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

#*****************************************************************************80
#
## pizza_kmeans does a simple clustering exercise.
#
#  Discussion:
#
#    There are three pizza trucks, red, green, and blue.
#    Each house is served by exactly one truck.
#    The cost of this system is the total of the distance between each
#    house and its truck.
#
#    Using k-means, we seek to minimize this cost, by adjusting the
#    assignment of trucks to houses, and by adjusting the location of
#    the trucks.
#
#  Licensing:
#
#    This code is distributed under the GNU LGPL license.
#
#  Modified:
#
#    24 September 2019
#
#  Author:
#
#    John Burkardt
#
  import matplotlib.pyplot as plt
  import numpy as np

  print ( '' )
  print ( 'pizza_kmeans:' )
  print ( '  MATH1900 Selected Topics in Mathematics: Machine Learning' )
#
#  Read the house data, x, y, truck
#
  house = np.loadtxt ( 'pizza_houses.txt' )
  x = house[:,0]
  y = house[:,1]
  t = house[:,2]
  n = len ( x )
  print ( '  Number of houses is %d' % ( n ) )
#
#  Read the truck data.
#
  c = np.loadtxt ( 'pizza_trucks.txt' )
#
#  Determine distance from each house to the red, green, and blue trucks.
#
  rd = np.sqrt ( ( x - c[0,0] )**2 + ( y - c[0,1] )**2 )
  gd = np.sqrt ( ( x - c[1,0] )**2 + ( y - c[1,1] )**2 )
  bd = np.sqrt ( ( x - c[2,0] )**2 + ( y - c[2,1] )**2 )
#
#  Create lists of the houses in the red, green, and blue clusters.
#
  rc = np.where ( t == 0 )
  gc = np.where ( t == 1 )
  bc = np.where ( t == 2 )
#
#  The cost is the sum of the distances between every house and its truck.
#
  cost = sum ( bd[bc] ) + sum ( rd[rc] ) + sum ( gd[gc] )
  print ( '  Cost of initial clustering is ', cost )
#
#  Step 0
#
  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Houses, Trucks, and Clusters, Step 0', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step0.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 1: Update houses to nearest truck.
#
  print ( '' )
  print ( '  Assign each house to nearest truck.' )

  rd = np.sqrt ( ( x - c[0,0] )**2 + ( y - c[0,1] )**2 )
  gd = np.sqrt ( ( x - c[1,0] )**2 + ( y - c[1,1] )**2 )
  bd = np.sqrt ( ( x - c[2,0] )**2 + ( y - c[2,1] )**2 )

  rc = np.where ( ( rd < bd ) & ( rd < gd ) )
  gc = np.where ( ( gd < bd ) & ( gd < rd ) )
  bc = np.where ( ( bd < rd ) & ( bd < gd ) )

  cost = sum ( bd[bc] ) + sum ( rd[rc] ) + sum ( gd[gc] )
  print ( '  Cost is ', cost )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Assign houses to nearest truck, Step 1', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step1.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 2: Move trucks to center of clusters.
#
  print ( '' )
  print ( '  Move each truck to cluster center.' )

  c[0,0] = np.mean ( x[rc] )
  c[0,1] = np.mean ( y[rc] )
  c[1,0] = np.mean ( x[gc] )
  c[1,1] = np.mean ( y[gc] )
  c[2,0] = np.mean ( x[bc] )
  c[2,1] = np.mean ( y[bc] )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Move Trucks to Centers, Step 2', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step2.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 3: Update houses to nearest truck.
#
  print ( '' )
  print ( '  Assign each house to nearest truck.' )

  rd = np.sqrt ( ( x - c[0,0] )**2 + ( y - c[0,1] )**2 )
  gd = np.sqrt ( ( x - c[1,0] )**2 + ( y - c[1,1] )**2 )
  bd = np.sqrt ( ( x - c[2,0] )**2 + ( y - c[2,1] )**2 )

  rc = np.where ( ( rd < bd ) & ( rd < gd ) )
  gc = np.where ( ( gd < bd ) & ( gd < rd ) )
  bc = np.where ( ( bd < rd ) & ( bd < gd ) )

