#! /usr/bin/perl -w
use strict;

#Script to return the percentage of each well's cell population BELOW the supplied threshold for P-S780RB1 per well as well as the mean value for nuclear antibody intensity per well.

#Settings are intended for reading data from Nuclei.csv as generated using the 3_channels_pipeline.cppipe file in conjunction with Cell Profiler.

sub get_line {
	print $_[0];
	chomp (my $line = <STDIN>);
	$line;
}

#Define 96 well plate addresses in @plate A1-H12

my (@column, @row, @plate, $addr, $row, $column);

@column = qw (1 2 3 4 5 6 7 8 9 10 11 12);
@row = qw (A B C D E F G H);

foreach $column (@column) {
	foreach $row (@row) {
		$addr = ("$row"."$column");
		push (@plate, $addr);
	}
}

#Get user input via questionnaire

my ($source, $dest, $threshold);

if ($#ARGV == 2) {
	$source = $ARGV[0];
	open IN, $source
		or die "Can't open '$source' for input: $!";
	$dest = $ARGV[1];
		die "Won't overwrite existing file"
	if -e $dest;
	open OUT, ">>$dest" #append destination file
		or die "Can't open '$dest' for output: $!";
		$threshold = $ARGV[2];
		} else {
	$source = &get_line("Which source file? ");
	open IN, $source
		or die "Can't open '$source' for input: $!";
	
	$dest = &get_line("Which destination file? ");
		die "Won't overwrite existing file"
	if -e $dest;
		open OUT, ">>$dest" #append destination file
			or die "Can't open '$dest' for output: $!";
			$threshold = &get_line("What threshold for hypo-P-S780 RB?: ");
}

# Read each line corresponding to each well address in turn and output a single line of summarized data for each well in the output file.

my ($well);
print OUT "Threshold\: $threshold\n";
print OUT "Well,Total Count,Threshold Passed,Percent Hypo-P,Mean Ab Intensity\n";

foreach $well (@plate) {
	seek(IN, 0, 0);
	my ($tcount, $gated, $found, $accum);
	while (<IN>) {
		if (/^\d{1,4},\d{1,4},Red,Blue,Green,\d{1,2},($well),[0-9]+,[A-H],\d{1,4},\d*.\d*,\d*.\d*,(\d*.\d*),\d*.\d*\r*\n/) { #refer to notes in 2gate_classifier.pl for modification guidelines for this regex line
						$tcount += 1;
						$accum += $2;
			if ($2 <= $threshold) { #NB MIN vs MAX gate adjusted here #gates applied here can use gt > lt < or equal to == symbols
				$gated += 1;
			}
			$found = 1;
		}
	}
	if (defined $found) {
		my $score = (($gated / $tcount) * 100);
		my $mean = ($accum / $tcount);
		print OUT "$well\,$tcount\,$gated\,$score\,$mean\n";
		}
}
	
close IN;
close OUT;