genomer-plugin-summary 0.0.2 → 0.0.3

data/VERSION

@@ -1 +1 @@
-0.0.2
+0.0.3

data/features/scaffold.feature

@@ -0,0 +1,122 @@
+Feature: Producing a summary of the scaffold
+  In order to have an overview of the scaffold
+  A user can use the "scaffod" command
+  to generate the a tabular output of the scaffold
+
+  Scenario: A scaffold with a single sequence
+    Given I create a new genomer project
+      And I write to "assembly/scaffold.yml" with:
+      """
+      ---
+        -
+          sequence:
+            source: contig0001
+      """
+      And I write to "assembly/sequence.fna" with:
+      """
+      >contig0001
+      ATGC
+      """
+     When I run `genomer summary scaffold`
+     Then the exit status should be 0
+      And the output should contain:
+      """
+      +--------------+-----------+
+      |         Scaffold         |
+      +--------------+-----------+
+      | Contigs (#)  |         1 |
+      | Gaps (#)     |         0 |
+      +--------------+-----------+
+      | Size (bp)    |         4 |
+      | Contigs (bp) |         4 |
+      | Gaps (bp)    |         0 |
+      +--------------+-----------+
+      | G+C (%)      |     50.00 |
+      | Contigs (%)  |    100.00 |
+      | Gaps (%)     |      0.00 |
+      +--------------+-----------+
+
+      """
+
+  Scenario: A scaffold with a two sequences
+    Given I create a new genomer project
+      And I write to "assembly/scaffold.yml" with:
+      """
+      ---
+        -
+          sequence:
+            source: contig0001
+        -
+          sequence:
+            source: contig0002
+      """
+      And I write to "assembly/sequence.fna" with:
+      """
+      >contig0001
+      ATGC
+      >contig0002
+      GGGC
+      """
+     When I run `genomer summary scaffold`
+     Then the exit status should be 0
+      And the output should contain:
+      """
+      +--------------+-----------+
+      |         Scaffold         |
+      +--------------+-----------+
+      | Contigs (#)  |         2 |
+      | Gaps (#)     |         0 |
+      +--------------+-----------+
+      | Size (bp)    |         8 |
+      | Contigs (bp) |         8 |
+      | Gaps (bp)    |         0 |
+      +--------------+-----------+
+      | G+C (%)      |     75.00 |
+      | Contigs (%)  |    100.00 |
+      | Gaps (%)     |      0.00 |
+      +--------------+-----------+
+
+      """
+
+  Scenario: A scaffold with a two sequences and a gap
+    Given I create a new genomer project
+      And I write to "assembly/scaffold.yml" with:
+      """
+      ---
+        -
+          sequence:
+            source: contig0001
+        -
+          unresolved:
+            length: 5
+        -
+          sequence:
+            source: contig0002
+      """
+      And I write to "assembly/sequence.fna" with:
+      """
+      >contig0001
+      ATGC
+      >contig0002
+      GGGC
+      """
+     When I run `genomer summary scaffold`
+     Then the exit status should be 0
+      And the output should contain:
+      """
+      +--------------+-----------+
+      |         Scaffold         |
+      +--------------+-----------+
+      | Contigs (#)  |         2 |
+      | Gaps (#)     |         1 |
+      +--------------+-----------+
+      | Size (bp)    |        13 |
+      | Contigs (bp) |         8 |
+      | Gaps (bp)    |         5 |
+      +--------------+-----------+
+      | G+C (%)      |     75.00 |
+      | Contigs (%)  |     61.54 |
+      | Gaps (%)     |     38.46 |
+      +--------------+-----------+
+
+      """

data/genomer-plugin-summary.gemspec

@@ -15,6 +15,7 @@ Gem::Specification.new do |gem|
   gem.version       = File.read 'VERSION'
 
   gem.add_dependency "genomer",        ">= 0.0.4"
+  gem.add_dependency "lazing",         ">= 0.1.1"
   gem.add_dependency "terminal-table", "~> 1.4.5"
 
   gem.add_development_dependency 'rake', '~> 0.9.0'

