unipept 1.1.0 → 1.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e4a1d5d01af5e00b706fe387a78a1a661d11a0ed
-  data.tar.gz: 22064ff35a1789a2cbc48fca2c4f4ec967c3d3a0
+  metadata.gz: f9e7aed1832989eddf44ce10e5ae42dc4976bf31
+  data.tar.gz: d23fd6021eeb04bc4c6cb9b077f0fca978422f38
 SHA512:
-  metadata.gz: e880e9f0737b91efae32a9724cb20ded630af7ad6bc3853af3b3993ebafea024d957f66d9fb14315f2670ade638dc7d708cafb92e54d540906ad1539a529c334
-  data.tar.gz: f7cd476ef08424f6a840033a33a025c70ce6eebeda8e71ea454c89246696678cb5049460bdc32904b54b72a447778d5b4aae01f134ac375153560bc5e9b1e684
+  metadata.gz: 6eeb2ca9e7dbbc13a0dd2c99a8aa3284cbca52250173060e4615f2b69d5d9b1cd27565c5cebc81275234d8f65542c07f4795f3d1b7bff71d89bb7b2bef9334f8
+  data.tar.gz: 6e79d4bf72af62e0c90c67e8f63a1eaa647850dfb616d225a2521a237274153e96887007f580bf83f5eab3112112f65bb2d6fdbde15a4630044bf3f63f2e6b04

data/.travis.yml

@@ -1,7 +1,8 @@
 language: ruby
 rvm:
-  - 2.2.2
-  - 2.1.6
+  - 2.3.1
+  - 2.2.5
+  - 2.1.9
   - 2.0.0
   - 1.9.3
 #  - jruby-19mode

data/VERSION

@@ -1 +1 @@
-1.1.0
+1.1.1

data/lib/commands/prot2pept.rb

@@ -20,6 +20,7 @@ module Unipept::Commands
       end
       run do |opts, _args, _cmd|
         pattern = opts.fetch(:pattern, '([KR])([^P])')
+        pattern = Regexp.compile(pattern)
 
         # decide if we have FASTA input
         first_char = $stdin.getc
@@ -33,7 +34,7 @@ module Unipept::Commands
               protein = ''
               puts line
             else
-              protein += line.chomp
+              protein << line.chomp
             end
           end
           puts Prot2pept.split(protein, pattern)
@@ -46,7 +47,7 @@ module Unipept::Commands
     end
 
     def self.split(protein, pattern)
-      protein.gsub(/\*/, "\n").gsub(/#{pattern}/, "\\1\n\\2").gsub(/#{pattern}/, "\\1\n\\2").split("\n").reject(&:empty?)
+      protein.tr('*', "\n").gsub(pattern, "\\1\n\\2").gsub(pattern, "\\1\n\\2").split("\n").reject(&:empty?)
     end
 
     # Invokes the uniprot command-line tool with the given arguments.

data/lib/commands/unipept/api_runner.rb

@@ -60,6 +60,11 @@ module Unipept
       end
     end
 
+    # returns the required fields to do any mapping
+    def required_fields
+      []
+    end
+
     # Returns a new batch_iterator based on the batch_size
     def batch_iterator
       Unipept::BatchIterator.new(batch_size)
@@ -79,7 +84,10 @@ module Unipept
 
     # Returns an array of regular expressions containing all the selected fields
     def selected_fields
-      @selected_fields ||= [*options[:select]].map { |f| f.split(',') }.flatten.map { |f| glob_to_regex(f) }
+      return @selected_fields unless @selected_fields.nil?
+      fields = [*options[:select]].map { |f| f.split(',') }.flatten
+      fields.concat(required_fields) unless fields.empty?
+      @selected_fields = fields.map { |f| glob_to_regex(f) }
     end
 
     # Returns a formatter, based on the format specified in the options

data/lib/commands/unipept/pept2lca.rb

@@ -1,6 +1,10 @@
 require_relative 'api_runner'
 module Unipept::Commands
   class Pept2lca < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
     def default_batch_size
       if options[:all]
         100

data/lib/commands/unipept/pept2prot.rb

@@ -2,6 +2,10 @@ require_relative 'api_runner'
 
 module Unipept::Commands
   class Pept2prot < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
     def default_batch_size
       if options[:all]
         5

data/lib/commands/unipept/pept2taxa.rb

@@ -1,6 +1,10 @@
 require_relative 'api_runner'
 module Unipept::Commands
   class Pept2taxa < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
     def default_batch_size
       if options[:all]
         5

data/lib/commands/unipept/taxonomy.rb

@@ -1,6 +1,10 @@
 require_relative 'api_runner'
 module Unipept::Commands
   class Taxonomy < ApiRunner
+    def required_fields
+      ['taxon_id']
+    end
+
     def default_batch_size
       100
     end

data/test/commands/unipept/test_api_runner.rb

@@ -103,6 +103,18 @@ module Unipept
       assert_equal(%w(a b c), output)
     end
 
+    def test_required_fields
+      assert_equal([], new_runner.required_fields)
+    end
+
+    def test_required_fields_configurable
+      r = new_runner
+      def r.required_fields
+        ['test']
+      end
+      assert_equal(['test'], r.required_fields)
+    end
+
     def test_default_batch_size
       assert_raises NotImplementedError do
         new_runner.default_batch_size
@@ -149,6 +161,22 @@ module Unipept
       assert_equal([], runner.selected_fields)
     end
 
+    def test_required_fields_are_not_selected_with_empty_selection
+      runner = new_runner
+      def runner.required_fields
+        ['test']
+      end
+      assert_equal([], runner.selected_fields)
+    end
+
+    def test_required_fields_are_selected
+      runner = new_runner('test',  host: 'http://param_host', select: 'field')
+      def runner.required_fields
+        ['test']
+      end
+      assert_equal([/^field$/, /^test$/], runner.selected_fields)
+    end
+
     def test_single_selected_fields
       runner = new_runner('test',  host: 'http://param_host', select: 'field')
       assert_equal([/^field$/], runner.selected_fields)

data/test/commands/unipept/test_pept2lca.rb

@@ -10,6 +10,12 @@ module Unipept
       assert_equal(100, pept2lca.default_batch_size)
     end
 
+    def test_required_fields
+      command = Cri::Command.define { name 'pept2lca' }
+      pept2lca = Commands::Pept2lca.new({ host: 'http://api.unipept.ugent.be' }, [], command)
+      assert_equal(['peptide'], pept2lca.required_fields)
+    end
+
     def test_argument_batch_size
       command = Cri::Command.define { name 'pept2lca' }
       pept2lca = Commands::Pept2lca.new({ host: 'http://api.unipept.ugent.be', batch: '123' }, [], command)
@@ -62,6 +68,19 @@ module Unipept
       assert_raises(StopIteration) { lines.next }
     end
 
+    def test_run_with_fasta_multiple_batches_and_select
+      out, err = capture_io_while do
+        Commands::Unipept.run(%w(pept2lca --host http://api.unipept.ugent.be --batch 2 --select taxon_id >test AALTER AALER >tost AALTER))
+      end
+      lines = out.each_line
+      assert_equal('', err)
+      assert(lines.next.start_with? 'fasta_header,peptide,taxon_id')
+      assert(lines.next.start_with? '>test,AALTER,1')
+      assert(lines.next.start_with? '>test,AALER,1')
+      assert(lines.next.start_with? '>tost,AALTER,1')
+      assert_raises(StopIteration) { lines.next }
+    end
+
     def test_run_with_fasta_multiple_batches_json
       out, err = capture_io_while do
         Commands::Unipept.run(%w(pept2lca --host http://api.unipept.ugent.be --batch 2 --format json >test AALTER AALER >tost AALTER))

data/test/commands/unipept/test_pept2prot.rb

@@ -10,6 +10,12 @@ module Unipept
       assert_equal(5, pept2prot.default_batch_size)
     end
 
+    def test_required_fields
+      command = Cri::Command.define { name 'pept2prot' }
+      pept2prot = Commands::Pept2prot.new({ host: 'http://api.unipept.ugent.be' }, [], command)
+      assert_equal(['peptide'], pept2prot.required_fields)
+    end
+
     def test_help
       out, _err = capture_io_while do
         assert_raises SystemExit do
@@ -48,6 +54,18 @@ module Unipept
       assert(lines.count { |line| line.start_with? '>tost,EGGAGSSTGQR,' } >= 1)
     end
 
+    def test_run_with_fasta_multiple_batches_and_select
+      out, err = capture_io_while do
+        Commands::Unipept.run(%w(pept2prot --host http://api.unipept.ugent.be --batch 2 --select uniprot_id >test EGGAGSSTGQR ENFVYIAK >tost EGGAGSSTGQR))
+      end
+      lines = out.each_line
+      assert_equal('', err)
+      assert(lines.next.start_with? 'fasta_header,peptide,uniprot_id')
+      assert(lines.count { |line| line.start_with? '>test,EGGAGSSTGQR,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>test,ENFVYIAK,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>tost,EGGAGSSTGQR,' } >= 1)
+    end
+
     def test_run_with_fasta_multiple_batches_json
       out, err = capture_io_while do
         Commands::Unipept.run(%w(pept2prot --host http://api.unipept.ugent.be --batch 2 --format json >test EGGAGSSTGQR ENFVYIAK >tost EGGAGSSTGQR))

data/test/commands/unipept/test_pept2taxa.rb

@@ -10,6 +10,12 @@ module Unipept
       assert_equal(5, pept2taxa.default_batch_size)
     end
 
+    def test_required_fields
+      command = Cri::Command.define { name 'pept2taxa' }
+      pept2taxa = Commands::Pept2taxa.new({ host: 'http://api.unipept.ugent.be' }, [], command)
+      assert_equal(['peptide'], pept2taxa.required_fields)
+    end
+
     def test_help
       out, _err = capture_io_while do
         assert_raises SystemExit do
@@ -48,6 +54,18 @@ module Unipept
       assert(lines.count { |line| line.start_with? '>tost,EGGAGSSTGQR,' } >= 1)
     end
 
+    def test_run_with_fasta_multiple_batches_and_select
+      out, err = capture_io_while do
+        Commands::Unipept.run(%w(pept2taxa --host http://api.unipept.ugent.be --batch 2 --select taxon_id >test EGGAGSSTGQR ENFVYIAK >tost EGGAGSSTGQR))
+      end
+      lines = out.each_line
+      assert_equal('', err)
+      assert(lines.next.start_with? 'fasta_header,peptide,taxon_id')
+      assert(lines.count { |line| line.start_with? '>test,EGGAGSSTGQR,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>test,ENFVYIAK,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>tost,EGGAGSSTGQR,' } >= 1)
+    end
+
     def test_run_with_fasta_multiple_batches_json
       out, err = capture_io_while do
         Commands::Unipept.run(%w(pept2taxa --host http://api.unipept.ugent.be --batch 2 --format json >test EGGAGSSTGQR ENFVYIAK >tost EGGAGSSTGQR))

data/test/commands/unipept/test_taxa2lca.rb

@@ -10,6 +10,12 @@ module Unipept
       end
     end
 
+    def test_required_fields
+      command = Cri::Command.define { name 'taxa2lca' }
+      taxa2lca = Commands::Taxa2lca.new({ host: 'http://api.unipept.ugent.be' }, [], command)
+      assert_equal([], taxa2lca.required_fields)
+    end
+
     def test_help
       out, _err = capture_io_while do
         assert_raises SystemExit do

data/test/commands/unipept/test_taxonomy.rb

@@ -8,6 +8,12 @@ module Unipept
       assert_equal(100, taxonomy.default_batch_size)
     end
 
+    def test_required_fields
+      command = Cri::Command.define { name 'taxonomy' }
+      taxonomy = Commands::Taxonomy.new({ host: 'http://api.unipept.ugent.be' }, [], command)
+      assert_equal(['taxon_id'], taxonomy.required_fields)
+    end
+
     def test_help
       out, _err = capture_io_while do
         assert_raises SystemExit do
@@ -46,6 +52,18 @@ module Unipept
       assert(lines.count { |line| line.start_with? '>tost,1,' } >= 1)
     end
 
+    def test_run_with_fasta_multiple_batches_and_select
+      out, err = capture_io_while do
+        Commands::Unipept.run(%w(taxonomy --host http://api.unipept.ugent.be --batch 2 --select taxon_name >test 1 216816 >tost 1))
+      end
+      lines = out.each_line
+      assert_equal('', err)
+      assert(lines.next.start_with? 'fasta_header,taxon_id,taxon_name')
+      assert(lines.count { |line| line.start_with? '>test,1,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>test,216816,' } >= 1)
+      assert(lines.count { |line| line.start_with? '>tost,1,' } >= 1)
+    end
+
     def test_run_with_fasta_multiple_batches_json
       out, err = capture_io_while do
         Commands::Unipept.run(%w(taxonomy --host http://api.unipept.ugent.be --batch 2 --format json >test 1 216816 >tost 1))

data/unipept.gemspec

@@ -2,16 +2,16 @@
 # DO NOT EDIT THIS FILE DIRECTLY
 # Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
 # -*- encoding: utf-8 -*-
-# stub: unipept 1.1.0 ruby lib
+# stub: unipept 1.1.1 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "unipept"
-  s.version = "1.1.0"
+  s.version = "1.1.1"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib"]
   s.authors = ["Toon Willems", "Bart Mesuere", "Tom Naessens"]
-  s.date = "2015-09-16"
+  s.date = "2016-06-12"
   s.description = "  Command line interface to the Unipept (http://unipept.ugent.be) web services\n  (pept2lca, taxa2lca, pept2taxa, pept2prot and taxonomy) and some utility\n  commands for handling proteins using the command line.\n"
   s.email = "unipept@ugent.be"
   s.executables = ["unipept", "prot2pept", "peptfilter", "uniprot"]

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: unipept
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.1.1
 platform: ruby
 authors:
 - Toon Willems
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-09-16 00:00:00.000000000 Z
+date: 2016-06-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cri