lederhosen 1.3.4 → 1.3.5

data/lederhosen.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "lederhosen"
-  s.version = "1.3.4"
+  s.version = "1.3.5"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Austin G. Davis-Richardson"]
-  s.date = "2012-12-03"
+  s.date = "2012-12-04"
   s.description = "Various tools for OTU clustering"
   s.email = "harekrishna@gmail.com"
   s.executables = ["lederhosen"]

data/lib/lederhosen/no_tasks.rb

@@ -76,6 +76,13 @@ module Lederhosen
 
       # parse a taxonomic description using the
       # taxcollector format returning name at each level (genus, etc...)
+      #
+      # - If the species names contains the word '_bacterium', use the strain
+      # name as the species name:
+      #
+      #   Escherichia_bacterium -> Escherichia_bacterium_strain_X123
+      #   Arcanobacterium_phocae -> Arcanobacterium_phocae (no change)
+      #
       def parse_taxonomy_taxcollector(taxonomy)
 
         levels = { 'domain'  => 0,
@@ -85,7 +92,8 @@ module Lederhosen
                    'order'   => 3,
                    'family'  => 4,
                    'genus'   => 5,
-                   'species' => 6 }
+                   'species' => 6,
+                   'strain'  => 7 }
 
         names = Hash.new
 
@@ -94,6 +102,12 @@ module Lederhosen
           names[level] = name
         end
 
+        # check if species name contains the word 'bacterium'
+        # if so, replace it with the strain name
+        if names['species'] =~ /_bacterium/
+          names['species'] = names['strain']
+        end
+
         # keep original taxonomic description
         names['original'] = taxonomy
 

data/lib/lederhosen/version.rb

@@ -3,7 +3,7 @@ module Lederhosen
     MAJOR = 1
     MINOR = 3
     CODENAME = 'Dirndl' # changes for minor versions
-    PATCH = 4
+    PATCH = 5
 
     STRING = [MAJOR, MINOR, PATCH].join('.')
   end

data/spec/no_tasks_spec.rb

@@ -6,7 +6,7 @@ describe 'no_tasks' do
 
   let(:greengenes_taxonomies) { ['124 U55236.1 Methanobrevibacter thaueri str. CW k__Archaea; p__Euryarchaeota; c__Methanobacteria; o__Methanobacteriales; f__Methanobacteriaceae; g__Methanobrevibacter; Unclassified; otu_127',
                                  'k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Rahnella;s__' ]}
-  let(:taxcollector_taxonomies) { ['[0]Bacteria;[1]Actinobacteria;[2]Actinobacteria;[3]null;[4]null;[5]null;[6]bacterium_TH3;[7]bacterium_TH3;[8]bacterium_TH3|M79434|8'] } 
+  let(:taxcollector_taxonomies) { ['[0]Bacteria;[1]Actinobacteria;[2]Actinobacteria;[3]null;[4]null;[5]null;[6]bacterium_TH3;[7]bacterium_TH3;[8]bacterium_TH3|M79434|8'] }
 
   it '#detect_taxonomy_format should recognize GreenGenes' do
     greengenes_taxonomies.each do |greengenes_taxonomy|
@@ -29,7 +29,7 @@ describe 'no_tasks' do
       taxonomy = lederhosen.parse_taxonomy_taxcollector(taxcollector_taxonomy)
       taxonomy['original'].should == taxcollector_taxonomy
 
-      levels = %w{domain phylum class order family genus species kingdom original}
+      levels = %w{domain phylum class order family genus species kingdom original strain}
 
       taxonomy.keys.each do |v|
         levels.should include v
@@ -59,4 +59,10 @@ describe 'no_tasks' do
       lederhosen.parse_taxonomy(taxcollector_taxonomy).should_not be_nil
     end
   end
+
+  it '#parse_taxonomy_taxcollector should replace unclassified species names with strain name' do
+    t = '[0]Bacteria;[1]Actinobacteria;[2]Actinobacteria;[3]Actinomycetales;[4]test;[5]null;[6]Propionibacterineae_bacterium;[7]Propionibacterineae_bacterium_870BRRJ;[8]Propionibacterineae_bacterium_870BRRJ|genus'
+    tax = lederhosen.parse_taxonomy(t)
+    tax['species'].should == tax['strain']
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: lederhosen
 version: !ruby/object:Gem::Version
-  version: 1.3.4
+  version: 1.3.5
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-12-03 00:00:00.000000000 Z
+date: 2012-12-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: dna
@@ -160,7 +160,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2143498304454718392
+      hash: 502535867079393177
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements: