#! /usr/bin/env python3
#
def faithful_kmeans2 ( ):

#*****************************************************************************80
#
## faithful_kmeans2() does a simple clustering exercise using scipy kmeans2().
#
#  Discussion:
#
#    Clustering data.
#
#  Licensing:
#
#    This code is distributed under the MIT license.
#
#  Modified:
#
#    22 March 2025
#
#  Author:
#
#    John Burkardt
#
  import matplotlib.pyplot as plt
  import numpy as np
  import platform
  from scipy.cluster.vq import kmeans2

  print ( '' )
  print ( 'faithful_kmeans2():' )
  print ( '  Python version: ' + platform.python_version ( ) )
  print ( '  Cluster Old Faithful data using scipy kmeans2().' )
#
#  Read the normalized data;
#  For convienence, make separate xn and yn copies.
#
  datan = np.loadtxt ( 'faithful_normalized.txt' )
  xn = datan[:,0]
  yn = datan[:,1]
  n = len ( xn )
  print ( '' )
  print ( '  Number of data values is ', n )
#
#  Display the data.
#
  plt.plot ( xn, yn, 'k.', markersize = 10 )
  plt.xlabel ( '<-- Duration (normalized) -->' )
  plt.ylabel ( '<-- Wait (normalized) -->' )
  plt.title ( 'Old Faithful eruption durations and waits' )
  plt.grid ( True )
  plt.axis ( 'equal' )
  filename = 'faithful_normalized_data.png'
  plt.savefig ( filename )
  print ( '  Graphics saved as "' + filename + '"' )
  plt.show ( )
  plt.clf ( )
#
#  Choose the number of clusters, and call kmeans2().
#
  k = 2
  c, label = kmeans2 ( datan, k )
#
#  Report the cluster centers.
#
  print ( '  Kmeans2 cluster centers C:' )
  print ( c )
#
#  Separate the data into "blue" and "red" groups, based on
#  which cluster they belong to.
#
#  Compute the separate energies.
#
  bd = ( xn - c[0,0] )**2 + ( yn - c[0,1] )**2
  rd = ( xn - c[1,0] )**2 + ( yn - c[1,1] )**2
#
#  Sum the blue and red cluster energies.
#
  cost0 = sum ( bd[label==0] )
  cost1 = sum ( rd[label==1] )
  cost = cost0 + cost1
  print ( '  Cluster variance = ', cost, '=', cost0, '+', cost1 )
#
#  Count the blue and red cluster elements.
#
  bn = sum ( label==0 )
  cn = sum ( label== 1)
  print ( '  Cluster size = ',  bn + cn, '=', bn, '+', cn )
#
#  Plot the red and blue data, and their cluster centers.
#
  plt.clf ( )
  plt.plot ( xn[label==0], yn[label==0], 'c.', markersize = 10 )
  plt.plot ( xn[label==1], yn[label==1], 'm.', markersize = 10 )
  plt.plot ( c[0,0], c[0,1], 'bo', markersize = 15 )
  plt.plot ( c[1,0], c[1,1], 'ro', markersize = 15 )
  plt.xlabel ( '<-- Duration -->' )
  plt.ylabel ( '<-- Wait -->' )
  plt.title ( 'Clusters using kmeans2()' )
  plt.grid ( True )
  filename = 'faithful_kmeans2.png'
  plt.savefig ( filename )
  print ( '  Graphics saved as "' + filename + '"' )
  plt.show ( )
  plt.clf ( )
#
#  Terminate.
#
  print ( '' )
  print ( 'faithful_kmeans2():' )
  print ( '  Normal end of execution.' )

  return

if ( __name__ == '__main__' ):
  faithful_kmeans2 ( )