unipept 2.0.0 → 2.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2f4204e204b0e0ab2cb65f750fb89e021d4d5cc1d0b845dd1e7d88e40fe221b0
-  data.tar.gz: 0f75df131260799611b2fab1d1be26ac248212bc7430b5845fc7afc97c70abfe
+  metadata.gz: f4e8010e58ccf55d028165182cb5ec241ce7501940d6cf3e80405660f9249076
+  data.tar.gz: 3d21f11d6b8b06ad60f781cb83d8e38a33c1d431854064bc1a69d072c7107559
 SHA512:
-  metadata.gz: 4d768d34c3b5aa94d507fd927235ba5347a7432c68e6d5a42cbb8efc360180dddccde15f3495c4717942a097ee538abde9cbf5edefefb2de485370f078bd396f
-  data.tar.gz: 61987f9e9f0b3be37801f7ef88ec39e4915207785fe294f781dafcdac7de8dd0041022497ec794b033022c132378f4393f40ed2d9eac55601b61d0398fe307a4
+  metadata.gz: af8d1bb017251e5bec0fb57217f26ae5c26e291c1e6e7cdddf62a248489539c023ecf3dc234d5126e992097bcf82b52e563bc908b0b8285a4d88326f142a3dcd
+  data.tar.gz: 25a8ef260a1de7a26d994912a90f73f807de1b4b7f53ca7963fe3ee3270324139de8353b61ba794b6d2bcc4ac3353d5f9f31827975b07edcb93d08527c771706

data/.github/workflows/ci.yml

@@ -0,0 +1,55 @@
+name: CI
+
+on: [push]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        ruby: ["2.4", "2.5", "2.6"]
+      fail-fast: false
+    name: Test Ruby ${{ matrix.ruby }}
+    steps:
+      - uses: actions/checkout@v1
+      - name: Setup Ruby
+        uses: actions/setup-ruby@v1
+        with:
+          ruby-version: ${{ matrix.ruby }}
+      - name: Cache Ruby dependencies
+        uses: actions/cache@v1
+        with:
+          path: vendor/bundle
+          key: ${{ runner.os }}-${{ matrix.ruby }}-gem-${{ hashFiles('**/Gemfile.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-${{ matrix.ruby }}-gem-
+      - name: Install dependencies
+        run: |
+          gem install bundler:1.17.2
+          bundle config path vendor/bundle
+          bundle install --jobs 4 --retry 3
+      - name: Test
+        run: |
+          bundle exec rake test
+  lint:
+    name: Lint
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v1
+      - name: Setup Ruby
+        uses: actions/setup-ruby@v1
+      - name: Cache Ruby dependencies
+        uses: actions/cache@v1
+        with:
+          path: vendor/bundle
+          key: ${{ runner.os }}-${{ matrix.ruby }}-gem-${{ hashFiles('**/Gemfile.lock') }}
+          restore-keys: |
+            ${{ runner.os }}-${{ matrix.ruby }}-gem-
+      - name: Install dependencies
+        run: |
+          gem install bundler:1.17.2
+          bundle config path vendor/bundle
+          bundle install --jobs 4 --retry 3
+      - name: Test
+        run: |
+          bundle exec rake test_style

data/.rubocop.yml

@@ -23,13 +23,17 @@ Metrics/AbcSize:
   Enabled: false
 Metrics/ClassLength:
   Enabled: false
-Metrics/LineLength:
+Layout/LineLength:
   Enabled: false
 Metrics/MethodLength:
   Enabled: false
 Metrics/BlockLength:
   Enabled: false
+Metrics/CyclomaticComplexity:
+  Enabled: false
+Metrics/PerceivedComplexity:
+  Enabled: false
 Naming/HeredocDelimiterNaming:
   Enabled: false
-Naming/UncommunicativeMethodParamName:
+Naming/MethodParameterName:
   Enabled: false

data/.ruby-version

@@ -1 +1 @@
-2.6.0
+2.4.6

data/Gemfile

@@ -1,16 +1,13 @@
 source 'http://rubygems.org'
 # Add dependencies required to use your gem here.
-# Example:
-#   gem "activesupport", ">= 2.3.5"
 
