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

#*****************************************************************************80
#
## cancer_classify_logistic() uses logistic regression to classify cancer data.
#
#  Licensing:
#
#    This code is distributed under the MIT license.
#
#  Modified:
#
#    05 July 2023
#
#  Author:
#
#    Andreas Mueller, Sarah Guido.
#    Modifications by John Burkardt.
#
#  Reference:
#
#    Andreas Mueller, Sarah Guido,
#    Introduction to Machine Learning with Python,
#    OReilly, 2017,
#    ISBN: 978-1-449-36941-5
#
  from sklearn.datasets import load_breast_cancer
  from sklearn.linear_model import LogisticRegression
  import matplotlib.pyplot as plt
  import numpy as np
  import platform
  import sklearn

  print ( '' )
  print ( 'cancer_classify_logistic():' )
  print ( '  Python version: ' + platform.python_version ( ) )
  print ( '  scikit-learn version: '+ sklearn.__version__ )
  print ( '  Use logistic regression on the breast cancer dataset.' )
  print ( '  Explore the influence of the C parameter.' )
#
#  Generate the dataset.
#
  print ( '' )
  print ( '  Retrieve the cancer dataset, (X, y).' )
  cancer = load_breast_cancer ( )

  from sklearn.model_selection import train_test_split
  X_train, X_test, y_train, y_test = train_test_split ( \
    cancer.data, cancer.target, stratify = cancer.target, random_state = 42 )
#
#  Compute training and testing accuracy for varying values of C.
#
  logreg = LogisticRegression().fit ( X_train, y_train )

  print ( '' )
  print ( 'With default C = 1:' )
  print ( '  Training set score ', logreg.score ( X_train, y_train ) )
  print ( '  Test     set score ', logreg.score ( X_test, y_test ) )

  logreg100 = LogisticRegression ( C = 100 ).fit ( X_train, y_train )

  print ( '' )
  print ( 'With C = 100:' )
  print ( '  Training set score ', logreg100.score ( X_train, y_train ) )
  print ( '  Test     set score ', logreg100.score ( X_test, y_test ) )

  logreg001 = LogisticRegression ( C = 0.001 ).fit ( X_train, y_train )

  print ( '' )
  print ( 'With C = 0.01:' )
  print ( '  Training set score ', logreg001.score ( X_train, y_train ) )
  print ( '  Test     set score ', logreg001.score ( X_test, y_test ) )

  plt.clf ( )
  plt.plot ( logreg.coef_.T, 'o', label = 'C=1' )
  plt.plot ( logreg100.coef_.T, '^', label = 'C=100' )
  plt.plot ( logreg001.coef_.T, 'v', label = 'C=0.001' )
  plt.xticks ( range ( cancer.data.shape[1] ), cancer.feature_names, rotation = 90 )
  plt.hlines ( 0, 0, cancer.data.shape[1] )
  plt.ylim ( -5.0, +5.0 )
  plt.xlabel ( 'Feature' )
  plt.ylabel ( 'Coefficient magnitude' )
  plt.title ( 'Coefficients learned by logistic regression, L2 norm' )
  plt.legend ( )
  filename = 'cancer_classify_logistic_l2.png' 
  plt.savefig ( filename )
  print ( '  Graphics saved as "' + filename + '"' )
  plt.close ( )
#
#  Repeat using L1 norm.
#
#  FAILS with error message:
#    "Solver lbfgs supports only 'l2' or 'none' penalties, got l1 penalty"
#
  print ( '' )
  print ( 'Repeat using l1 norm.' )
  print ( '  This exercise omitted; it fails with "l1" penalty.' )

  if ( false ):
    print ( '' )
    plt.clf ( )
    for C, marker in zip ( [ 0.001, 1, 100 ], [ 'o', '^', 'v' ] ):
      lr_l1 = LogisticRegression ( C = C, penalty = "l1" ).fit ( X_train, y_train )
      print ( 'With default C = 1:' )
      print ( '  Training set score, l1 ', lr_l1.score ( X_train, y_train ) )
      print ( '  Test     set score, l1 ', lr_l1.score ( X_test, y_test ) )
      plt.plot ( lr_l1.coef_.T, marker, label = "C = " + str ( C ) )

    plt.xticks ( range ( cancer.data.shape[1] ), cancer.feature_names, rotation = 90 )
    plt.hlines ( 0, 0, cancer.data.shape[1] )
    plt.ylim ( -5.0, +5.0 )
    plt.xlabel ( 'Feature' )
    plt.ylabel ( 'Coefficient magnitude' )
    plt.title ( 'Coefficients learned by logistic regression, L1 norm' )
    plt.legend ( )
    filename = 'cancer_classify_logistic_l1.png' 
    plt.savefig ( filename )
    print ( '  Graphics saved as "' + filename + '"' )
    plt.close ( )
#
#  Terminate.
#
  print ( '' )
  print ( 'cancer_classify_logistic():' )
  print ( '  Normal end of execution.' )

  return

def timestamp ( ):

#*****************************************************************************80
#
## timestamp() prints the date as a timestamp.
#
#  Licensing:
#
#    This code is distributed under the MIT license. 
#
#  Modified:
#
#    21 August 2019
#
#  Author:
#
#    John Burkardt
#
  import time

  t = time.time ( )
  print ( time.ctime ( t ) )

  return

if ( __name__ == '__main__' ):
  timestamp ( )
  cancer_classify_logistic ( )
  timestamp ( )