genomer-plugin-validate 0.0.1

data/features/annotations/missing_attributes.feature

@@ -0,0 +1,75 @@
+Feature: Validating annotation files for missing attributes
+  In order to submit genome annotations
+  A user can use the "annotation" command to detect missing ID, Name, product
+  to ensure that their annotation file contains no errors
+
+  @disable-bundler
+  Scenario: Validating an annotations file with a missing ID attribute
+    Given I successfully run `genomer init project`
+      And I cd to "project"
+      And I write to "assembly/scaffold.yml" with:
+        """
+        ---
+          - sequence:
+              source: contig1
+        """
+      And I write to "assembly/sequence.fna" with:
+        """
+        >contig1
+        AAAAATTTTTGGGGGCCCCC
+        """
+      And I write to "assembly/annotations.gff" with:
+        """
+        ##gff-version 3
+        contig1	.	gene	1	3	.	+	1	ID=gene1;Name=something
+        contig1	.	gene	4	6	.	+	1	Name=something
+        """
+      And I append to "Gemfile" with:
+        """
+        gem 'genomer-plugin-validate', :path => '../../../'
+        """
+     When I run `genomer validate annotations`
+     Then the exit status should be 0
+      And the output should contain:
+        """
+        Annotations found with missing ID attribute
+
+        """
+
+  @disable-bundler
+  Scenario: Validating an annotations file with a missing Name or product attributes
+    Given I successfully run `genomer init project`
+      And I cd to "project"
+      And I write to "assembly/scaffold.yml" with:
+        """
+        ---
+          - sequence:
+              source: contig1
+        """
+      And I write to "assembly/sequence.fna" with:
+        """
+        >contig1
+        AAAAATTTTTGGGGGCCCCC
+        """
+      And I write to "assembly/annotations.gff" with:
+        """
+        ##gff-version 3
+        contig1	.	gene	1	2	.	+	1	ID=gene1;Name=something
+        contig1	.	gene	3	4	.	+	1	ID=gene2;product=something
+        contig1	.	gene	5	6	.	+	1	ID=gene3;product=something;Name=else
+        contig1	.	gene	7	8	.	+	1	ID=gene4
+        """
+      And I append to "Gemfile" with:
+        """
+        gem 'genomer-plugin-validate', :path => '../../../'
+        """
+     When I run `genomer validate annotations`
+     Then the exit status should be 0
+      And the output should not contain "gene1"
+      And the output should not contain "gene2"
+      And the output should not contain "gene3"
+      And the output should contain:
+        """
+        No 'Name' or 'product' attribute for annotation 'gene4'
+        """
+

data/features/annotations/name.feature

@@ -0,0 +1,40 @@
+Feature: Validating annotation files for incorrect names
+  In order to submit genome annotations
+  A user can use the "annotation" command to detect uppercase gene names
+  to ensure that their annotation file contains none of these
+
+  @disable-bundler
+  Scenario: Validating an annotations file with an uppercase name attribute
+    Given I successfully run `genomer init project`
+      And I cd to "project"
+      And I write to "assembly/scaffold.yml" with:
+        """
+        ---
+          - sequence:
+              source: contig1
+        """
+      And I write to "assembly/sequence.fna" with:
+        """
+        >contig1
+        AAAAATTTTTGGGGGCCCCC
+        """
+      And I write to "assembly/annotations.gff" with:
+        """
+        ##gff-version 3
+	contig1	.	gene	1	3	.	+	1	Name=Uppercase;ID=1
+	contig1	.	gene	4	6	.	+	1	Name=lowercase;ID=2
+        """
+      And I append to "Gemfile" with:
+        """
+        gem 'genomer-plugin-validate', :path => '../../../'
+        """
+     When I run `genomer validate annotations`
+     Then the exit status should be 0
+      And the output should contain:
+        """
+        Illegal capitalised Name attribute 'Uppercase' for '1'
+        """
+      And the output should not contain:
+        """
+        Illegal capitalised Name attribute 'lowercase' for '2'
+        """

data/features/command-line-interface.feature

@@ -0,0 +1,36 @@
+Feature: The validator command line interface
+  In order to generate correct genomer builds
+  A user can use the "validator" plugin at the command line
+  to validate their genome build
+
+  @disable-bundler
+  Scenario: Running with just the 'validate' command
+    Given I successfully run `genomer init project`
+      And I cd to "project"
+      And I append to "Gemfile" with:
+        """
+        gem 'genomer-plugin-validate', :path => '../../../'
+        """
+     When I run `genomer validate`
+     Then the exit status should be 0
+      And the output should contain:
+        """
+        USAGE: genomer validate <GROUP>
+
+        Available validation groups:
+        """
+
+  @disable-bundler
+  Scenario: Running with an unknown validation group
+    Given I successfully run `genomer init project`
+      And I cd to "project"
+      And I append to "Gemfile" with:
+        """
+        gem 'genomer-plugin-validate', :path => '../../../'
+        """
+     When I run `genomer validate unknown`
+     Then the exit status should be 1
+      And the output should contain:
+        """
+        Error. Unknown validation group 'unknown'
+        """