-gem 'cri', '~>2.15.3'
-gem 'typhoeus', '~> 1.1'
+gem 'cri', '~>2.15.10'
+gem 'typhoeus', '~> 1.3.1'
 
 group :development do
-  gem 'coveralls', '~> 0', require: false
-  gem 'minitest', '~> 5.7'
-  gem 'rake', '~> 10.4'
-  gem 'rubocop', '~>0.32'
+  gem 'minitest', '~> 5.14'
+  gem 'rake', '~> 13.0.1'
+  gem 'rubocop', '~>0.79.0'
 end
 
 group :release do

data/Gemfile.lock

@@ -3,24 +3,15 @@ GEM
   specs:
     addressable (2.4.0)
     ast (2.4.0)
-    builder (3.2.3)
-    colored (1.2)
-    coveralls (0.8.22)
-      json (>= 1.8, < 3)
-      simplecov (~> 0.16.1)
-      term-ansicolor (~> 1.3)
-      thor (~> 0.19.4)
-      tins (~> 1.6)
-    cri (2.15.3)
-      colored (~> 1.2)
+    builder (3.2.4)
+    cri (2.15.10)
     descendants_tracker (0.0.4)
       thread_safe (~> 0.3, >= 0.3.1)
-    docile (1.3.1)
     ethon (0.12.0)
       ffi (>= 1.3.0)
     faraday (0.9.2)
       multipart-post (>= 1.2, < 3)
-    ffi (1.10.0)
+    ffi (1.12.1)
     git (1.5.0)
     github_api (0.16.0)
       addressable (~> 2.4.0)
@@ -29,9 +20,9 @@ GEM
       hashie (>= 3.4)
       mime-types (>= 1.16, < 3.0)
       oauth2 (~> 1.0)
-    hashie (3.6.0)
-    highline (2.0.0)
-    jaro_winkler (1.5.2)
+    hashie (4.0.0)
+    highline (2.0.3)
+    jaro_winkler (1.5.4)
     jeweler (2.3.9)
       builder
       bundler
@@ -43,66 +34,53 @@ GEM
       rake
       rdoc
       semver2
-    json (2.1.0)
-    jwt (2.1.0)
+    jwt (2.2.1)
     mime-types (2.99.3)
     mini_portile2 (2.4.0)
-    minitest (5.11.3)
-    multi_json (1.13.1)
+    minitest (5.14.0)
+    multi_json (1.14.1)
     multi_xml (0.6.0)
-    multipart-post (2.0.0)
-    nokogiri (1.10.1)
+    multipart-post (2.1.1)
+    nokogiri (1.10.7)
       mini_portile2 (~> 2.4.0)
-    oauth2 (1.4.1)
-      faraday (>= 0.8, < 0.16.0)
+    oauth2 (1.4.2)
+      faraday (>= 0.8, < 2.0)
       jwt (>= 1.0, < 3.0)
       multi_json (~> 1.3)
       multi_xml (~> 0.5)
       rack (>= 1.2, < 3)
-    parallel (1.13.0)
-    parser (2.6.0.0)
+    parallel (1.19.1)
+    parser (2.7.0.2)
       ast (~> 2.4.0)
-    powerpack (0.1.2)
     psych (3.1.0)
-    rack (2.0.6)
+    rack (2.1.1)
     rainbow (3.0.0)
-    rake (10.5.0)
-    rdoc (6.1.1)
-    rubocop (0.63.0)
+    rake (13.0.1)
+    rdoc (6.2.1)
+    rubocop (0.79.0)
       jaro_winkler (~> 1.5.1)
       parallel (~> 1.10)
-      parser (>= 2.5, != 2.5.1.1)
-      powerpack (~> 0.1)
+      parser (>= 2.7.0.1)
       rainbow (>= 2.2.2, < 4.0)
       ruby-progressbar (~> 1.7)
-      unicode-display_width (~> 1.4.0)
-    ruby-progressbar (1.10.0)
+      unicode-display_width (>= 1.4.0, < 1.7)
+    ruby-progressbar (1.10.1)
     semver2 (3.4.2)
-    simplecov (0.16.1)
-      docile (~> 1.1)
-      json (>= 1.8, < 3)
-      simplecov-html (~> 0.10.0)
-    simplecov-html (0.10.2)
-    term-ansicolor (1.7.1)
-      tins (~> 1.0)
-    thor (0.19.4)
     thread_safe (0.3.6)
-    tins (1.20.2)
     typhoeus (1.3.1)
       ethon (>= 0.9.0)
-    unicode-display_width (1.4.1)
+    unicode-display_width (1.6.1)
 
