#
# Written by:
# -- 
# John L. Weatherwax                2009-04-21
# 
# email: wax@alum.mit.edu
# 
# Please send comments and especially bug reports to the
# above email address.
#
# EPage 787
# 
#-----

restaurant_problem_gen_data = function( N ){
  #
  # A function to generate data from the decision tree given on Figure~18.2
  #
  
  alternate = sample( c("No","Yes"), N, replace=TRUE )
  bar = sample( c("No","Yes"), N, replace=TRUE )
  fri_sat = sample( c("No","Yes"), N, replace=TRUE )
  hungry = sample( c("No","Yes"), N, replace=TRUE )
  patrons = sample( c("None","Some","Full"), N, replace=TRUE )
  price = sample( c("$","$$","$$$"), N, replace=TRUE ) 
  raining = sample( c("No","Yes"), N, replace=TRUE )
  reservation = sample( c("No","Yes"), N, replace=TRUE )
  type = sample( c("French","Italian","Thai","Burger"), N, replace=TRUE )
  waitestimate = sample( c("0-10","10-30","30-60",">60"), N, replace=TRUE )

  # Compute the result given the known truth model:
  # 
  willwait = as.logical( rep(1,N) ) # by default we will wait ==True (won't wait ==False)

  inds = patrons == "None"
  willwait[inds] = FALSE 

  inds = ( patrons == "Full" ) & ( waitestimate == ">60" )
  willwait[inds] = FALSE

  inds = ( patrons == "Full" ) & ( waitestimate == "30-60" ) & ( alternate == "No" ) & ( reservation == "No" ) & ( bar == "No" ) 
  willwait[inds] = FALSE

  inds = ( patrons == "Full" ) & ( waitestimate == "30-60" ) & ( alternate == "Yes" ) & ( fri_sat == "No" ) 
  willwait[inds] = FALSE
  
  inds = ( patrons == "Full" ) & ( waitestimate == "10-30" ) & ( hungry == "Yes" ) & ( alternate == "Yes" ) & ( raining == "No" )
  willwait[inds] = FALSE

  df = data.frame( alternate=alternate, bar=bar, fri_sat=fri_sat, hungry=hungry, patrons=patrons, price=price, raining=raining, reservation=reservation, type=type, waitestimate=waitestimate )
  list( data=df, willwait=willwait ) 
  
}

map_features_to_doubles = function( X ){
  #
  # Maps the given True/False, string, and character features to real numbers [0,1] 
  #
  for( jj in 1:dim(X)[2] ){
    num_column = as.numeric(X[,jj]) 
    mx = max(num_column)
    X[,jj] = num_column/mx 
  }
  X
}