data/features/support/env.rb

@@ -0,0 +1,13 @@
+require 'bundler'
+begin
+  Bundler.setup(:default, :development)
+rescue Bundler::BundlerError => e
+  $stderr.puts e.message
+  $stderr.puts "Run `bundle install` to install missing gems"
+  exit e.status_code
+end
+
+$LOAD_PATH.unshift(File.dirname(__FILE__) + '/../../lib')
+$LOAD_PATH.unshift(File.dirname(__FILE__) + '/../../spec')
+
+require 'aruba/cucumber'

data/genomer-plugin-validate.gemspec

@@ -0,0 +1,28 @@
+# -*- encoding: utf-8 -*-
+
+Gem::Specification.new do |s|
+  s.name        = "genomer-plugin-validate"
+  s.version     = File.read('VERSION')
+  s.authors     = ["Michael Barton"]
+  s.email       = ["mail@michaelbarton.me.uk"]
+  s.homepage    = ""
+  s.summary     = %q{Validate assembly files for errors}
+  s.description = %q{Test assembly files for common errors which may lead to incorrect assembly}
+
+  s.rubyforge_project = "genomer-plugin-validate"
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = ["lib"]
+
+  s.add_runtime_dependency "genomer",          ">= 0.0.4"
+  s.add_runtime_dependency "heredoc_unindent", "~> 1.1.0"
+
+  s.add_development_dependency "rake"
+  s.add_development_dependency "rspec",                   "~> 2.8.0"
+  s.add_development_dependency "scaffolder-test-helpers", "~> 0.4.1"
+  s.add_development_dependency "cucumber",                "~> 1.1.4"
+  s.add_development_dependency "aruba",                   "~> 0.4.11"
+  s.add_development_dependency "rr",                      "~> 1.0.4"
+end

data/lib/extensions/string.rb

@@ -0,0 +1,12 @@
+class String
+
+  # Taken from Rails
+  def underscore
+    self.gsub(/::/, '/').
+    gsub(/([A-Z]+)([A-Z][a-z])/,'\1_\2').
+    gsub(/([a-z\d])([A-Z])/,'\1_\2').
+    tr("-", "_").
+    downcase
+  end
+
+end

data/lib/genomer-plugin-validate.rb

@@ -0,0 +1,33 @@
+require "genomer"
+require "heredoc_unindent"
+
+class GenomerPluginValidate < Genomer::Plugin
+  require 'genomer-plugin-validate/validator'
+  require 'genomer-plugin-validate/group'
+
+  def run
+    name = arguments.shift
+    return self.class.help_message if name.nil?
+
+    group = Group.groups[name]
+    raise Genomer::Error, "Unknown validation group '#{name}'" if group.nil?
+
+    group.validators.map{|i| Validator.validators[i]}.map do |v|
+      v.new(arguments,flags).run
+    end.flatten * "\n"
+  end
+
+  def self.help_message
+    msg = <<-EOS.unindent
+      USAGE: genomer validate <GROUP>
+      
+      Available validation groups:
+    EOS
+    msg << Group.groups.map do |(k,v)|
+      str =  '  '
+      str << k.ljust(15)
+      str << v.description
+    end * "\n"
+  end
+
+end

data/lib/genomer-plugin-validate/group.rb

@@ -0,0 +1,17 @@
+module GenomerPluginValidate::Group
+
+  def self.load
+    path = File.join(File.dirname(__FILE__),'..','genomer-plugin-validate','group','*')
+    Dir[path].each do |i|
+      require i if i =~ /\.rb/
+    end
+  end
+
+  def self.groups
+    load
+    Hash[constants.map do |name|
+      [name.to_s.downcase,const_get(name)]
+    end]
+  end
+
+end

data/lib/genomer-plugin-validate/group/annotations.rb

@@ -0,0 +1,20 @@
+class GenomerPluginValidate::Group::Annotations
+
+  def self.description
+    "Validate GFF3 annotations file"
+  end
+
+  def self.validators
+    [
+      :duplicate_id,
+      :missing_id,
+      :no_name_or_product,
+      :gff3_attributes,
+      :view_attributes,
+      :duplicate_coordinates,
+      :uppercase_name,
+      :bad_product_field
+    ]
+  end
+
+end

data/lib/genomer-plugin-validate/validator.rb

@@ -0,0 +1,27 @@
+require 'extensions/string'
+
+module GenomerPluginValidate::Validator
+
+  def self.load
+    path = File.join(File.dirname(__FILE__),'..','genomer-plugin-validate','validator','*')
+    Dir[path].each do |i|
+      require i if i =~ /\.rb/
+    end
+  end
+
+  def self.validators
+    load
+    Hash[constants.map do |name|
+      [name.to_s.underscore.to_sym, const_get(name)]
+    end]
+  end
+
+  def annotations_by_attribute(attr)
+    annotations.inject(Hash.new{|h,k| h[k] = []}) do |hash,attn|
+      attr_value = attn.get_attribute(attr) ? attn.get_attribute(attr).to_s : nil
+      hash[attr_value] <<= attn
+      hash
+    end
+  end
+
+end

