update svn ignore

[function2gene.git] / bin / combine_results

diff --git a/bin/combine_results b/bin/combine_results
index 666fe7c81a70808c6bf01b69ed4952f0d6a5a156..49c1ca6da1b32aea859b0daa132d1648888c0ed5 100755 (executable)
--- a/bin/combine_results
+++ b/bin/combine_results
@@ -152,11 +152,37 @@ if (@{$options{keywords}}) {
 
 
 
+my %alias_reverse;
+
 for my $file_name (@ARGV) {
      my $file = new IO::File $file_name, 'r' or die "Unable to open file $file_name $!";
      while (<$file>) {
          next if /^"Name"/;
          my @gene = map {s/^\"//; s/\"$//; $_;} split /(?<=\")\,(?=\")/, $_;
+         # check to see if there's a different name we should be using
+         if (not exists $genes{$gene[NAME]}) {
+              # if the gene has a valid name, we do at least one test.
+              my $num_tested = $gene[NAME] ne 'NO NAME' ? 1 : 0;
+              my %candidates;
+              if ($gene[NAME] ne 'NO NAME' and exists $alias_reverse{$gene[NAME]} and $alias_reverse{$gene[NAME]} ne '') {
+                   $candidates{$alias_reverse{$gene[NAME]}}++;
+              }
+              else {
+                   for my $alias (grep {$_ !~ /^NO (ALIASES|NAME)$/} split(/; /, $gene[ALIAS])) {
+                        if (exists $alias_reverse{$alias} and $alias_reverse{$alias} ne '') {
+                             $candidates{$alias_reverse{$alias}}++;
+                        }
+                        $num_tested++;
+                   }
+              }
+              #print STDERR "Choosing $alias_reverse{$gene[NAME]} for $gene[NAME]\n";
+              for my $candidate (keys %candidates) {
+                   if ($candidates{$candidate} > ($num_tested/2)) {
+                        print STDERR "Choosing $candidate for '$gene[NAME]', as it matched $candidates{$candidate} of $num_tested tests\n";
+                        $gene[NAME] = $candidate;
+                   }
+              }
+         }
          $genes{$gene[NAME]}{name} = $gene[NAME];
          $genes{$gene[NAME]}{database}{$gene[DBNAME]}++;
          $genes{$gene[NAME]}{hits}++;
@@ -166,7 +192,19 @@ for my $file_name (@ARGV) {
          add_if_better($genes{$gene[NAME]},'description',$gene[DESCRIPTION]);
          add_if_better($genes{$gene[NAME]},'location',$gene[LOCATION]);
          add_unique_parts($genes{$gene[NAME]},'function',split(/; /, $gene[FUNCTION]));
-         add_unique_parts($genes{$gene[NAME]},'alias', split(/; /, $gene[ALIAS]));
+         my @aliases = grep {$_ ne 'NO ALIASES'} split(/; /, $gene[ALIAS]);
+         add_unique_parts($genes{$gene[NAME]},'alias', @aliases);
+         if ($gene[NAME] ne 'NO NAME') {
+              for my $alias (@aliases) {
+                   if (not exists $alias_reverse{$alias}) {
+                        $alias_reverse{$alias} = $gene[NAME];
+                   }
+                   elsif ($alias_reverse{$alias} ne $gene[NAME]) {
+                        print STDERR "Alias $alias for $gene[NAME] also points at $alias_reverse{$alias} [".join(',',@aliases).".]\n";
+                        $alias_reverse{$alias} = '';
+                   }
+              }
+         }
      }
 }
 
@@ -238,6 +276,11 @@ for my $keyword (keys %keyword_keyword) {
      $auto_weight{$keyword} = $results_by_this_keyword/$results_combined;
 }
 
