# prom1.pl (GetMapURLs)
#
# Daniel Westreich, October 2002
# Mullins Lab, Microbiology Dept.
# University of Washington
#
# Usage:	prom1.pl filename
# Alt usage:	prom1.pl filename > out.htm
#

use IO::Socket;

#http://www.ncbi.nlm.nih.gov/mapview/map_search.cgi?chr=hum_chr.inf&query=
#&qchr=&advsrch=off

$prefix = "http://";
$host = "www.ncbi.nlm.nih.gov";

$cgiinfo = "/mapview/";
$frontdoc = "map_search.cgi?chr=hum_chr.inf&query=";
$enddoc =  "&qchr=&advsrch=off";

$error = false;
open(OUT, ">mapurls.out");

while(<>) {
    if (/^\#/)
    {
	print "Skipping line...\n";
	next;
    }
    chomp($_);
    $gene = $_;
    print "Beginning our search for gene ".$gene."\n";

# 2.pl
    $mapurl = GetMapURLFromGeneName($gene);
    print OUT $mapurl."\n\n";
    if ($mapurl =~ /Error/)
    {
	$error = true;
    }
}

if ($error eq true)
{
    print "** ERROR.  Check log file mapurls.out for details.\n\n";
    print "When errors are corrected, invoke next step with 'prom2.pl mapurls.out'\n";
    exit();
}
else
{
    print "\n\nNo errors.  Continue with step 2 (command line 'prom2.pl mapurls.out')? [Y/n]"; 

    if (<STDIN> =~/^[nN]/)
    {
	print "Invoke the next step of this process at any time with 'prom2.pl mapurls.out'";
    }
    else
    {
	system("prom2.pl mapurls.out");
    }
}

######################

sub GetMapURLFromGeneName {
    $gene = $_[0];
    $docurl = $cgiinfo.$frontdoc.$gene.$enddoc;
    $cSeqs = 0;

    $remote = IO::Socket::INET->new( Proto => "tcp", PeerAddr => $host, PeerPort => "http(80)", );
    unless($remote)  { die "cannot connect to http daemon on $host" }
    $remote->autoflush(1);
    print $remote "GET $docurl HTTP/1.0\n\n";
    
    while (<$remote>)
    {
	if ($_ =~ /Genes_seq/)
	{
	    $cSeqs++;
	    @words = split(/\s+/);
	    chomp(@words);
	    $length = @words;
	    
	    for ($i = 0; $i < $length; $i++)
	    {
		if (@words[$i] =~ /Genes_seq/) # now at "href=...Genes_seq..."
		{
		    @words = split(/>/,@words[$i]); # now at "href=maps...fill=40"
		    @words = split(/href=/,@words[0]); #now at "maps...fill=40"
		    $Genes_seq_url = $prefix.$host.$cgiinfo.@words[1];
		}
	    }
	}
	if ($cSeqs > 1)
	{
	    last;
	}
    }
    -close $remote;  
    if ($cSeqs > 1)
    {
	return "# $gene:  Error - user must disambiguate Genes_seq tags in ".$prefix.$host.$docurl;
    }
    if ($cSeqs == 0)
    {
	return "# $gene:  Error - did not find a Genes_seq tag in ".$prefix.$host.$docurl;
    }
    print OUT "#Successfully got map URL for $gene\n";
    return $gene."|".$Genes_seq_url;
}