data/lib/genomer-plugin-validate/validator/bad_product_field.rb

@@ -0,0 +1,45 @@
+class GenomerPluginValidate::Validator::BadProductField < Genomer::Plugin
+  include GenomerPluginValidate::Validator
+
+  ERROR = "Bad product field for '%s:' "
+
+  def run
+    [
+      hypothetical_products,
+      domain_related_like_ending_products,
+      nterm_products,
+      all_caps_products
+    ].flatten
+  end
+
+  def products_matching(re)
+    annotations_by_attribute("product").
+      map{|(product,entries)| entries.map{|i| [i.id,product]}}.
+      flatten(1).
+      map{|(id,product)| [id, re.match(product)]}.
+      select{|(_,match)| match}.
+      map{|(id,match)| [id,match.to_a[1].downcase]}
+  end
+
+  def hypothetical_products
+    products_matching(/^([Hh]ypothetical)(?! protein)/).
+      map{|i| (ERROR + "start with 'putative' instead of '%s.'") % i}
+  end
+
+  def domain_related_like_ending_products
+    products_matching(/([Dd]omain|[Rr]elated|[Ll]ike).?$/).
+      map{|i| (ERROR + "products ending with '%s' are not allowed.") % i}
+  end
+
+  def nterm_products
+    products_matching(/(?!\B)([Nn][-\s][Tt]erm(inal)?)/).
+      map{|i| (ERROR + "'N-terminal' or variations are not allowed.") % i}
+  end
+
+  def all_caps_products
+    products_matching(/^([A-Z\s-]+)$/).
+      map(&:first).
+      map{|i| (ERROR + "all caps product fields are not allowed.") % i}
+  end
+
+end

data/lib/genomer-plugin-validate/validator/duplicate_coordinates.rb

@@ -0,0 +1,12 @@
+class GenomerPluginValidate::Validator::DuplicateCoordinates < Genomer::Plugin
+
+  def run
+    annotations.
+      group_by{|attn| [attn.start, attn.end].sort }.
+      select{|_,v| v.length > 1}.
+      map{|(coords,attns)| attns.map(&:id).sort }.
+      map{|attns| attns.map{|attns| "'#{attns}'"}.join(', ') }.
+      map{|attns| "Identical locations for #{attns}" }
+  end
+
+end

data/lib/genomer-plugin-validate/validator/duplicate_id.rb

@@ -0,0 +1,11 @@
+class GenomerPluginValidate::Validator::DuplicateID < Genomer::Plugin
+  include GenomerPluginValidate::Validator
+
+  def run
+    annotations_by_attribute('ID').
+      select{|_,v| v.length > 1}.
+      select{|k,_| ! k.nil? }.
+      map{|(id,_)| "Duplicate ID '#{id}'" }
+  end
+
+end

data/lib/genomer-plugin-validate/validator/gff3_attributes.rb

@@ -0,0 +1,16 @@
+class GenomerPluginValidate::Validator::Gff3Attributes < Genomer::Plugin
+
+  def valid_gff3_attributes
+    %w|ID Name Alias Parent Target Gap Derives_from Note
+      Dbxref Ontology_term Is_circular|
+  end
+
+  def run
+    annotations.
+      map{|attn| attn.attributes.map{|(k,v)| [k,attn] }}.
+      flatten(1).
+      select{|(term,_)| term =~ (/^[A-Z]/) }.
+      reject{|(term,_)| valid_gff3_attributes.include? term }.
+      map{|(term,attn)| "Illegal GFF3 attribute '#{term}' for '#{attn.id}'"}
+  end
+end

data/lib/genomer-plugin-validate/validator/missing_id.rb

@@ -0,0 +1,13 @@
+class GenomerPluginValidate::Validator::MissingID < Genomer::Plugin
+  include GenomerPluginValidate::Validator
+
+  def run
+    missing = annotations_by_attribute('ID').detect{|k,_| k.nil? }
+    if missing
+      ["Annotations found with missing ID attribute"]
+    else
+      []
+    end
+  end
+
+end

data/lib/genomer-plugin-validate/validator/no_name_or_product.rb

@@ -0,0 +1,13 @@
+class GenomerPluginValidate::Validator::NoNameOrProduct < Genomer::Plugin
+  include GenomerPluginValidate::Validator
+
+  def run
+    no_product = annotations_by_attribute('product')[nil].map(&:id)
+    no_name    = annotations_by_attribute('Name')[nil].map(&:id)
+
+    (no_name & no_product).map do |id|
+      "No 'Name' or 'product' attribute for annotation '#{id}'"
+    end
+  end
+
+end

data/lib/genomer-plugin-validate/validator/uppercase_name.rb

@@ -0,0 +1,13 @@
+class GenomerPluginValidate::Validator::UppercaseName < Genomer::Plugin
+  include GenomerPluginValidate::Validator
+
+  def run
+    annotations_by_attribute('Name').
+      select{|(name,_)| name =~ /^[A-Z]/}.
+      map{|(_,entry)| entry}.
+      flatten.
+      map{|i| [i.get_attribute('Name'),i.id]}.
+      map{|i| "Illegal capitalised Name attribute '%s' for '%s'" % i}
+  end
+
+end