 PLATFORMS
   ruby
 
 DEPENDENCIES
-  coveralls (~> 0)
-  cri (~> 2.15.3)
+  cri (~> 2.15.10)
   jeweler
-  minitest (~> 5.7)
-  rake (~> 10.4)
-  rubocop (~> 0.32)
-  typhoeus (~> 1.1)
+  minitest (~> 5.14)
+  rake (~> 13.0.1)
+  rubocop (~> 0.79.0)
+  typhoeus (~> 1.3.1)
 
 BUNDLED WITH
    1.17.2

data/README.md

@@ -1,9 +1,6 @@
 # unipept-cli
 
 [![Gem Version](https://badge.fury.io/rb/unipept.svg)](http://badge.fury.io/rb/unipept)
-[![Build Status](https://api.travis-ci.org/unipept/unipept-cli.svg)](https://travis-ci.org/unipept/unipept-cli)
-[![Coverage Status](https://coveralls.io/repos/unipept/unipept-cli/badge.svg)](https://coveralls.io/r/unipept/unipept-cli)
-[![Code Climate](https://codeclimate.com/github/unipept/unipept-cli/badges/gpa.svg)](https://codeclimate.com/github/unipept/unipept-cli)
 
 Unipept-cli offers a command line interface to the [Unipept](http://unipept.ugent.be) web service.
 Documentation about the web service can be found at [http://unipept.ugent.be/apidocs](http://unipept.ugent.be/apidocs), documentation about the command line tools at [http://unipept.ugent.be/clidocs](http://unipept.ugent.be/clidocs).
@@ -19,7 +16,7 @@ ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-darwin14]
 
 More information on installing Ruby can be found at https://www.ruby-lang.org/en/installation/
 
