From 0738a7b4b8b35d7470e4e75591845cf06ccca19e Mon Sep 17 00:00:00 2001
From: Brian Osborne <briano@bioteam.net>
Date: Mon, 21 Sep 2015 10:26:14 -0400
Subject: [PATCH] Add 'desc' to table format

---
 Bio/SeqIO/table.pm   | 23 +++++++++++++++++++----
 t/SeqIO/table.t      | 39 ++++++++++++++++++++++++++++++++++++---
 t/data/test-1.tab    |  3 +++
 t/data/test-1.tab.gb | 16 ++++++++++++++++
 4 files changed, 74 insertions(+), 7 deletions(-)
 create mode 100644 t/data/test-1.tab
 create mode 100644 t/data/test-1.tab.gb

diff --git a/Bio/SeqIO/table.pm b/Bio/SeqIO/table.pm
index 33ce7fc2d..64ff8825a 100644
--- a/Bio/SeqIO/table.pm
+++ b/Bio/SeqIO/table.pm
@@ -23,12 +23,11 @@ Bio::SeqIO::table - sequence input/output stream from a delimited table
 
 =head1 SYNOPSIS
 
-  # It is probably best not to use this object directly, but
-  # rather go through the SeqIO handler system. Go:
+  # Do not to use this object directly, use Bio::SeqIO, for example:
 
-  $stream = Bio::SeqIO->new(-file => $filename, -format => 'table');
+  $in = Bio::SeqIO->new(-file => $filename, -format => 'table');
 
-  while ( my $seq = $stream->next_seq() ) {
+  while ( my $seq = $in->next_seq() ) {
 	# do something with $seq
   }
 
@@ -126,6 +125,8 @@ use base qw(Bio::SeqIO);
                     containing the accession number of the sequence
  -seq               The one-based index of the column containing
                     the sequence string of the sequence
+ -desc              The one-based index of the column containing
+                    the description of the sequence
  -species           The one-based index of the column containing the
                     species for the sequence record; if not a
                     number, will be used as the static species
@@ -178,6 +179,7 @@ sub _initialize {
         $header,
         $delim,
         $display_id,
+        $desc,
         $accnr,
         $seq,
         $taxon,
@@ -188,6 +190,7 @@ sub _initialize {
                                   HEADER
                                   DELIM
                                   DISPLAY_ID
+                                  DESC
                                   ACCESSION_NUMBER
                                   SEQ
                                   SPECIES
@@ -209,6 +212,7 @@ sub _initialize {
     $attrs->{-display_id} = $display_id if defined($display_id);
     $attrs->{-accession_number} = $accnr if defined($accnr);
     $attrs->{-seq} = $seq if defined($seq);
+    $attrs->{-desc} = $desc if defined($desc);
     if (defined($taxon)) {
         if (ref($taxon) || ($taxon =~ /^\d+$/)) {
             # either a static object, or a column reference
@@ -582,6 +586,17 @@ sub trim_values{
     return $self->{'trim_values'};
 }
 
+=head2 write_seq
+
+ Title: write_seq
+ Usage: write_seq() is not implemented for table format output.
+
+=cut
+
+sub write_seq {
+    shift->throw("write_seq() not implemented for 'table' format");
+}
+
 =head1 Internal methods
 
 All methods with a leading underscore are not meant to be part of the
diff --git a/t/SeqIO/table.t b/t/SeqIO/table.t
index 0cb17fa90..5a15eec12 100644
--- a/t/SeqIO/table.t
+++ b/t/SeqIO/table.t
@@ -1,5 +1,4 @@
 # -*-Perl-*- Test Harness script for Bioperl
-# $Id$
 
 use strict;
 
@@ -7,7 +6,7 @@ BEGIN {
     use lib '.';
     use Bio::Root::Test;
     
-    test_begin(-tests => 450,
+    test_begin(-tests => 464,
 			   -requires_module => 'IO::Scalar');
 	
 	use_ok('Bio::Tools::CodonTable');
@@ -78,9 +77,43 @@ ok $seqin = Bio::SeqIO->new(-file => test_input_file("test.tsv"),
                          -trim => 1);
 run_tests([@names],[@accs],[4,4,4,4,4,5,5,5,5,5],[@psg],[@rs]);
 
+# Tests to check that 'description' is read from 'table' format
+ok $seqin = Bio::SeqIO->new(
+    -file   => test_input_file("test-1.tab"),
+    -format => 'table',
+    -header => 1,
+    -display_id => 1, 
+    -accession_number => 1, 
+    -seq => 3, 
+    -desc => 2
+);
+ok($seqin);
+my $seq = $seqin->next_seq;
+ok($seq);
+is( $seq->desc, 'd1');
+is( $seq->display_id, 'n1');
+is( $seq->seq, 'aaaa');
+$seq = $seqin->next_seq;
+ok($seq);
+is( $seq->desc, 'd2');
+is( $seq->display_id, 'n2');
+is( $seq->seq, 'tttt');
+
 $seqin->close();
 
-# need Spreadsheet::ParseExcel installed for testing Excel format
+# Tests to check that we can _not_ write to 'table' format
+ok $seqin = Bio::SeqIO->new(
+    -file   => test_input_file("test-1.tab.gb"),
+    -format => 'genbank'
+);
+ok($seqin);
+my $seq = $seqin->next_seq;
+ok($seq);
+my $tmpfile = test_output_file();
+my $seqout = Bio::SeqIO->new( -format => 'table', -file => ">$tmpfile" );
+dies_ok { $seqout->write_seq($seq) } "write_seq() not implemented";
+
+# Need Spreadsheet::ParseExcel installed for testing Excel format
 SKIP: {
 	test_skip(-tests => 112, -requires_module => 'Spreadsheet::ParseExcel');
 
diff --git a/t/data/test-1.tab b/t/data/test-1.tab
new file mode 100644
index 000000000..2830d331a
--- /dev/null
+++ b/t/data/test-1.tab
@@ -0,0 +1,3 @@
+N	D	S
+n1	d1	aaaa
+n2	d2	tttt
diff --git a/t/data/test-1.tab.gb b/t/data/test-1.tab.gb
new file mode 100644
index 000000000..aa2bfcfdc
--- /dev/null
+++ b/t/data/test-1.tab.gb
@@ -0,0 +1,16 @@
+LOCUS       n1                         4 bp    dna     linear   UNK 
+DEFINITION  d1
+ACCESSION   n1
+KEYWORDS    .
+FEATURES             Location/Qualifiers
+ORIGIN      
+        1 aaaa
+//
+LOCUS       n2                         4 bp    dna     linear   UNK 
+DEFINITION  d2
+ACCESSION   n2
+KEYWORDS    .
+FEATURES             Location/Qualifiers
+ORIGIN      
+        1 tttt
+//
-- 
2.11.4.GIT