lederhosen 1.3.7 → 1.3.8

data/Gemfile

@@ -6,10 +6,12 @@ gem 'thor'
 
 group :test do
   gem 'rspec'
+  gem 'rspec-prof'
 end
 
 group :development do
   gem 'rdoc', '~> 3.12'
   gem 'bundler'
   gem 'jeweler'
+  gem 'ruby-prof'
 end

data/lederhosen.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "lederhosen"
-  s.version = "1.3.7"
+  s.version = "1.3.8"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Austin G. Davis-Richardson"]
-  s.date = "2012-12-06"
+  s.date = "2012-12-07"
   s.description = "Various tools for OTU clustering"
   s.email = "harekrishna@gmail.com"
   s.executables = ["lederhosen"]
@@ -64,6 +64,7 @@ Gem::Specification.new do |s|
       s.add_development_dependency(%q<rdoc>, ["~> 3.12"])
       s.add_development_dependency(%q<bundler>, [">= 0"])
       s.add_development_dependency(%q<jeweler>, [">= 0"])
+      s.add_development_dependency(%q<ruby-prof>, [">= 0"])
     else
       s.add_dependency(%q<dna>, ["= 0.0.12"])
       s.add_dependency(%q<progressbar>, [">= 0"])
@@ -71,6 +72,7 @@ Gem::Specification.new do |s|
       s.add_dependency(%q<rdoc>, ["~> 3.12"])
       s.add_dependency(%q<bundler>, [">= 0"])
       s.add_dependency(%q<jeweler>, [">= 0"])
+      s.add_dependency(%q<ruby-prof>, [">= 0"])
     end
   else
     s.add_dependency(%q<dna>, ["= 0.0.12"])
@@ -79,6 +81,7 @@ Gem::Specification.new do |s|
     s.add_dependency(%q<rdoc>, ["~> 3.12"])
     s.add_dependency(%q<bundler>, [">= 0"])
     s.add_dependency(%q<jeweler>, [">= 0"])
+    s.add_dependency(%q<ruby-prof>, [">= 0"])
   end
 end
 

data/lib/lederhosen/no_tasks.rb

@@ -1,5 +1,8 @@
 module Lederhosen
   class CLI
+
+    attr_accessor :taxonomy_format
+
     no_tasks do
 
       # parse a line of usearch prefix
@@ -18,7 +21,7 @@ module Lederhosen
         identity = str[3].to_f
 
         # parse taxonomic_description
-        taxonomies = parse_taxonomy(taxonomic_description)
+        taxonomies = parse_taxonomy(taxonomic_description) rescue { 'original' => str[9] }
 
         { :identity => identity }.merge(taxonomies)
       end
@@ -28,49 +31,57 @@ module Lederhosen
       #
       # - :taxcollector
       # - :greengenes
+      # - :qiime (subset of greengenes)
       #
       def detect_taxonomy_format(taxonomy)
         # taxcollector taxonomy starts with a open square bracked
         if taxonomy =~ /^\[/
           :taxcollector
-        else
+        elsif taxonomy =~ /^\d/
           :greengenes
+        else
+          :qiime
         end
       end
 
       def parse_taxonomy(taxonomy)
-        format = detect_taxonomy_format(taxonomy)
+        @taxonomy_format ||= detect_taxonomy_format(taxonomy)
 
-        case format
+        case @taxonomy_format
         when :greengenes
           parse_taxonomy_greengenes(taxonomy)
         when :taxcollector
           parse_taxonomy_taxcollector(taxonomy)
-        else
-          fail 'unknown format!'
+        when :qiime
+          parse_taxonomy_qiime(taxonomy)
+        else # return original string
+          { :original => taxonomy }
         end
       end
 
-      def parse_taxonomy_greengenes(taxonomy)
-
-        levels = { 'domain'  => /k__(\w*)/,
-                   'kingdom' => /k__(\w*)/,
-                   'phylum'  => /p__(\w*)/,
-                   'class'   => /c__(\w*)/,
-                   'order'   => /o__(\w*)/,
-                   'family'  => /f__(\w*)/,
-                   'genus'   => /g__(\w*)/,
-                   'species' => /s__(\w*)/
-                  }
+      def parse_taxonomy_qiime(taxonomy)
+        levels = %w{kingdom phylum class order family genus species}
+        match_data = taxonomy.match(/k__(\w*);p__(\w*);c__(\w*);o__(\w*);f__(\w*);g__(\w*);s__(\w*)/)
+        match_data = match_data[1..-1]
 
         names = Hash.new
-
-        levels.each_pair do |level, regexp|
-          names[level] = taxonomy.match(regexp)[1] rescue nil
-        end
+        # for some reason Hash[*levels.zip(match_data)] ain't working
+        levels.zip(match_data).each { |l, n| names[l] = n }
 
         names['original'] = taxonomy
+        names
+      end
+
+      def parse_taxonomy_greengenes(taxonomy)
+        levels = %w{kingdom phylum class order family genus species}
+        match_data = taxonomy.match(/k__(\w*); ?p__(\w*); ?c__(\w*); ?o__(\w*); ?f__(\w*); ?g__(\w*); ?(\w*);/)
+        match_data = match_data[1..-1]
+
+        names = Hash.new
+        # for some reason Hash[*levels.zip(match_data)] ain't working
+        levels.zip(match_data).each { |l, n| names[l] = n }
 
+        names['original'] = taxonomy
         names
       end
 
@@ -85,25 +96,21 @@ module Lederhosen
       #
       def parse_taxonomy_taxcollector(taxonomy)
 
-        levels = { 'domain'  => 0,
-                   'kingdom' => 0,
-                   'phylum'  => 1,
-                   'class'   => 2,
-                   'order'   => 3,
-                   'family'  => 4,
-                   'genus'   => 5,
-                   'species' => 6,
-                   'strain'  => 7 }
+        levels = %w{kingdom phylum class order family genus species strain}
 
-        names = Hash.new
+        match_data =
+          begin
+            taxonomy.match(/\[0\](.*);\[1\](.*);\[2\](.*);\[3\](.*);\[4\](.*);\[5\](.*);\[6\](.*);\[7\](.*);\[8\](.*)/)[1..-1]
+          rescue
+            $stderr.puts taxonomy.inspect
+            return nil
+          end
 
-        levels.each_pair do |level, num|
-          name = taxonomy.match(/\[#{num}\](\w*)[;\[]/)[1] rescue nil
-          names[level] = name
-        end
+        names = Hash.new
+        # for some reason Hash[*levels.zip(match_data)] ain't working
+        levels.zip(match_data).each { |l, n| names[l] = n }
 
         # 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

data/lib/lederhosen/tasks/otu_table.rb

@@ -39,6 +39,7 @@ module Lederhosen
       pbar = ProgressBar.new "loading", input.size
 
       # Load cluster table
+
       input.each do |input_file|
         pbar.inc
         File.open(input_file) do |handle|
@@ -50,11 +51,9 @@ module Lederhosen
                 if dat.nil?
                   'unclassified_reads'
                 else
-                  dat[level]
+                  dat[level] || 'unparsed_name'
                 end
 
-              name = 'unparsed_name' if name.nil?
-
               level_sample_cluster_count[level][input_file][name] += 1
               all_names[level] << name
             end

data/lib/lederhosen/version.rb

@@ -3,7 +3,7 @@ module Lederhosen
     MAJOR = 1
     MINOR = 3
     CODENAME = 'Dirndl' # changes for minor versions
-    PATCH = 7
+    PATCH = 8
 
     STRING = [MAJOR, MINOR, PATCH].join('.')
   end

data/lib/lederhosen.rb

@@ -1,8 +1,6 @@
 require 'rubygems'
-require 'thor'
-require 'progressbar'
-require 'zlib'
-require 'dna'
+require 'bundler'
+Bundler.require :default
 require 'set'
 
 Dir.glob(File.join(File.dirname(__FILE__), 'lederhosen', '*.rb')).each { |f| require f }

data/spec/no_tasks_spec.rb

@@ -1,12 +1,13 @@
 require 'spec_helper'
 
-lederhosen = Lederhosen::CLI.new
-
 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(:greengenes_taxonomies) { ['124 U55236.1 Methanobrevibacter thaueri str. CW k__Archaea; p__Euryarchaeota; c__Methanobacteria; o__Methanobacteriales; f__Methanobacteriaceae; g__Methanobrevibacter; Unclassified; otu_127']}
+  let(:qiime_taxonomies) { [ '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(:lederhosen) { Lederhosen::CLI.new }
+
+  it '#parse_usearch_line should parse a line of usearch output'
 
   it '#detect_taxonomy_format should recognize GreenGenes' do
     greengenes_taxonomies.each do |greengenes_taxonomy|
@@ -48,6 +49,17 @@ describe 'no_tasks' do
     end
   end
 
+  it '#parse_taxonomy_greengenes should parse qiime taxonomy' do
+    qiime_taxonomies.each do |qiime_taxonomy|
+      taxonomy = lederhosen.parse_taxonomy_qiime(qiime_taxonomy)
+      levels = %w{domain phylum class order family genus species kingdom original}
+
+      taxonomy.keys.each do |v|
+        levels.should include v
+      end
+    end
+  end
+
   it '#parse_taxonomy should automatically detect and parse greengenes taxonomy' do
     greengenes_taxonomies.each do |greengenes_taxonomy|
       lederhosen.parse_taxonomy(greengenes_taxonomy).should_not be_nil
@@ -65,4 +77,5 @@ describe 'no_tasks' do
     tax = lederhosen.parse_taxonomy(t)
     tax['species'].should == tax['strain']
   end
+
 end

data/spec/spec_helper.rb

@@ -1,6 +1,8 @@
 $:.unshift File.join(File.dirname(__FILE__), '..')
+
 require 'lederhosen'
-require 'rspec'
+
+Bundler.require :test, :development
 
 $test_dir = ENV['TEST_DIR'] || "/tmp/lederhosen_test_#{(0...8).map{65.+(rand(25)).chr}.join}/"
 `mkdir -p #{$test_dir}`

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: lederhosen
 version: !ruby/object:Gem::Version
-  version: 1.3.7
+  version: 1.3.8
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-12-06 00:00:00.000000000 Z
+date: 2012-12-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: dna
@@ -107,6 +107,22 @@ dependencies:
     - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: ruby-prof
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Various tools for OTU clustering
 email: harekrishna@gmail.com
 executables:
@@ -160,7 +176,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 3956169746555075526
+      hash: 1569227273029021963
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements: