#!/usr/bin/perl
#Take variables from the fields filled out on the previous web page 
require 'miner-lib';

my $dir = "/share/bread/www-data/data/webminer";

&ReadParse;
foreach (keys(%in)) {
	$in{$_} =~ s/[^A-Za-z0-9\.\,\_\-\[\]\{\}\*\?\+\\]//g;
	}
#Print appropriate header information for HTML or spreadsheet output
if ($in{tabdelimited}) {
	print "Content-type: spreadsheet/tab-delimited\n\n";
	}
else {
	print &PrintHeader;
	}
#Decide whether to print second page (where you set thresholds)
#or final output page (with results of the search)
if ($in{makejspage}) {
	&MakeJSPage;
	}
else {
	&DoSearch;
	}
#Generate second page, which is a form listing the criteria selected on
#the first page and a field to set the expression threshold (or motif
#to search for, etc.) for each criterion
sub MakeJSPage {
	print '<script language="JavaScript1.1">';
	print "\nvar myPopup\nvar popupSize = ";
	print '"toolbar=1,scrollbars=1,location=1,status=1,menubar=1,resizable=1,width=400,height=250"';
	print "\n\n";
	print "function popup () {\n";
	print 'myPopup = window.open ("", "popup", popupSize)';
	print "\n}\n\n";
	print "</script>\n";
	@choices = ("empty", split(/\\0/, $in{chosen}));
	$number_of_crits = $#choices;
	print "<img src=http://genome-www.stanford.edu/webminer/images/wmtr.gif>";
	print "<form method=post action=http://genome-www.stanford.edu/cgi-bin/webminer/miner>";
	foreach $x (1..$#choices) {
		if ($choices[$x] =~ /^exp\.orfs\.maxh/) {
			&PrintMaxh;
			}
		elsif ($choices[$x] =~ /^exp\.orfs\.pep/) {
			&PrintMotif;
			}
		elsif ($choices[$x] =~ /^exp\.orfs\.prom/) {
			&PrintPromoter;
			}
		elsif ($choices[$x] =~ /^exp\.orfs\.iso/) {
			&PrintIso;
			}
		elsif ($choices[$x] =~ /^exp\.orfs\.mw/) {
			&PrintMass;
			}
		elsif ($choices[$x] =~ /^exp\.orfs\.dups/) {
			&PrintDups;
			}
		elsif ($choices[$x] =~ /^exp\.([^\.]*)\.([^\.]*)/) {
			&LookupExpNames;
			$class = $1;
			$expt = $2;
			&PrintExpJS;
			}
		print "<p>";
		}
	
	print "<input type=hidden name=critnumber value=$number_of_crits>";
	print "<input type=checkbox name=tabdelimited>Output as tab-delimited instead of HTML table<br>";
	print "<input type=submit>";
	}
#Print the criterion fields for a MaxH search
sub PrintMaxh {
	print "Criterion $x: <b>Maximum hydrophobicity</b> value of the protein";
	print "<br><i><a href=http://genome-www.stanford.edu/webminer/help/maxh.html ";
	print 'target="popup" onClick=popup()';
	print ">Help with MaxH searches</i></a><br>";
	print "<input type=hidden name=crit$x value=maxh>";
	print "Proteins with MaxH values ";
	print "<select name=moreless$x>";
	print '<option value="more">above';
	print '<option value="less">below';
	print "</select>";
	print "<input type=text size=3 name=maxh$x>";
	}
#Print the criterion fields for a protein motif search
sub PrintMotif {
	print "Criterion $x: Proteins containing the following <b>motif</b>";
	print "<br><i><a href=http://genome-www.stanford.edu/webminer/help/protein.html";
	print ' target="popup" onClick=popup()';
	print ">Help with protein motif searches</i></a><br>";
	print "<input type=hidden name=crit$x value=motif>";
	print "Proteins with the motif ";
	print "<input type=text size=40 name=motif$x>";
	}
#Print the criterion fields for a promoter motif search
sub PrintPromoter {
	print "Criterion $x: Genes with the following <b>promoter sequence</b> within 1 kb upstream of their start codons";
	print "<br><i><a href=http://genome-www.stanford.edu/webminer/help/promoter.html ";
	print 'target="popup" onClick=popup()';
	print ">Help with promoter searches</i></a><br>";
	print "<input type=hidden name=crit$x value=promoter>";
	print "Promoters that contain ";
	print "<input type=text size=40 name=promoter$x>";
	}
#Print the criterion fields for an isoelectric point search
sub PrintIso {
	print "Criterion $x: Predicted <b>isoelectric point</b> of the protein<br>";
	print "<input type=hidden name=crit$x value=iso>";
	print "Proteins with isoelectric points ";
	print "<select name=moreless$x>";
	print '<option value="more">higher than';
	print '<option value="less">lower than';
	print "</select>";
	print " pH <input type=text size=3 name=iso$x>";
	}
#Print the criterion fields for a molecular weight search
sub PrintMass {
	print "Criterion $x: Predicted <b>molecular weight</b> of the protein<br>";
	print "<input type=hidden name=crit$x value=mass>";
	print "Proteins with molecular weight ";
	print "<select name=moreless$x>";
	print '<option value="more">greater than';
	print '<option value="less">less than';
	print "</select>";
	print "<input type=text size=3 name=mass$x> kDa";
	}
#Print the criterion fields for a gene duplicates search
sub PrintDups {
	print "Criterion $x: Genes contained in <b>duplicated blocks</b> within the genome";
	print "<br><i><a href=http://genome-www.stanford.edu/webminer/help/dups.html ";
	print 'target="popup" onClick=popup()';
	print ">Help with duplicate gene searches</i></a><br>";
	print "<input type=hidden name=crit$x value=dups>";
	print "Genes that ";
	print "<select name=dups$x>";
	print '<option value="dup">are';
	print '<option value="not_dup">are not';
	print "</select>";
	print " contained in a duplicated block in the genome";
	}
#Get the experiment descriptors for the gene expression searches that 
#were chosen as criteria (builds the %classdescs and %exptdescs arrays)
sub LookupExpNames {
	open(IN, "<$dir/classdescs");
	while (<IN>) {
		/^([^\t]*)\t([^\t]*)/;
		$class = $1;
		$longcrit = $2;
		$classdescs{$class} = $longcrit;
		}
	open(IN, "<$dir/exptdescs");
	while (<IN>) {
		/^([^\t]*)\t([^\t]*)/;
		$expt = $1;
		$longcrit = $2;
		$exptdescs{$expt} = $longcrit;
		}
	}
#Print the criterion fields for a gene expression search
sub PrintExpJS {
	if (!$classdescs{$class}) {$classdescs{$class} = $class;}
	print "Criterion $x: Expression in <b>$classdescs{$class}</b><br>";
	print "<input type=hidden name=crit$x value=exp>";
	print "Genes which are induced ";
	print "<select name=moreless$x>";
	print "<option value=more>more than";
	print "<option value=less>less than";
	print "</select>";
	print "<input type=text size=3 name=induction$x>";
	print "-fold in the experiment ";
	print "<input type=hidden name=expt$x value=$class.$expt>";
	if (!$exptdescs{$class.".".$expt}) {$exptdescs{$class.".".$expt} = $expt;}
	print "<b>" . $exptdescs{$class . "." . $expt} . "</b><br>";
	}
#Go through each criterion and score every ORF as passing it or not
#and then print the results table with every ORF (linked to YPD), its
#gene name, the matching criteria values for all search criteria, and
#its description and pathway information
sub DoSearch {
	foreach $i (1..$in{critnumber}) {
		$crit = "crit" . $i;
		if ($in{$crit} eq "exp") {
			&DoExpSearch;
			}
		if ($in{$crit} eq "maxh") {
			&DoMaxhSearch;
			}
		if ($in{$crit} eq "motif") {
			&DoMotifSearch;
			}
		if ($in{$crit} eq "promoter") {
			&DoPromoterSearch;
			}
		if ($in{$crit} eq "iso") {
			&DoIsoSearch;
			}
		if ($in{$crit} eq "mass") {
			&DoMassSearch;
			}
		if ($in{$crit} eq "dups") {
			&DoDupsSearch;
			}
		}
	&GetNames;
	&GetPaths;
	&GetDescs;
	foreach $matched_orf (keys(%matches)) {
		if (!($names{$matched_orf})) {
			$names{$matched_orf} = $matched_orf;
			}
		$matchcounts[$matches{$matched_orf}]++;
		push(@matched_names, $names{$matched_orf});
		$orf_by_name{$names{$matched_orf}} = $matched_orf;
		}
	if ($in{tabdelimited}) {
		$format = "<pre>";
		$endformat = "</pre>";
		$line = "\n";
		$sep = "\t";
		}
	else {
		print "<table><tr>";
		print "<td><img src=http://genome-www.stanford.edu/webminer/images/wmtr.gif>";
		print "<td><center><font size=+2>Webminer has found<br>";
		$i = $in{critnumber}-3;
		if ($i<1) {$i=1};
		foreach $j ($i..$in{critnumber}-1) {
			$num_matches = 0;
			foreach $k ($j..$#matchcounts) {
				$num_matches += $matchcounts[$k];
				}
			print "$num_matches ORFs meeting $j ";
			if ($j > 1) {
				print "criteria,<br> ";
				}
			else {
				print "criterion,<br>";	
				}
			}
		if ($#matchcounts>1) {
			print "and ";
			}
		print "<font color=red>";
		if (!$matchcounts[$in{critnumber}]) {
			$matchcounts[$in{critnumber}] = 0;
			}
		print "$matchcounts[$#matchcounts] ORFs";
		print "</font><br>";
		print "meeting "; 
		if ($#matchcounts>1) {
			print "all of ";
			}
		print "your search criteria.</font></table>";
		$format="<table>";
		$endformat = "</table>";
		$line = "<tr>";
		$sep = "<td>";
		}		
	print "$format";
	print "$line$sep" . "ORF" . $sep . "Name$sep";
	&LookupExpNames;
	foreach $z (1..$in{critnumber}) {
		if ($in{"expt" . $z})	{
			print $exptdescs{$in{"expt" . $z}} . "$sep";
			}
		if ($in{"crit" . $z} =~ /maxh/) {
			print "MaxH$sep";
			}
		if ($in{"crit" . $z} =~ /motif/) {
			$motif = $in{"motif" . $z};
			print "Motif '$motif'$sep";
			}
		if ($in{"crit" . $z} =~ /promoter/) {
			$promoter = $in{"promoter" . $z};
			print "Promoter '$promoter'$sep";
			}
		if ($in{"crit" . $z} =~ /mass/) {
			print "M.W.$sep";
			}
		if ($in{"crit" . $z} =~ /iso/) {
			print "Isoelectric Pt.$sep";
			}
		if ($in{"crit" . $z} =~ /dups/) {
			print "Duplicated Block$sep";
			}
		}
	print "Pathway $sep";
	print "Description $sep";
	print "$line";
	foreach $matched_name (sort(@matched_names)) {
		$matched_orf = $orf_by_name{$matched_name};
		if ($matches{$matched_orf} == $in{critnumber}) {
			print "$sep";
			if (!$in{tabdelimited}) {
				print "<a href=http://genome-www4.stanford.edu/cgi-bin/SGD/locus.pl?locus=$matched_orf>";
				}
			print "$matched_orf";
			if (!$in{tabdelimited}) {
				print "</a>";
				}
			print "$sep";
			print "$names{$matched_orf}$sep";
			foreach $z (1..$in{critnumber}) {
				$crit = "crit" . $z;
				print "$$crit{$matched_orf}$sep";
				}
			print "$paths{$matched_orf}$sep";
			print "$descs{$matched_orf}$sep";
			print "$line";
			}
		}
	print $endformat;
	}
#Do a gene expression search (open the appropriate data file and read in
#all ORFs and their expression values.  ORFs matching the criterion have
#their value in the array %matches increased by 1.)  
sub DoExpSearch {
	$datafile = $in{"expt" . $i};
	$moreless = $in{"moreless" . $i};
	$induction = $in{"induction" . $i};
	open(IN, "<$dir/minerdata/$datafile");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$value = $2;
		$orf =~ tr/a-z/A-Z/;
		$orf =~ s/([WC])([AB])/$1-$2/;
		if ($value && (!$$crit{$orf})) {
			if ($moreless =~ /more/) {
				if ($value > $induction) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			else {
				if ($value < $induction) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			}
		}
	}
#Do a MaxH search (open the MaxH data file and read in all ORFs and their
#MaxH values.  ORFs matching the criterion have their value in the array
#%matches increased by 1.)  
sub DoMaxhSearch {
	$moreless = $in{"moreless" . $i};
	$maxh = $in{"maxh" . $i};
	open(IN, "<$dir/minerdata/orfs.maxh");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$value = $2;
		if ($value && (!$$crit{$orf})) {
			if ($moreless =~ /more/) {
				if ($value > $maxh) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			else {
				if ($value < $maxh) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			}
		}
	}	
#Do a protein motif search (open the protein sequence data file and read
#in all ORFs and their predicted protein sequences.  ORFs matching the
#criterion have their value in the array %matches increased by 1.)  
sub DoMotifSearch {
	$motif = $in{"motif" . $i};
	$motif =~ tr/a-z/A-Z/;
	$motif =~ s/[^A-Z0-9\,\.\*\-\{\}\[\]\^\+\?]//g;
	open(IN, "<$dir/minerdata/orfs.pep");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$pep = $2;
		if ($pep && (!$$crit{$orf})) {
			if ($pep =~ /($motif)/) {
				$matches{$orf}++;
				$$crit{$orf} = $1;
				}
			}
		}
	}	
#Do a promoter motif search (open the promoter sequence data file and read
#in all ORFs and their predicted promoter sequences.  ORFs matching the
#criterion have their value in the array %matches increased by 1.)  
sub DoPromoterSearch {
	$promoter = $in{"promoter" . $i};
	$promoter =~ tr/a-z/A-Z/;
	$promoter =~ s/[^0-9A-Z\,\.\*\-\{\}\[\]\^\+\?]//g;
	$promoter =~ s/N/./g;
	$promoter =~ s/Y/[CT]/g;
	$promoter =~ s/R/[AG]/g;
	$promoter =~ s/M/[AC]/g;
	$promoter =~ s/K/[GT]/g;
	$promoter =~ s/S/[GC]/g;
	$promoter =~ s/W/[AT]/g;
	
	open(IN, "<$dir/minerdata/orfs.prom");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$prom = $2;
		if ($prom && (!$$crit{$orf})) {
			if ($prom =~ /($promoter)/) {
				$matches{$orf}++;
				$$crit{$orf} = $1;
				}
			}
		}
	}	
#Do an isoelectric pt search (open the iso pt data file and read in
#all ORFs and their products' predicted isoelectric points.  ORFs matching
#the criterion have their value in the array %matches increased by 1.)  
sub DoIsoSearch {
	$moreless = $in{"moreless" . $i};
	$iso = $in{"iso" . $i};
	open(IN, "<$dir/minerdata/orfs.iso");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$value = $2;
		if ($value && (!$$crit{$orf})) {
			if ($moreless =~ /more/) {
				if ($value > $iso) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			else {
				if ($value < $iso) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			}
		}
	}	
#Do a molecular weight search (open the mol wt data file and read in
#all ORFs and their products' predicted masses.  ORFs matching the
#criterion have their value in the array %matches increased by 1.)  
sub DoMassSearch {
	$moreless = $in{"moreless" . $i};
	$mass = $in{"mass" . $i};
	open(IN, "<$dir/minerdata/orfs.mw");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$value = $2;
		if ($value && (!$$crit{$orf})) {
			if ($moreless =~ /more/) {
				if ($value > $mass) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			else {
				if ($value < $mass) {
					$matches{$orf}++;
					$$crit{$orf} = $value;
					}
				}
			}
		}
	}	
#Do a gene duplicates search (open the duplicates data file and read in
#all ORFs and what duplicate block, if any, they are in.  ORFs matching
#the criterion have their value in the array %matches increased by 1.)  
sub DoDupsSearch {
	$want_dups = $in{"dups" . $i};
	open(IN, "<$dir/minerdata/orfs.dups");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf = $1;
		$orf =~ tr/a-z/A-Z/;
		$dup_block = $2;
		if ((($want_dups eq "dup") && $dup_block && (!$$crit{$orf})) || (($want_dups eq "not_dup") && (!($dup_block)) && (!$$crit{$orf}))) {
			$matches{$orf}++;
			$$crit{$orf} = $dup_block;
			}
		}
	}	
sub PrintExp {
	open(LS, "ls $dir/minerdata/$dataclass*|");
	@expts = "";
	while (<LS>) {
		chop;
		/minerdata\/$dataclass\.(.*)/;
		$suffix = $1;
		push(@expts, $suffix);
		}
	print "Criterion $x: Expression in <b>$longcrit</b> dataset<br>";
	print "<input type=hidden name=crit$x value=exp>";
	print "Genes which are induced ";
	print "<select name=moreless$x>";
	print "<option value=more>more than";
	print "<option value=less>less than";
	print "</select>";
	print "<input type=text size=3 name=induction$x>";
	print "-fold in the experiment ";
	print "<select name=expt$x>";
	foreach $expt (@expts) {
		if ($expt) {
			print "<option value=$dataclass.$expt>$expt";
			}
		}
	print "</select>";
	}
#Get gene names for all ORFs, store them in the %names array
sub GetNames {
	open(IN, "<$dir/orfs.names");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf=$1;
		$name=$2;
		$names{$orf} = $name;
		}
	}
#Get descriptions of what pathway each gene functions in, store them in
#the %paths array
sub GetPaths {
	open(IN, "<$dir/orfs.paths");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf=$1;
		$path=$2;
		$paths{$orf} = $path;
		}
	}
#Get descriptions of the function of each gene, store them in the %descs
#array
sub GetDescs {
	open(IN, "<$dir/orfs.descs");
	while (<IN>) {
		chop;
		/^(.*)\t(.*)/;
		$orf=$1;
		$desc=$2;
		$descs{$orf} = $desc;
		}
	}