[med-svn] r1289 - trunk/community/infrastructure

[smoe-guest at alioth.debian.org]

Author: smoe-guest
Date: 2008-02-05 16:15:41 +0000 (Tue, 05 Feb 2008)
New Revision: 1289

Modified:
   trunk/community/infrastructure/getData.pl
Log:
Works!  Added KEGG and IntACT. Reads /etc/default/getData.



Modified: trunk/community/infrastructure/getData.pl
===================================================================
--- trunk/community/infrastructure/getData.pl	2008-02-04 13:04:05 UTC (rev 1288)
+++ trunk/community/infrastructure/getData.pl	2008-02-05 16:15:41 UTC (rev 1289)
@@ -28,11 +28,28 @@
 # Or else go to GNU Web pages http://www.gnu.org and follow the white rabbit.
 #
 
-my $rootdir = "/var/lib/mirrored";
+my $mirrordir;
 
+my $configfile="/etc/default/getData";
+if ( -r $configfile ) {
+   open(FH,"<$configfile") or die "Could not open config data at '$configfile'.\n";
+   while(<FH>) {
+      next if /^\s*#/;
+      next unless /\s*(\S.*\S)\s*=\s*"([^"]*)"\s*/;
+      #print STDERR "Read: $1\n";
+      if ($1 eq "mirrordir") {
+      	$mirrordir=$2;
+	print STDERR "Set mirrordir to '$mirrordir'.\n";
+      }
+   }
+   close FH;
+}
+
+$mirrordir = "/var/lib/mirrored" unless defined($mirrordir);
+
 =head1 SYNOPSIS
 
-getData.pl [ --rootdir <path> ] <list of db names>
+getData.pl [ --mirrordir <path> ] <list of db names>
 getData.pl --list
 
 =head1 DESCRIPTION
@@ -67,9 +84,9 @@
 
     Present a more detailed description in form of a man page.
 
-=item --rootdir <path>
+=item --mirrordir <path>
 
-    Specifies destination directory. The data will be mirrored to $rootdir/$dbname/
+    Specifies destination directory. The data will be mirrored to $mirrordir/$dbname/
 
 =item --list
 
@@ -111,6 +128,8 @@
 
 my %toBeMirrored = (
 
+#                     A  S  T  R  O  N  O  M  Y
+
 #	"tycho2" => {
 #		name => "Tycho2 Star Coordinates",
 #		source => "wget --mirror ftp://cdsarc.u-strasbg.fr/pub/cats/I/259/tyc2.dat*",
@@ -127,7 +146,31 @@
 #		   . "-a ftp.lowell.edu/pub/elgb/astorb.dat.gz -nt astorb.dat ] "
 #		   . "|| zcat ftp.lowell.edu/pub/elgb/astorb.dat.gz > astorb.dat"
 #	},
-#
+
+#                    B  I  O  I  N  F  O  R  M  A  T  I  C  S
+
+	"intact.psimitab" => {
+		name  => "IntACT Protein Interaction Database",
+		source => "wget --mirror ftp://ftp.ebi.ac.uk/pub/databases/intact/current/psimitab/intact.zip",
+		"post-download" => "unzip `find . -name '*.zip'`"
+	},
+
+	"kegg.brite" => {
+		name => "Kegg Brite Database",
+		source => "wget --mirror ftp://ftp.genome.jp/pub/kegg/release/current/brite.tar.gz"
+	},
+
+	"kegg.pathway" => {
+		name => "Kegg Pathway Database",
+		source => "wget --mirror ftp://ftp.genome.jp/pub/kegg/release/current/pathway.tar.gz"
+	},
+
+	"kegg.pathway.hsa" => {
+		name => "Kegg Pathway Database (HSA)",
+		source => "wget --mirror ftp://ftp.genome.jp/pub/kegg/pathway/organisms/hsa"
+	},
+
+
 #	"DE405" => {
 #		name => "DE405",
 #		source => "wget --mirror ftp://ssd.jpl.nasa.gov/pub/eph/export/unix/unxp2[01]*.405",
@@ -163,51 +206,8 @@
 
 my ($list,$help,$man)=(0,0,0);
 
-my %options = (
-	"rootdir:s" => \$rootdir,
-	"list" => \$list,
-	"help"=>\$help,
-	"man"=>\$man);
+my %options=("mirrordir:s" => \$mirrordir, "list" => \$list, "help"=>\$help, "man"=>\$man);
 
-sub help () {
-    require "Pod/Usage.pm";
-    import Pod::Usage;
-    pod2usage() unless defined($man);
-    pod2usage( -verbose => 2 );
-}
-
-
-help() if !GetOptions(%options) or $help or $man;
-
-
-if ($list) {
-	foreach my $db (sort keys %toBeMirrored) {
-		printf "%-10s",$db;
-		# more information like the expected size should be printed here.
-		print "\t".$toBeMirrored{$db}{"name"};
-		print "\n";
-	}
-	exit 0;
-}
-
-die "Cannot write to root destination directory at '$rootdir'.\n"
-	unless  -w "$rootdir" ;
-
-#print join(", ", at ARGV)."\n";
-
-my $d=getcwd();
-
-foreach my $db (@ARGV) {
-	unless(exists($toBeMirrored{$db})) {
-		print STDERR "Unknown database: '$db'\n";
-		next;
-	}
-};
-
-my ($list,$help,$man)=(0,0,0);
-
-my %options=("rootdir:s" => \$rootdir, "list" => \$list, "help"=>\$help, "man"=>\$man);
-
 sub myhelp () {
     require "Pod/Usage.pm";
     import Pod::Usage;
@@ -229,7 +229,7 @@
 	exit 0;
 }
 
-die "Cannot write to root destination directory at '$rootdir'.\n" unless ( -w "$rootdir" );
+die "Cannot write to root destination directory at '$mirrordir'.\n" unless ( -w "$mirrordir" );
 
 #print join(", ", at ARGV)."\n";
 
@@ -241,21 +241,25 @@
 		next;
 	}
 
-	print STDERR "\"$db\" -> \"$rootdir\"\n";
+	print STDERR "\"$db\" -> \"$mirrordir\"\n";
 	print "Mirroring ".$toBeMirrored{$db}{"name"}." ($db)\n";
 
-	unless ( -d "$rootdir/$db" ) {
-		print "  creating directory $rootdir/$db\n";
-		mkdir("$rootdir/$db") or die "Could not create directory \"$rootdir/$db\"\n";
+	unless ( -d "$mirrordir/$db" ) {
+		print "  creating directory $mirrordir/$db\n";
+		mkdir("$mirrordir/$db") or die "Could not create directory \"$mirrordir/$db\"\n";
 	}
-	chdir("$rootdir/$db") or die "Could not change directory to \"$rootdir/$db\"\n";
+	chdir("$mirrordir/$db") or die "Could not change directory to \"$mirrordir/$db\"\n";
 
-	my $cmd = $toBeMirrored{$db}{"source"}."\n";
+	my $cmd = $toBeMirrored{$db}{"source"};
+	if (!defined ($cmd) or "" eq $cmd) {
+		print STDERR "$db: download instructions not specified - skipping.\n";
+		next;
+	}
 	print STDERR "$cmd\n";
 	system($cmd) and die "Experienced problem.";
 
-	$cmd = $toBeMirrored{$db}{"post-download"}."\n";
-	if ( defined($cmd) and "" != $cmd) {
+	$cmd = $toBeMirrored{$db}{"post-download"};
+	if ( defined($cmd) and "" ne $cmd) {
 		print STDERR "$cmd\n";
 		system($cmd) and die "Experienced problem.";
 	}