data/lib/genomer-plugin-summary/metrics.rb

@@ -0,0 +1,41 @@
+require 'genomer'
+require 'lazing'
+
+module GenomerPluginSummary::Metrics
+
+  ALL = :all
+
+  def gc_content(type,scfd)
+    gc   = enumerator_for(type,scfd).mapping{|i|   gc(i)}.inject(:+) || 0.0
+    atgc = enumerator_for(type,scfd).mapping{|i| atgc(i)}.inject(:+) || 0.0
+    gc / atgc * 100
+  end
+
+  def count(type,scfd)
+    enumerator_for(type,scfd).count
+  end
+
+  def percent(type,scfd)
+    length(type,scfd) / length(ALL,scfd).to_f * 100
+  end
+
+  def length(type,scfd)
+    enumerator_for(type,scfd).
+      mapping(&:sequence).
+      mapping(&:length).
+      inject(:+) || 0
+  end
+
+  def gc(entry)
+    entry.sequence.gsub(/[^GCgc]/,'').length.to_f
+  end
+
+  def atgc(entry)
+    entry.sequence.gsub(/[^ATGCatgc]/,'').length.to_f
+  end
+
+  def enumerator_for(type,scaffold)
+    scaffold.selecting{|i| [ALL,i.entry_type].include? type }
+  end
+
+end

data/lib/genomer-plugin-summary/scaffold.rb

@@ -0,0 +1,56 @@
+require 'genomer'
+require 'genomer-plugin-summary/metrics'
+require 'terminal-table'
+
+class GenomerPluginSummary::Scaffold < Genomer::Plugin
+  include GenomerPluginSummary::Metrics
+
+  LAYOUT = [
+    {:name => 'Contigs (#)',  :entry_type => :sequence,   :method => :count},
+    {:name => 'Gaps (#)',     :entry_type => :unresolved, :method => :count},
+    :separator,
+    {:name => 'Size (bp)',    :entry_type => :all,        :method => :length},
+    {:name => 'Contigs (bp)', :entry_type => :sequence,   :method => :length},
+    {:name => 'Gaps (bp)',    :entry_type => :unresolved, :method => :length},
+    :separator,
+    {:name => 'G+C (%)',      :entry_type => :all,        :method => :gc_content},
+    {:name => 'Contigs (%)',  :entry_type => :sequence,   :method => :percent},
+    {:name => 'Gaps (%)',     :entry_type => :unresolved, :method => :percent}
+  ]
+
+  def run
+    tabulate calculate_metrics(LAYOUT, scaffold)
+  end
+
+  def title
+    'Scaffold'
+  end
+
+  def tabulate(data)
+    table = Terminal::Table.new(:title => title) do |t|
+      data.each do |(k,v)|
+        t << if k == :separator
+               :separator
+             else
+               v = sprintf('%#.2f',v) if v.class == Float
+               [k.ljust(12),v.to_s.rjust(9)]
+             end
+      end
+    end
+
+    table.align_column 0, :left
+    table.align_column 1, :right
+    table.to_s
+  end
+
+  def calculate_metrics(specs,scaffold)
+    specs.map do |spec|
+      if spec == :separator
+        spec
+      else
+        [spec[:name], send(spec[:method],spec[:entry_type],scaffold)]
+      end
+    end
+  end
+
+end

data/lib/genomer-plugin-summary/sequences.rb

@@ -1,7 +1,9 @@
 require 'genomer'
+require 'genomer-plugin-summary/metrics'
 require 'terminal-table'
 
 class GenomerPluginSummary::Sequences < Genomer::Plugin
+  include GenomerPluginSummary::Metrics
 
   def run
     sequences = calculate(scaffold)
@@ -51,22 +53,22 @@ class GenomerPluginSummary::Sequences < Genomer::Plugin
   end
 
   def calculate(scaffold)
-    total_length = scaffold.map(&:sequence).join.length.to_f
+    total_length   = length(:all,scaffold).to_f
+    running_length = 0
 
