Added GPML read-in, STDOUT/file out/manipulation tests.

svn path=/; revision=136

diff --git a/preecej/perl_singletons/pathway_gene_swapper.pl b/preecej/perl_singletons/pathway_gene_swapper.pl
index b6134a0545374a5c6b4ed7a5c6a14355efc63a09..f907f456f7af47aad7877f1d86b048d9545fb5c5 100644 (file)
--- a/preecej/perl_singletons/pathway_gene_swapper.pl
+++ b/preecej/perl_singletons/pathway_gene_swapper.pl
@@ -119,8 +119,8 @@ Returns a hash with config values set.
 # ---------------------------------------------------------------------------
 sub config($)
 {
-    if ($debug) { print "...Config file path: $_[0]\n"; }
-
+    
+    print "Opening configuration file and reading data...\n\n";
     my %local_config_hash;
 
     open(CONFIG_FILE, $_[0]) or die("Could not open $_[0]");
@@ -136,6 +136,8 @@ sub config($)
        $local_config_hash{$data_field} = $data_val; 
     }
     
+    close(CONFIG_FILE);
+    
     return %local_config_hash;
 }
 
@@ -210,31 +212,23 @@ sub init
         . "\n";
 
     %configs = config($input_config_file);
-    
-    if ($debug) { print "...<DEBUG: \%configs>\n"
-        . Dumper(\%configs) . "\n\n"; }
 }
 
 
 
 # ---------------------------------------------------------------------------
 =item B<void import_data()>
-Reads, parses, and stores source GPML.
+Reads, parses, and stores gene mapping file and source GPML.
 =cut
 # ---------------------------------------------------------------------------
 sub import_data
 {
-    print "Opening input file and reading header info...\n\n";
-    #open the file                                                                                                          
-    #read a line
-    #skip header line
-    #split line on comma
-    #get original gene and swapped gene
-    #add genes to %swap_genes
-    open(CONFIG_FILE, $input_gene_file) or die("Could not open $input_gene_file"); 
+    print "Opening gene mapping file and reading data...\n\n";
+
+    open(GENE_FILE, $input_gene_file) or die("Could not open $input_gene_file"); 
     
-    my $line = <CONFIG_FILE>;
-    while (<CONFIG_FILE>)
+    my $line = <GENE_FILE>;
+    while (<GENE_FILE>)
     {
        $line = $_;
        chomp $line;
@@ -244,12 +238,13 @@ sub import_data
 
        $swap_genes{$data_field} = $data_val; 
     } 
-    #test:display %swap_genes
-     if ($debug) { print "...<DEBUG: \%swap_genes>\n"
-        . Dumper(\%swap_genes) . "\n\n"; }
-     
-    
-    
+
+    close(GENE_FILE);
+
+    print "Opening GPML pathway file and reading data...\n\n";
+
+    my $parser = new XML::DOM::Parser;
+    $gpml_doc = $parser->parsefile($input_gpml_file);
 }
 
 
@@ -261,7 +256,16 @@ Verbose only.
 # ---------------------------------------------------------------------------
 sub show_input
 {
+    print "[Configuration Settings]\n";
+    print Dumper(\%configs) . "\n\n";
+    print "\n";
+
+    print "[Gene Mappings]\n";
+    print Dumper(\%swap_genes) . "\n\n";
+    print "\n";
+
     print "[Source GPML]\n";
+    print $gpml_doc->toString;
     print "\n";
 }
 
@@ -272,17 +276,25 @@ Substitutes gene data.
 # ---------------------------------------------------------------------------
 sub swap_genes
 {
-    print "Swapping gene data...\n\n";
+    print "Swapping gene data and making other modifications...\n\n";
         
     # -------------------------------------------------------------------------
     # [PathVisio Perl Pseudo-Script]
     # -------------------------------------------------------------------------
     # .remove all <BiopaxRef> and <bp:PublicationXref> elements and children
+    # my $pub_nodes = $gpml_doc->getElementsByTagName("bp:PublicationXref");
+    # for my $pub_node (\$pub_nodes)
+    # {
+    #     $pub_node->removeChild();
+    # }
+    
     # .identify mapped genes
     #   .if mapped gene did not belong to a group, create new group node
     #   .duplicate <DataNode> for each mapping
     #     .create new 5-digit hex code "GraphId"
-    #   .if needed, add new group node (create new 5-digit hex code)
+    #   .if needed, add new group node
+    #     .create new, random 5-digit hex code
+    #       .make sure it doesn't already exist
     #     .add "GroupRef" on each added node, including the orig.
     #   .rename TextLabel (prefix: Eu-, suffix: -#?)
     #   .add Comment back-referencing TAIR locus id (use "source" attribute)
@@ -291,36 +303,21 @@ sub swap_genes
     #   .decrement CenterX and CenterY by 10px each
     #   .change box Width if needed
     #   .color the box (<Graphics...Color="4488ff" />)
-    # .After PathVisio sanity check:
+    # .After PathVisio sanity check, before communication:
     #   .remove back-ref TAIR comments
 
     
-    # -------------------------------------------------------------------------
-    # [XML::DOM] 
-    # -------------------------------------------------------------------------
-    # my $parser = new XML::DOM::Parser;
-    # my $doc = $parser->parsefile ("file.xml");
     
-    # # print all HREF attributes of all CODEBASE elements
-    # my $nodes = $doc->getElementsByTagName ("CODEBASE");
+    # [usage example]
+    # my $nodes = $gpml_doc->getElementsByTagName("DataNode");
     # my $n = $nodes->getLength;
     
     # for (my $i = 0; $i < $n; $i++)
     # {
-    #  my $node = $nodes->item ($i);
-    #  my $href = $node->getAttributeNode ("HREF");
-    #  print $href->getValue . "\n";
+    #     my $node = $nodes->item ($i);
+    #     my $gene_label = $node->getAttributeNode("TextLabel");
+    #     print $gene_label->getValue . "\n";
     # }
-    
-    # # Print doc file
-    # $doc->printToFile ("out.xml");
-    
-    # # Print to string
-    # print $doc->toString;
-    
-    # # Avoid memory leaks - cleanup circular references for garbage collection
-    # $doc->dispose;
-
 }
 
 # ---------------------------------------------------------------------------
@@ -330,7 +327,8 @@ Displays the transformed data. Verbose only.
 # ---------------------------------------------------------------------------
 sub show_output
 {
-    print "[Modified GPML]\n";
+    print "[Modified GPML Output]\n";
+    print $gpml_doc->toString;
     print "\n";
 }
 
@@ -341,8 +339,12 @@ Writes the transformed GPML doc out to the specified output file.
 # ---------------------------------------------------------------------------
 sub export_data
 {
-    print "Writing GPML to new output file...\n";
-    print "\n";
+    print "Writing GPML to output file...\n\n";
+
+    # ensures utf-8 encoding (for accent marks, etc.)
+    open my $out_file_handle, ">:utf8", "$output_file" or die $!;
+
+    $gpml_doc->print($out_file_handle);
 }
 
 =back
@@ -358,5 +360,6 @@ if ($verbose) { show_input; }
 swap_genes();
 if ($verbose) { show_output; }
 export_data;
+$gpml_doc->dispose; # cleanup
 exit;