use LWP 5.64;
use URI;
use LWP::Simple;

$dirname = "output/";
get_PDB_protein_names("input/RS126_PDB_temp.txt"); #call subroutine


sub get_PDB_protein_names {# take in 1 variables, inputfilename, save to 3 files
	my($inputfile) = @_;   # pass in 1 varialbe
	
	#open data file
	if (open(MyInputFile, $inputfile)) {
		print "InputFile '",$inputfile,"' opened OK \n";
	}
	else {
		print "Cannot open InputFile '",$inputfile,"'\n";
		exit;
	}
	
	while(defined($lineName=<MyInputFile>)) {
		chomp $lineName;
		#print $lineName."\n";
		
		$lineSeq=<MyInputFile>;
		chomp $lineSeq;
		$lineSecStru=<MyInputFile>;
		chomp $lineSecStru;
		
		print "Step Blast for ".$lineName.": \n";
		
		my $stepurl = "http://www.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Put&QUERY=".$lineSeq."&DATABASE=nr&PROGRAMS=psiBlast&SERVICE=plain&PAGE=Proteins&ALIGNMENTS=2000&DESCRIPTIONS=2000
";
		my $content = get $stepurl;
		
		if (defined $content) {	#check if any problem, mainly due to protein is too long
		# my $content = get $step10url;
		# die "Couldn't get $url" unless defined $content;

			if($content =~ m/RID = \w{11}/g) {
				# print $&."\n";
				$step10RID = substr($&, 6, length($&)-6); 
				print $step10RID."\n";
		
			}
			else {
				$step10RID = "";
				print "They do not have RID information.\n";
			}
		} #end of if
		
		if ($step10RID ne "") { #no RID found for whatever reason, check 
			my $step10url_2 = "http://www.ncbi.nlm.nih.gov/blast/Blast.cgi?RID=".$step10RID."&CMD=Get&ALIGNMENTS=2000&DESCRIPTIONS=2000";
			print $step10url_2."\n";
			sleep 3;
			my $step10Status = "";
	
			while ($step10Status ne "READY") {
				$content = get $step10url_2;
				die "Couldn't get $url" unless defined $content;
				if($content =~ m/Status=\w{5}/g) {
					print $&."...\n";
					$step10Status = substr($&, 7, length($&)-7); 
					#print $step10Status."...\n";			
				}
				sleep 3;
			}
			print "\n";
		}	#end of if

		# write $content to a temporary file for checking
		open(tempOutputFile, ">".$dirname."temp".substr($lineName,0,4).".htm");	
		print tempOutputFile $content;
		close tempOutputFile;
		
		
		
	} #end of while
		
}