#!/usr/bin/perl

# Stacia Wyman, 5 Jan 05, Winter Study CS011 

# HW_1_1
#
# The following code allows you to read all the file names from a particular
# directory into an array. Note that you can use variable names in the double
# quotes around the name of the directory to be opened.
$prefix = "/home/faculty/stacia/ws05/data/metazoa/";
opendir(METAZOA,"$prefix") || die "Couldn't open dir";
@ALL_FILES = readdir(METAZOA);
closedir(METAZOA);

foreach $file (@ALL_FILES){
    chomp($file);    # I wasn't sure if this was needed (it wasn't), 
		     # but it doesn't hurt to have it.
    if ($file =~ /\.gbk$/) {	    # make sure it's a GenBank file
	$filename = $prefix.$file;  # $file doesn't contain path, so add it 
	open(GBFILE,$filename) || die "Couldn't open $!";
	while (<GBFILE>){
	    if(/^DEFINITION\s+(.+) mitochon/){
		print "$1\n";
	    }
	}
	close(GBFILE);
    } 
}
close(OUTFILE);




#!/usr/bin/perl

# Stacia Wyman, 5 Jan 05, Winter Study CS011 

# HW_1_2
# 
# Very similar to number one, but we also record coordinates whenever we
# see:	gene	num..num
# Then, as soon as we see /gene="COX1", we output the saved coords.

$prefix = "/home/faculty/stacia/ws05/data/metazoa/";
opendir(METAZOA,"$prefix") || die "Couldn't open dir";
@ALL_FILES = readdir(METAZOA);
closedir(METAZOA);

# I chose to use coordstr as boolean flag which is 1 when we've 
# read coordinates, *except* if we just printed out COX1's coordinates,
# then it is zero so it won't print out the coordinates twice for each time
# it sees the text /gene="COX1". There are many different and correct ways to
# do this.
$coordstr = 0;
foreach $file (@ALL_FILES){
    chomp($file);
    if ($file =~ /\.gbk$/) {
	$filename = $prefix.$file;
	open(GBFILE,$filename) || die "Couldn't open $!";
	while (<GBFILE>){
	    if(/^DEFINITION\s+(.+) mitochon/){
		print "$1, ";
	    }
	    if(/gene\s+(\d+\.\.\d+)/){  # have to escape the . here
		$coordstr = $1;
	    }
	    if (/gene="COX1"/ && $coordstr){
		print "$coordstr\n";
		$coordstr = 0;
	    }
	}
	close(GBFILE);
    } 
}
close(OUTFILE);