Bio::EMBLDB::Common#references: don't drop RG (consortium name)

RG lines (reference group / consortium name, e.g. "International
Human Genome Sequencing Consortium" and "Genome Reference
Consortium", both real examples from human GRCh38 chromosome 1,
accession CM000663) are common in large-scale genome assembly
submissions, and increasingly so since 2013. #ref already captured
RG into the raw reference Hash, but #references had no 'RG' branch,
so the Bio::Reference built from it silently lost all authorship
information whenever a reference was attributed to a consortium
alone (authors ended up as an empty Array).

Append the RG value to hash['authors'] (which 'RA' -- processed
first, per the Hash key order in #ref -- has already set to an
Array) so the consortium name is preserved instead of discarded.

Author: ngoto

lib/bio/db/embl/common.rb

@@ -311,6 +311,16 @@ def references
               tag, xref = item.split(/\; /).map {|i| i.strip.sub(/\.\z/, '') }
               hash[ tag.downcase ]  = xref
             }
+          when 'RG'
+            # RG (reference group / consortium name), e.g.
+            # "International Human Genome Sequencing Consortium",
+            # common in large-scale genome assembly submissions.
+            # Note that 'RA' is always processed before 'RG' (see the
+            # 'raw' Hash key order in #ref), so hash['authors'] is
+            # already set to an Array at this point.
+            unless value.to_s.strip.empty?
+              hash['authors'].push value
+            end
           end
         }
         Reference.new(hash)

test/unit/bio/db/embl/test_embl_new_part.rb

@@ -118,4 +118,46 @@ def test_co_absent
       assert_equal('', @obj.co)
     end
   end # class TestEMBL_CO_line_absent
+
+  class TestEMBL_RG_line < Test::Unit::TestCase
+    def setup
+      # Real example pattern (Homo sapiens chromosome 1, GRCh38,
+      # accession CM000663): a reference authored solely by a
+      # consortium (no RA line at all), and one with both individual
+      # authors (RA) and a consortium name (RG).
+      text = <<~THE_END_OF_THE_TEXT
+        ID   AB123456; SV 1; linear; genomic DNA; STD; PRO; 4 BP.
+        XX
+        AC   AB123456;
+        XX
+        RN   [1]
+        RA   Doe J., Roe R.;
+        RG   Example Genome Consortium
+        RT   ;
+        RL   Submitted (01-JAN-2020) to the INSDC.
+        XX
+        RN   [2]
+        RG   Example Genome Consortium
+        RT   ;
+        RL   Submitted (01-JAN-2020) to the INSDC.
+        XX
+        SQ   Sequence 4 BP; 1 A; 1 C; 1 G; 1 T; 0 other;
+             acgt                                                               4
+        //
+      THE_END_OF_THE_TEXT
+
+      @obj = Bio::EMBL.new(text)
+    end
+
+    def test_references_authors_include_consortium_alongside_individuals
+      # Regression test: RG (reference group/consortium name) used to
+      # be silently dropped by Common#references.
+      expected = ['Doe, J.', 'Roe, R.', 'Example Genome Consortium']
+      assert_equal(expected, @obj.references[0].authors)
+    end
+
+    def test_references_authors_is_consortium_only_when_no_ra
+      assert_equal(['Example Genome Consortium'], @obj.references[1].authors)
+    end
+  end # class TestEMBL_RG_line
 end # module Bio