#!/usr/local/bin/perl 

$file1 = $ARGV[0];
$file2 = $ARGV[1];
$file3 = $ARGV[2];

if ($#ARGV != 2)
{
    print "Usage: $0 needs 3 arguments: Naive-Bayes.pl lastnames keynames trueornot\n";
    exit();
}

# 1 = IS a South asian name
# 0 = IS NOT a South asian name

@lastnames = ( );                           #array for set of last names
$counter_lastnames = 0;                     #Counts the No. of last names
$counter_no_sa_names = 0;                   #Counts the No. of S. Asian Names in our training data set
$counter_no_NONsa_names = 0;                #Counts the No. of NON S. Asian Names in our training data
@lngthnames = ( );                          #Array for length of each last name
@lngthkeys = ( );                           #Array for length of each key
                                            #@NEWKEY = array for set of ID keys
@trueornotSA = ( );                         #True case of whether its SA or NOT
@condprobabilitiesYTrue = ( );              #Array for Prob{X_1 = 1 | Y = 1}
@condprobabilitiesYFalse = ( );             #Array for Prob{X_1 = 1 | Y = 0}


open(FILE1,"<$file1");                      #First open file of all possible LAST NAMES
while ($line = <FILE1>)                     
        {	
           chomp $line;                     #getting read of "\n"
           push(@lastnames, $line);         #Building last name array from each line
           $namelngth = length($line);      #determining length of each last name 
           push(@lngthnames, $namelngth);   #Building length of last names     
           $counter_lastnames++;            #keeping count of the # of last names in the data set
	}
close(FILE1);


open(FILE2,"<$file2");                      #Opening file containing ID keys "chat" "jee" etc.
while ($line = <FILE2>)      
  {
    if ($line =~ /(\S+)/)                   #ensuring that file is of the "proper" type
    {
       @NEWKEY = split /\s+/, $line;        #Splits all the ID keys into NEWKEY array
    }
  }
close(FILE2);


open(FILE3,"<$file3");                      #Opening File of 1's or 0's (T or F) for South Asian Names
while ($line = <FILE3>)                     
  {
    chomp $line;                            #getting read of "\n"
    push(@trueornotSA, $line);              #Building list of S. Asian True or False
    if ($line == 1)
      {
	  $counter_no_sa_names++;           #Number of true South Asian Names
      }
    else
      {
          $counter_no_NONsa_names++;        #Number of true NON South Asian Names
      }
  }
close(FILE3);


$number_of_keys = scalar(@NEWKEY);          #Determines the size of the array NEWKEY

for ($j=0; $j < $number_of_keys; $j++)      #Make the array for length of each ID key
    {
	$keycount = length(@NEWKEY[$j]);    #getting the length of each tuple (ID key)
        push(@lngthkeys, $keycount);        #Building an array for the length of each ID Key
    }

@scoreindex = ();                           #keeps track of the Bayesian score for each last name
@maxscorename = ();                         #maximum possible attainable score per name --> name length - 3

for ($i=0; $i < $counter_lastnames; $i++)                          #Looping through ALL the lastname lists
    {
        $score = 0;                                                #Variable for assigning score to each test sample
	for ($j=0; $j < $number_of_keys; $j++)                     #Loop through each ID KEY (tuple)
	    {
                $counter_name_length = @lngthnames[$i];            #Going through array of Last Name Lengths
                $counter_key_length =  @lngthkeys[$j];             #Going through array of ID Key Lengths


                #Next we need to ENSURE that each LAST NAME length is >= the length of each ID KEY (maximum of 4)
                #i.e. Name is "Ray" and ID is "chakr" -- we have a problem then!

                if (($counter_name_length - $counter_key_length) >= 0)
                {
                   for ($k=0; $k <= ($counter_name_length - $counter_key_length); $k++)
		      {
			$substringkey = substr(@lastnames[$i], $k, $counter_key_length);      #start @ $k and move right $counter_key_length units
                        if ($substringkey eq @lastnames[$i])                                  #Next determine: names are size of the ID Keys
                         { 
                           #print "The short lastname here is  @lastnames[$i]\n";
                         }
                        if (@NEWKEY[$j] eq $substringkey)
                          {
			      $score++;                                                       #update the score matrix
                          }
                      }
                } 
            }
        push(@scoreindex, $score);                                                            #Keeping track of the final score matrix
        push(@maxscorename, @lngthnames[$i]-3);                                               #keep track of MAX. possible score attainable per name = name length - min(ID KEY)
     } 


######################TESTING PURPOSES#####################################
                                                                         
print "Name\t\t Score\t SAorNO\n";                                                                                                
for ($k=0; $k < $counter_lastnames; $k++)                                                                                
{                                                                                                                        
    print "@lastnames[$k]\t@scoreindex[$k]\t@trueornotSA[$k]\n";                                                                         
    print "@lastnames[$k] is @trueornotSA[$k] name\n";                                          
}                                                                                                                        
                                                                         
######################END OF TESTING PURPOSES##############################