#!/usr/bin/perl
use strict;
use warnings;
use Data::Dumper;

my %operons;
my %ontologyHash = ();
my $gene;
my @line; 
my @sliceOntology; 

#MAKE HASH OPERONS => GENES  (FROM OPERONSETTEXT.TXT FILE) 

open (DATA, "OperonSetText") || die "Couldnt open operonSet";

while (<DATA>) {
	chomp;
	my ($operon, undef, undef, $genes) = split(/\s+/);
	my @items = split(/,/, $genes);
	foreach (@items) {
	my $gene = (split(/\|/))[0];
	$operon = (split(/-/,$operon))[0]; 
	push @{$operons{$operon}}, $gene;
	}
}

print "$_:", join(',', @{$operons{$_}}), "\n" foreach (sort keys %operons);
#print Dumper(%operons);

#MAKE HASH GENES => ONTOLOGY  (FROM ONTOLOGY.TXT FILE) 


open DATA2, "ontology.txt";


while (<DATA2>) {
	chomp;
	@line = split(/\t+/);							 #Get tab deliminted components into an array. 
	my (undef, undef, $gene) = @line;				 #Get gene name from line array. 
	#print $gene . "  "; 
	@sliceOntology = @line;  		 				
	@sliceOntology = @sliceOntology[12..$#sliceOntology]; #Keep only those after the 10th one. 
	#print " $#sliceOntology @sliceOntology \n ";
	push @{$ontologyHash{$gene}}, @sliceOntology;
	#$ontologyHash{$gene} = \@sliceOntology;
	 
}

#print "$_:", join(',', @{$ontologyHash{$_}}), "\n" foreach (sort keys %ontologyHash);
#print Dumper(%ontologyHash);


#MAKE HASH OPERONS => ONTOLOGY by going through Uri Alon's file, matching operon terms
#from that file, with operon terms in the Operon=>Gene hash above. 


open (DATA3, "coliInterFullNames.txt") || die "Couldnt open 3";

my %opToGo; 
my $count = 0; 
my $opMatch; 

while (<DATA3>) {
	chomp;
	my ($num, $op) = split(/\s+/);
	
	
	#print "$num  $op\n";
	
	#WE NEED TO CONSIDER THE MATCHING PROBLEMS HERE AND FIX THEM. 
	
	#my $opMatch = $op; #(1) If this is a acs-yjcH-actP for example, and op = acs, we still want it to match
					   #therefore, remove anything after the - in the earlier function that creates 
					   #operon names. 
					   # (2) When $op = ada_alkB, this did not match ada-alkB. Since above, we are removing anything after the
					   #-, we can also remove anything after the _ in op. 
					
	
						
					   
	#print $op, "= PRE SPLIT "; 
	$opMatch = (split(/_/, $op))[0]; 
	
	
	
	#print $opMatch , "= SPLIT \n"; 	
	if($operons{$opMatch}){ 
		my @genes = @{$operons{$opMatch}};  
		
		foreach (@genes){ 
			
			if($ontologyHash{$_}){ 
				#print " @{$ontologyHash{$_}} 	, "  ;  
				push @{$opToGo{$num}}, @{$ontologyHash{$_}}  ;

				
			}
		
		} 
	
	}
	else{ 
		print "OPERON FROM ALON FILE: ",$op, " OR " , $opMatch, " DID not match OperonSetText operonName\n"; 
		#print $count += 1; 
		#die; 
		
		   #(3) When $op = 	araFG_araH_1H_2 and $opMatch = araFG this does not match araFGH
					   # in the operons file. We should adjust this so that if a substring in Alon matches a larger string in 
					   # operon, thats fine too. 
					   # Therefore, if there is no match, 
					   # pattern match through the whole of operon keys for this substring.
					   # if a larger string is found that contains it, then turn opMatch into that larger, 
					   # key. 
		
		foreach my $t (sort (keys (%operons))){ 
			if($t =~ /$opMatch/){ 
				print "BUT $t contains $opMatch\n"; 
				#############
						$opMatch = $t;  
						
						#print $opMatch , "= SPLIT \n"; 	
						if($operons{$opMatch}){ 
							my @genes = @{$operons{$opMatch}};  
							
							foreach (@genes){ 
								
								if($ontologyHash{$_}){ 
									#print " @{$ontologyHash{$_}} 	, "  ;  
									push @{$opToGo{$num}}, @{$ontologyHash{$_}}  ;
					
									
								}
							
							} 
									
						}				
									
				############
			} 
		
		
		}

		
		
	}
	
	}

my $k; 
my $multi; 
my $cata;
my $cata2; 

my %opToCat;
my $ma = 0; 

foreach $k (sort (keys (%opToGo)))
{
	 my @a = @{$opToGo{$k}} ;  
	 foreach my $mat (@a){ 
	 	#print " *$mat* ";
	 	#Search the multifun2go file for the multifun catgaory associated with this word.  

		open (DATA4, "multifun2go.txt") || die "Couldnt open 3";
		my @line2 = <DATA4>;
	 		
	 	$ma = 0; 	
		foreach $multi (@line2) {
			#print " $multi "; 
    		if ($multi =~ /\b$mat\b/i) {		#Case insensitive matching is required because there are often capitalized entries 
    										#in the multiFun2Go file, but non-capitalized in the ontology.txt file from Ecocyc. 
    										
       		
       			  $cata = $`; 
       			  if($cata) { 
       			  	if( $cata =~/MultiFun:(\d).+/) { 
       			  		 #print "$k => $1 \n";
 	      				 push @{$opToCat{$k}}, $1 ;		
 	   				 	 $ma = $ma +1; 
      	
       			  	}
       			  	 
       			  }
      		}	
      		
   		}
   		if($ma == 0){ 
   			print "NO MATCH BETWEEN $mat AND MULTIFUNC FUNCTION FILE OR RETRIVAL OF CATAGORY FAILED: Operon = $k, \n"; 
   			#die; 
   			
   			
   		} 
   		
   		
	 	
	 } 
	 #print "\n"; 
	 
	

}
#print Dumper(%opToCat); 

#OpToCat contains a hash of operon number to go Top Level MultiFun Number; 

#Now write an EXCEL flat file containing the data we have isolated. 
#Do this in the order of Uri Alon's members list. 


open (DATA3, "coliInterFullNames.txt") || die "Couldnt open 3";
open (MYFILE, '>>finalData.txt');

while (<DATA3>) {
	chomp;
	my ($num, $op) = split(/\s+/);

	if($opToCat{$num}){
		print MYFILE "$num\t$op\t", join(',', @{$opToGo{$num}}), "\t",  join(',', @{$opToCat{$num}}). "\n"; 
		#print "$num\t$op\t", join(',', @{$opToGo{$num}}), "\t",  join(',', @{$opToCat{$num}}). "\n"; 
	}
	else { 
		print MYFILE "$num\t$op\tNO ONTOLOGY DATA\n";  	
		print  "$num\t$op\tNO ONTOLOGY DATA\n";  	
	}
}


close (MYFILE);