  cost = sum ( bd[bc] ) + sum ( rd[rc] ) + sum ( gd[gc] )
  print ( '  Cost is ', cost )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Assign houses to nearest truck, Step 3', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step3.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 4: Move trucks to center of clusters.
#
  print ( '' )
  print ( '  Move each truck to cluster center.' )

  c[0,0] = np.mean ( x[rc] )
  c[0,1] = np.mean ( y[rc] )
  c[1,0] = np.mean ( x[gc] )
  c[1,1] = np.mean ( y[gc] )
  c[2,0] = np.mean ( x[bc] )
  c[2,1] = np.mean ( y[bc] )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Move Trucks to Centers, Step 4', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step4.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 5: Update houses to nearest truck.
#
  print ( '' )
  print ( '  Assign each house to nearest truck.' )

  rd = np.sqrt ( ( x - c[0,0] )**2 + ( y - c[0,1] )**2 )
  gd = np.sqrt ( ( x - c[1,0] )**2 + ( y - c[1,1] )**2 )
  bd = np.sqrt ( ( x - c[2,0] )**2 + ( y - c[2,1] )**2 )

  rc = np.where ( ( rd < bd ) & ( rd < gd ) )
  gc = np.where ( ( gd < bd ) & ( gd < rd ) )
  bc = np.where ( ( bd < rd ) & ( bd < gd ) )

  cost = sum ( bd[bc] ) + sum ( rd[rc] ) + sum ( gd[gc] )
  print ( '  Cost is ', cost )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Assign houses to nearest truck, Step 5', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step5.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 6: Move trucks to center of clusters.
#
  print ( '' )
  print ( '  Move each truck to cluster center.' )

  c[0,0] = np.mean ( x[rc] )
  c[0,1] = np.mean ( y[rc] )
  c[1,0] = np.mean ( x[gc] )
  c[1,1] = np.mean ( y[gc] )
  c[2,0] = np.mean ( x[bc] )
  c[2,1] = np.mean ( y[bc] )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Move Trucks to Centers, Step 6', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step6.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 7: Update houses to nearest truck.
#
  print ( '' )
  print ( '  Assign each house to nearest truck.' )

  rd = np.sqrt ( ( x - c[0,0] )**2 + ( y - c[0,1] )**2 )
  gd = np.sqrt ( ( x - c[1,0] )**2 + ( y - c[1,1] )**2 )
  bd = np.sqrt ( ( x - c[2,0] )**2 + ( y - c[2,1] )**2 )

  rc = np.where ( ( rd < bd ) & ( rd < gd ) )
  gc = np.where ( ( gd < bd ) & ( gd < rd ) )
  bc = np.where ( ( bd < rd ) & ( bd < gd ) )

  cost = sum ( bd[bc] ) + sum ( rd[rc] ) + sum ( gd[gc] )
  print ( '  Cost is ', cost )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Assign houses to nearest truck, Step 7', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step7.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Step 8: Move trucks to center of clusters.
#
  print ( '' )
  print ( '  Move each truck to cluster center.' )

  c[0,0] = np.mean ( x[rc] )
  c[0,1] = np.mean ( y[rc] )
  c[1,0] = np.mean ( x[gc] )
  c[1,1] = np.mean ( y[gc] )
  c[2,0] = np.mean ( x[bc] )
  c[2,1] = np.mean ( y[bc] )

  plt.plot ( x[rc], y[rc], 'r.', markersize = 20 )
  plt.plot ( x[gc], y[gc], 'g.', markersize = 20 )
  plt.plot ( x[bc], y[bc], 'b.', markersize = 20 )
  plt.plot ( c[0,0], c[0,1], 'r*', markersize = 25 )
  plt.plot ( c[1,0], c[1,1], 'g*', markersize = 25 )
  plt.plot ( c[2,0], c[2,1], 'b*', markersize = 25 )
  plt.xlabel ( '<-- X -->', fontsize = 16 )
  plt.ylabel ( '<-- Y -->', fontsize = 16 )
  plt.title ( 'Move Trucks to Centers, Step 8', fontsize = 16 )
  plt.grid ( True )
  filename = 'pizza_step8.png'
  plt.savefig ( filename )
  plt.show ( )
  plt.clf ( )
#
#  Terminate.
#
  print ( '' )
  print ( 'pizza_kmeans:' )
  print ( '  Normal end of execution.' )

  return

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