]  ) )
        {
        std::cerr << "kd-tree knn search result:" << std::endl
                  << "query point = [" << queryPoint << "]" << std::endl
                  << "k = " << numberOfNeighbors << std::endl;
        std::cerr << "measurement vector : distance_by_distMetric : distance_by_tree" << std::endl;
        std::cerr << "[" << tree->GetMeasurementVector( neighbors[i] )
                  << "] : "
                  << distanceFromMetric
                  << " : "
                  << searchDistance[i] << std::endl;
        numberOfFailedPoints1++;
        }
      }
    }

  unsigned int numberOfFailedPoints2 = 0;

  //
  // Generate a second sample of random points
  // and use them to query the tree
  //
  for (unsigned int j = 0; j < numberOfTestPoints; ++j )
    {

    double min_dist = itk::NumericTraits< double >::max();

    queryPoint[0] = randomNumberGenerator->GetNormalVariate( 0.0, 1.0 );
    queryPoint[1] = randomNumberGenerator->GetNormalVariate( 0.0, 1.0 );

    tree->Search( queryPoint, numberOfNeighbors, neighbors );

    //
    // The first neighbor should be the closest point.
    //
    result = tree->GetMeasurementVector( neighbors[0] );

    //
    // Compute the distance to the "presumed" nearest neighbor
    //
    double result_dist = std::sqrt(
          (result[0] - queryPoint[0]) *
          (result[0] - queryPoint[0]) +
          (result[1] - queryPoint[1]) *
          (result[1] - queryPoint[1])
          );

    //
    // Compute the distance to all other points, to verify
    // whether the first neighbor was the closest one or not.
    //
    for( unsigned int i = 0; i < numberOfDataPoints; ++i )
      {
      test_point = tree->GetMeasurementVector( i );

      const double dist = std::sqrt(
          (test_point[0] - queryPoint[0]) *
          (test_point[0] - queryPoint[0]) +
          (test_point[1] - queryPoint[1]) *
          (test_point[1] - queryPoint[1])
          );

      if( dist < min_dist )
        {
        min_dist = dist;
        min_point = test_point;
        }
      }

    if( std::fabs(min_dist - result_dist) > 10.0*itk::NumericTraits<double>::epsilon()*min_dist )
      {
      std::cerr << "Problem found " << std::endl;
      std::cerr << "Query point " << queryPoint << std::endl;
      std::cerr << "Reported closest point " << result
                << " distance " << result_dist << std::endl;
      std::cerr << "Actual   closest point " << min_point
                << " distance " << min_dist << std::endl;
      std::cerr << std::endl;
      std::cerr << "Test FAILED." << std::endl;
      numberOfFailedPoints2++;
      }

    }


  if( argc > 4 )
    {
    //
    // Plot out the tree structure to the console in the format used by Graphviz dot
    //
    std::ofstream plotFile;
    plotFile.open( argv[4] );
    tree->PlotTree( plotFile );
    plotFile.close();
    }


  if( numberOfFailedPoints1 )
    {
    std::cerr << numberOfFailedPoints1 << " out of " << sample->Size();
    std::cerr << " points failed to find themselves as closest-point" << std::endl;
    }

  if( numberOfFailedPoints2 )
    {
    std::cerr << numberOfFailedPoints2 << " out of " << numberOfTestPoints;
    std::cerr << " points failed to find the correct closest point." << std::endl;
    }


  if( numberOfFailedPoints1 || numberOfFailedPoints2 )
    {
    return EXIT_FAILURE;
    }


  std::cout << "Test PASSED." << std::endl;
  return EXIT_SUCCESS;
}