+my $avg_weight = sum(values %auto_weight) / scalar keys %auto_weight;
+for my $keyword (keys %auto_weight) {
+     $auto_weight{$keyword} = $auto_weight{$keyword}/$avg_weight;
+}
+
 print {$results_fh} join(',',map {qq("$_")} @csv_fields),qq(\n);
 for my $gene (keys %genes) {
      $genes{$gene}{rzscore} = scalar grep {$_ !~ /\[/} keys %{$genes{$gene}{terms}};
@@ -280,6 +323,7 @@ sub add_unique_parts{
          $$hr{$key} = [@values];
      }
      else {
+         return unless @values;
          my %temp_hash;
          @temp_hash{@{$$hr{$key}}} = (1) x scalar @{$$hr{$key}};
          $temp_hash{@values} = (1) x scalar @values;
@@ -297,18 +341,48 @@ sub add_if_better{
      }
 }
 
+sub space_fill{
+     my ($value,$length,$right) = @_;
+     $right ||= 0;
+     if (length($value) > $length) {
+         $value =~ m/(.{$length})/;
+         return $1;
+     }
+     if (length($value) == $length) {
+         return $value
+     }
+     if ($right) {
+         return join('',
+                     ' ' x ($length - length($value)),
+                     $value,
+                    );
+     }
+     else {
+         return join('',
+                     $value,
+                     ' ' x ($length - length($value)),
+                    );
+     }
+}
 
+sub results_table_line {
+     my ($keyword,@fields) = @_;
+     return join( ' & ',
+                 space_fill($keyword,23),
+                 map {space_fill($_,11,1)} @fields
+               )."\n";
+}
+
+my @database_order = grep {lc($_) ne 'total'} keys %databases;
 if (defined $results_table_fh) {
-     our ($keyword,$weight,$autoweight,$gct,$hvt,$nct,$t) = ('Keyword','Weight','Autoweight','GeneCards','Harvester','NCBI','Total');
-     format RESULTS_TABLE =
-@<<<<<<<<<<<<<<<<<<<<<< & @>>>>>>>>>> & @>>>>>>>>>> & @>>>>>>>>>> & @>>>>>>>>>> & @>>>>>>>>>> & @>>>>>>>>>> \\
-$keyword,                 $weight,      $autoweight,  $gct,         $hvt,         $nct,         $t
-.
-     $results_table_fh->format_name('RESULTS_TABLE');
-     write $results_table_fh;
+     my $keyword;
+     print {$results_table_fh} results_table_line('Keyword','Weight','Autoweight',
+                                                 map {ucfirst($_)} @database_order,
+                                                 'Total',
+                                                );
 
      for $keyword (sort keys %terms) {
-         ($gct,$hvt,$nct,$t) =
+         my @fields =
               map {
                    if (not defined $_) {
                         '$-$';
@@ -317,15 +391,17 @@ $keyword,                 $weight,      $autoweight,  $gct,         $hvt,
                         $_->{unique} ||= 0;
                         "$_->{count} ($_->{unique})";
                    }
-              } @{$terms{$keyword}}{qw(genecard harvester ncbi total)};
-         $weight = $keyword_weight{$keyword} || 1;
-         $autoweight = $auto_weight{$keyword};
-         write $results_table_fh;
+              } @{$terms{$keyword}}{@database_order,'total'};
+         unshift @fields, $auto_weight{$keyword};
+         unshift @fields, $keyword_weight{$keyword} || 1;
+         print {$results_table_fh} results_table_line($keyword,
+                                                      @fields
+                                                     );
 
      }
-
      $keyword = 'Total';
-     ($gct,$hvt,$nct,$t) =
+     my @fields = ('','');
+     push @fields,
          map {
               if (not defined $_) {
                    '$-$';
@@ -334,11 +410,10 @@ $keyword,                 $weight,      $autoweight,  $gct,         $hvt,
                    $_->{unique} ||= 0;
                    "$_->{count} ($_->{unique})";
               }
-         } map {$_->{total}} @databases{qw(genecard harvester ncbi total)};
-     #($gct,$hvt,$nct,$t) = map {$_->{total}} @databases{qw(genecard harvester ncbi total)};
-     $weight = '';
-     $autoweight = '';
-     write $results_table_fh;
+         } map {$_->{total}} @databases{@database_order,'total'};
+     print {$results_table_fh} results_table_line($keyword,
+                                                 @fields
+                                                );
 }
 
 __END__