-    length = 0
     scaffold.map do |entry|
       i = nil
       if entry.entry_type != :unresolved
         entry_length = entry.sequence.length
         i = { :sequence => entry.source,
-              :start    => length + 1, 
-              :end      => length + entry_length,
+              :start    => running_length + 1,
+              :end      => running_length + entry_length,
               :size     => entry_length,
               :percent  => entry_length / total_length * 100,
-              :gc       => gc_content(entry.sequence) }
+              :gc       => gc(entry) / atgc(entry) * 100 }
       end
         
-      length += entry.sequence.length
+      running_length += entry.sequence.length
       i
     end.compact
   end
@@ -87,9 +89,4 @@ class GenomerPluginSummary::Sequences < Genomer::Plugin
     totals
   end
 
-  def gc_content(sequence)
-    nucleotides = sequence.gsub(/[^ATGCatgc]/,'')
-    nucleotides.gsub(/[^GCgc]/,'').length.to_f / nucleotides.length * 100
-  end
-
 end

data/spec/genomer-plugin-summary_spec/metrics_spec.rb

@@ -0,0 +1,255 @@
+require 'spec_helper'
+require 'genomer-plugin-summary/metrics'
+
+describe GenomerPluginSummary::Metrics do
+
+  let(:metric) do
+   o = Object.new
+   o.extend described_class
+   o
+  end
+
+  describe "#gc_content" do
+
+    subject do
+      metric.gc_content entry_type, scaffold
+    end
+
+    context "an empty scaffold" do
+      let(:scaffold){ [] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should be_nan }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == be_nan }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == be_nan }
+      end
+
+    end
+
+    context "a single contig scaffold" do
+      let(:scaffold){ [sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 50.0 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should be_nan }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 50.0 }
+      end
+
+    end
+
+    context "a mixed scaffold" do
+      let(:scaffold){ [sequence('ATGC'),unresolved('NNNN'),sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 50.0 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == be_nan }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 50.0 }
+      end
+
+    end
+
+  end
+  describe "#count" do
+
+    subject do
+      metric.count entry_type, scaffold
+    end
+
+    context "an empty scaffold" do
+      let(:scaffold){ [] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 0 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 0 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 0 }
+      end
+
+    end
+
+    context "a single contig scaffold" do
+      let(:scaffold){ [sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 1 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 0 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 1 }
+      end
+
+    end
+
+    context "a mixed scaffold" do
+      let(:scaffold){ [sequence('ATGC'),unresolved('NNNN'),sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 2 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 1 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 3 }
+      end
+
+    end
+
+  end
+
+  describe "#length" do
+
+    subject do
+      metric.length entry_type, scaffold
+    end
+
+    context "an empty scaffold" do
+      let(:scaffold){ [] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 0 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 0 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 0 }
+      end
+
+    end
+
+    context "a single contig scaffold" do
+      let(:scaffold){ [sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 4 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 0 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 4 }
+      end
+    end
+
+    context "a mixed scaffold" do
+      let(:scaffold){ [sequence('ATGC'),unresolved('NNNN'),sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 8 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 4 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 12 }
+      end
+    end
+
+  end
+
+  describe "#percent" do
+
+    subject do
+      metric.percent entry_type, scaffold
+    end
+
+    context "an empty scaffold" do
+      let(:scaffold){ [] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should be_nan }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == be_nan }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == be_nan }
+      end
+
+    end
+
+    context "a single contig scaffold" do
+      let(:scaffold){ [sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 100.0 }
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 0.0 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 100.0 }
+      end
+    end
+
+    context "a mixed scaffold" do
+      let(:scaffold){ [sequence('ATGC'),unresolved('NNNN'),sequence('ATGC')] }
+
+      context "contigs" do
+        let(:entry_type){ :sequence }
+        it{ should == 8 / 12.0 * 100}
+      end
+      context "gaps" do
+        let(:entry_type){ :unresolved }
+        it{ should == 4 / 12.0 * 100 }
+      end
+      context "everything" do
+        let(:entry_type){ :all }
+        it{ should == 100.0 }
+      end
+    end
+
+  end
+
+end

data/spec/genomer-plugin-summary_spec/scaffold_spec.rb

@@ -0,0 +1,76 @@
+require 'spec_helper'
+require 'genomer-plugin-summary/scaffold'
+
+describe GenomerPluginSummary::Scaffold do
+
+  describe "#tabulate" do
+
+    subject do
+      described_class.new([],{}).tabulate(data) + "\n"
+    end
+
+    context "passed table data" do
+
+      let(:data) do
+        [['Contigs (#)',1.0],
+          :separator,
+         ['Gaps (#)',0]]
+      end
+
+      it do
+        should ==<<-EOS.unindent!
+      +--------------+-----------+
+      |         Scaffold         |
+      +--------------+-----------+
+      | Contigs (#)  |      1.00 |
+      +--------------+-----------+
+      | Gaps (#)     |         0 |
+      +--------------+-----------+
+        EOS
+      end
+    end
+  end
+
+  describe "#calculate_metrics" do
+
+    subject do
+      described_class.new([],{}).calculate_metrics(specs,scaffold)
+    end
+
+    context "should calculate a single metrics for the scaffold" do
+
+      let(:scaffold) do
+        [sequence('ATGC')]
+      end
+
+      let(:specs) do
+        [{:name => 'Contigs (%)',  :entry_type => :sequence,   :method => :percent}]
+      end
+
+      it do
+        should == [['Contigs (%)',100.0]]
+      end
+    end
+
+    context "should calculate a single metrics with separators" do
+
+      let(:scaffold) do
+        [sequence('ATGC')]
+      end
+
+      let(:specs) do
+        [:separator,
+         {:name => 'Contigs (%)',  :entry_type => :sequence,   :method => :percent}]
+      end
+
+      it do
+        should == [
+          :separator,
+          ['Contigs (%)',100.0]
+        ]
+      end
+    end
+
+  end
+
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: genomer-plugin-summary
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.0.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-10-12 00:00:00.000000000 Z
+date: 2012-10-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: genomer
@@ -27,6 +27,22 @@ dependencies:
     - - ! '>='
       - !ruby/object:Gem::Version
         version: 0.0.4
+- !ruby/object:Gem::Dependency
+  name: lazing
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.1.1
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.1.1
 - !ruby/object:Gem::Dependency
   name: terminal-table
   requirement: !ruby/object:Gem::Requirement
@@ -153,17 +169,22 @@ files:
 - Rakefile
 - VERSION
 - features/gaps.feature
+- features/scaffold.feature
 - features/sequences.feature
 - features/support/env.rb
 - features/support/genomer_steps.rb
 - genomer-plugin-summary.gemspec
 - lib/genomer-plugin-summary.rb
 - lib/genomer-plugin-summary/gaps.rb
+- lib/genomer-plugin-summary/metrics.rb
+- lib/genomer-plugin-summary/scaffold.rb
 - lib/genomer-plugin-summary/sequences.rb
 - man/genomer-summary-gaps.ronn
 - man/genomer-summary.ronn
 - spec/genomer-plugin-summary_spec.rb
 - spec/genomer-plugin-summary_spec/gaps_spec.rb
+- spec/genomer-plugin-summary_spec/metrics_spec.rb
+- spec/genomer-plugin-summary_spec/scaffold_spec.rb
 - spec/genomer-plugin-summary_spec/sequences_spec.rb
 - spec/spec_helper.rb
 homepage: ''
@@ -180,7 +201,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 1551182150525719363
+      hash: 4220695143332306969
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -189,7 +210,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 1551182150525719363
+      hash: 4220695143332306969
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.23
@@ -198,10 +219,13 @@ specification_version: 3
 summary: Generates reports on the status of the genomer project
 test_files:
 - features/gaps.feature
+- features/scaffold.feature
 - features/sequences.feature
 - features/support/env.rb
 - features/support/genomer_steps.rb
 - spec/genomer-plugin-summary_spec.rb
 - spec/genomer-plugin-summary_spec/gaps_spec.rb
+- spec/genomer-plugin-summary_spec/metrics_spec.rb
+- spec/genomer-plugin-summary_spec/scaffold_spec.rb
 - spec/genomer-plugin-summary_spec/sequences_spec.rb
 - spec/spec_helper.rb