#!/usr/bin/perl -w
$too_many_cels = 100; ##more than this is too much.
$numfiles=@ARGV;
$default_output_file = "cel2r.tab";
##first we figure out what are the cdf file and the cel files.
if($numfiles>2){ ##this means arguments were give
    @cels = @ARGV;
    $cdf = shift(@cels);
    $outfile = pop(@cels);
}
elsif($numfiles==2){ #this means arguments were given, put not enough
       die "You must specify at least one cdf, one cel file, and one output file.\n";
   }
else{ ##if no arguments are given we assume all .cel files are of interest. if 1 arg this is the outputfile
    ## and that there exactly one cdf file exists
    @cdfs = glob("*.[Cc][Dd][Ff]");
    $ncdf = @cdfs;
    if($ncdf != 1){
	die "There are $ncdf cdf files in this directory.\nIf you don't pass arguments there must be exactly one.\n";
    }	
    else{
	$cdf = $cdfs[0];
    }
    @cels = glob("*.[Cc][Ee][Ll]");
    $numcel = @cels;
    if($numcel<1 || $numcel>$too_many_cels){
	die "You specified $numcel cel files.\nYou must specify between 1 and $too_many_cels.\n";
    }
    if($numfiles){ ##if 1, its the output
	$outfile = $ARGV[0];
    }
    else{
	$outfile = $default_output_file;    
    }
}
###now we now that $cdf is the cdf file and @cels contains the cel files
##to be put together
$info = $cdf;
$info =~ s/[Cc][Dd][Ff]/info/;

open(INFO,">$info") || die "Sorry, I cound't open a Info file!\n";
open(CDF,$cdf) || die "Sorry, I coulnd't open a CDF file!\n";
$found = 0;
##first get number of columns and rows
while(defined($_=<CDF>) & $found< 2){
    chomp;
    if (/Cols./){
	s/Cols=//;
	$cols = $_;
	chop($cols);
	$found +=1;
    }
    elsif (/Rows./){
	s/Rows=//;
	$rows = $_;
	chop($rows);
	$found +=1;
    }
}
##result will contain all the information of the pm and mm
@result = (0 .. ($cols*$rows-1));
$gene_id_counter = 0;
while(<CDF>){
    chomp;
    if (/\[.+\]/) {  ##if surounded by [ ] it is a new qc or new block.
	s/\[|\]//g; ##get rid of [ ]
    }
    if (/QC./){ ##this is a control probe
	chop; ##get rid of anoying ^M
	$gene_name = $_;
	$pm_or_mm = -1;
	while (<CDF>){ ##now we go get the index for this QC
	    chomp;
	    if (/NumberCells./){ ##this says how many cell, so we can loop
		$n = $_;
		$n =~ s/NumberCells=//;
		last; ##done, get out of looking for NumAtoms;
	    }
	}
	while (<CDF>){ ##now look for CellHeader
	    chomp;
	    if(/CellHeader./){
		for($i=0; $i<$n;$i++){ ##we found it, now we extract names
		    $check = 0;
		    while($check<=5){ ##make sure you find something and not a hole!
			@line = split("\t",<CDF>);
			chomp(@line);
			$check = @line; 
		    }
		    $x = $line[0];
		    $x =~ s/Cell\d+=//;
		    $y = $line[1];
		    $result[$line[5]]= ##$line[5] has the index
			join("\t",0,$i,$pm_or_mm,$x,$y);
		}
		last; ##we are done, get out of this QC
	    }
	}
    }
    elsif (/Unit\d+_Block\d+./){ ##this is a block with gene info
	while(<CDF>){ ##look for name
	    chomp;
	    if(/Name=./){
		$gene_id_counter = $gene_id_counter + 1;
		$gene_name = $_; ##next line is gene name
		chop($gene_name); ##get rid of anoying ^M
		$gene_name =~ s/Name=//; #take name= out
		print INFO "$gene_name\n";
		last;
	    }
	}
	while(<CDF>) { ##go until you fin NumAtoms
	    chomp;
	    if(/NumAtoms./){
		$n = $_;
		$n =~ s/NumAtoms=//;
		last; ##done, get out of looking for NumAtoms;
	    }
	}
	while (<CDF>){ ##now look for CellHeader
	    chomp;
	    if(/CellHeader./){
		for($i=0; $i<$n;$i++){ ##we found it, now we extract names
		    $check = 0;
		    while($check<=5){ ##make sure you find something and not a hole!
			@line1 = split("\t",<CDF>);
			chomp(@line1);
			$check = @line1; 
		    }
		    $x1 = $line1[0];
		    $x1 =~ s/Cell\d+=//;
		    $y1 = $line1[1];
		    $check = 0;
		    while($check<=5){ ##make sure you find something and not a hole!
			@line2 = split("\t",<CDF>);
			chomp(@line2); 
			$check = @line2; 
		    }
		    $x2 = $line2[0];
		    $x2 =~ s/Cell\d+=//;
		    $y2 = $line2[1];
##this assumes they come in pairs
		    if($y1 < $y2){
			$pm_or_mm1 = 1;
			$pm_or_mm2 = 0;
		    }
		    else{
			$pm_or_mm1 = 0;
			$pm_or_mm2 = 1;
		    }
		    $result[$line1[11]]= ##$line[11] has the index
			join("\t",$gene_id_counter,$i,$pm_or_mm1,$x1,$y1);
		    $result[$line2[11]]= 
			join("\t",$gene_id_counter,$i,$pm_or_mm2,$x2,$y2);
		}
		last; #and we are done
	    }
	}
    }
}
close(CDF);
close(INFO);
##check for spots on array with no gene
$total = 0;
for ($i=0; $i<($cols*$cols); $i++){
    if ($result[$i] eq $i){
	$x = $i % $cols;
	$y = ($i - $x)/$cols;
	$result[$i] = join("\t",-1,1,-1,$x,$y);
	$total += 1;
    }
}
##NOW WE ARE READY TO START WITH THE CEL FILES
$names = "";
foreach $cel (@cels){
    open(CEL,$cel) || die "Sorry, I coulnd't open $cel!\n";
    $chip_name = $cel;
    chomp($chip_name);
    $chip_name =~ s/\.[Cc][Ee][Ll]//;
    $names = join("\t",$names,$chip_name);
    while(<CEL>){
	chomp;
	if (/CellHeader./){ ##here we go;
	    while(<CEL>){
		@line=split("\t",$_);
		$length = @line;
		if($length > 2){
		    $key= $line[0]+$line[1]*$cols;
		    $result[$key] = join("\t",$result[$key],$line[2]);
		}
		else{
		    ##we are done
		    last;
		}
	    }  		
	}     
    }
    close(CEL);
}
$names = "$names\n";

open(OUT,">$outfile")  || die "Sorry, I coulnd't open $outfile!\n";;
$temp = join("\t","id","number","pm.or.mm","x","y");
$temp = "$temp $names";
print OUT $temp;
for ($i=0; $i<($cols*$rows); $i++){
    print OUT "$result[$i]\n";
}
close(OUT);