-The Unipept CLI is available as a *gem*. This means it can easily be installed with the following command:
+The Unipept CLI is available as a _gem_. This means it can easily be installed with the following command:
 
 ```bash
 $ gem install unipept

data/Rakefile

@@ -25,7 +25,7 @@ begin
     commands for handling proteins using the command line.
     EOS
     gem.email = 'unipept@ugent.be'
-    gem.authors = ['Toon Willems', 'Bart Mesuere', 'Tom Naessens']
+    gem.authors = ['Bart Mesuere', 'Pieter Verschaffelt', 'Toon Willems', 'Tom Naessens']
     gem.required_ruby_version = '>= 2.0.0'
   end
   Jeweler::RubygemsDotOrgTasks.new
@@ -43,7 +43,7 @@ end
 
 RuboCop::RakeTask.new(:test_style)
 
-task test: %i[test_unit test_style]
+task test: %i[test_unit]
 
 task default: :test
 

data/VERSION

@@ -1 +1 @@
-2.0.0
+2.1.0

data/lib/batch_iterator.rb

@@ -46,7 +46,7 @@ module Unipept
         input_set = Set.new
 
         slice.each do |line|
-          line.chomp!
+          line = line.chomp
           if fasta? line
             current_fasta_header = line
           else

data/lib/commands/unipept.rb

@@ -12,6 +12,7 @@ require_relative 'unipept/config'
 require_relative 'unipept/pept2ec'
 require_relative 'unipept/pept2funct'
 require_relative 'unipept/pept2go'
+require_relative 'unipept/pept2interpro'
 require_relative 'unipept/pept2lca'
 require_relative 'unipept/pept2prot'
 require_relative 'unipept/pept2taxa'
@@ -28,6 +29,7 @@ module Unipept
       add_pept2ec_command
       add_pept2funct_command
       add_pept2go_command
+      add_pept2interpro_command
       add_pept2lca_command
       add_peptinfo_command
       add_taxa2lca_command
@@ -149,8 +151,8 @@ module Unipept
 
     def add_pept2funct_command
       @root_command.define_command('pept2funct') do
-        usage 'pept2ec[options]'
-        summary 'Fetch EC numbers and GO terms of UniProt entries that match tryptic peptides.'
+        usage 'pept2funct[options]'
+        summary 'Fetch EC numbers, GO terms and InterPro codes of UniProt entries that match tryptic peptides.'
         description <<-EOS
         For each tryptic peptide the unipept pept2funct command retrieves from Unipept the set of EC numbers and GO terms from all UniProt entries whose protein sequence contains an exact matches to the tryptic peptide. The command expects a list of tryptic peptides that are passed
 
@@ -164,7 +166,7 @@ module Unipept
         EOS
 
         flag :e, :equate, 'equate isoleucine (I) and leucine (L) when matching peptides'
-        flag :a, :all, 'Also return the names of the EC numbers and GO terms. Note that this may have a performance penalty.'
+        flag :a, :all, 'Also return the names of the EC numbers, GO terms and InterPro codes. Note that this may have a performance penalty.'
         option :s, :select, 'select the information fields to return. Selected fields are passed as a comma separated list of field names. Multiple -s (or --select) options may be used.', argument: :required, multiple: true
 
         runner Commands::Pept2funct
@@ -195,6 +197,30 @@ module Unipept
       end
     end
 
+    def add_pept2interpro_command
+      @root_command.define_command('pept2interpro') do
+        usage 'pept2interpro [options]'
+        summary 'Fetch InterPro entries of UniProt entries that match tryptic peptides.'
+        description <<-EOS
+        For each tryptic peptide the unipept pept2interpro command retrieves from Unipept the set of InterPro entries from all UniProt entries whose protein sequence contains an exact matches to the tryptic peptide. The command expects a list of tryptic peptides that are passed
+
+        - as separate command line arguments
+
+        - in a text file that is passed as an argument to the -i option
+
+        - to standard input
+
+        The command will give priority to the first way tryptic peptides are passed, in the order as listed above. Text files and standard input should have one tryptic peptide per line.
+        EOS
+
+        flag :e, :equate, 'equate isoleucine (I) and leucine (L) when matching peptides'
+        flag :a, :all, 'Also return the names and types of the InterPro entries. Note that this may have a performance penalty.'
+        option :s, :select, 'select the information fields to return. Selected fields are passed as a comma separated list of field names. Multiple -s (or --select) options may be used.', argument: :required, multiple: true
+
+        runner Commands::Pept2interpro
+      end
+    end
+
     def add_pept2lca_command
       @root_command.define_command('pept2lca') do
         usage 'pept2lca [options]'

data/lib/commands/unipept/api_runner.rb

@@ -190,10 +190,74 @@ module Unipept
     def filter_result(json_response)
       result = JSON[json_response] rescue []
       result = [result] unless result.is_a? Array
+      key_order = result.first.keys if result.first
+      result = flatten_functional_fields(result) if formatter.instance_of?(Unipept::CSVFormatter)
       result.map! { |r| r.select! { |k, _v| selected_fields.any? { |f| f.match k } } } unless selected_fields.empty?
+      result = inflate_functional_fields(result, key_order) if formatter.instance_of?(Unipept::CSVFormatter) && result.first
       result
     end
 
+    # Transforms the hierarchical input to something without hierarchy. All fields
+    # associated with functional annotations are transformed to a flat alternative.
+    # Example: {"go" => {"go_term": xxx, "protein_count": yyy}} --> {"go_term" => [xxx], "protein_count" => [yyy]}
+    def flatten_functional_fields(data)
+      output = []
+      data.each do |row|
+        output_row = {}
+        row.each do |k, v|
+          if %w[ec go ipr].include? k
+            v.each do |item|
+              item.each do |field_name, field_value|
+                new_field_name = %w[ec_number go_term ipr_code].include?(field_name) ? field_name : k + '_' + field_name
+                output_row[new_field_name] = [] unless output_row.key? new_field_name
+                output_row[new_field_name] << field_value
+              end
+            end
+          else
+            output_row[k] = v
+          end
+        end
+        output << output_row
+      end
+      output
+    end
+
+    # Transforms a flattened input created by flatten_functional_fields to the original
+    # hierarchy.
+    def inflate_functional_fields(data, original_key_order)
+      output = []
+      data.each do |row|
+        output_row = {}
+
+        processed_keys = []
+        original_key_order.each do |original_key|
+          if %w[ec go ipr].include? original_key
+            # First, we take all distinct keys that start with "ec", "go" or "ipr"
+            annotation_keys = row.keys.select { |key| key.start_with? original_key }
+            processed_keys += annotation_keys
+            unless annotation_keys.empty?
+              # Each of the values of the annotation_keys is an array. All respective values of each of
+              # these arrays need to be put together into one hash. (E.g. {a => [1, 2], b=> [x, y]} --> [{a: 1, b: x}, {a: 2, b: y}])
+              reconstructed_objects = []
+              (0..row[annotation_keys[0]].length).each do |i|
+                reconstructed_object = {}
+                annotation_keys.each do |annotation_key|
+                  reconstructed_object[%w[ec_number go_term ipr_code].include?(annotation_key) ? annotation_key : annotation_key[annotation_key.index('_') + 1, annotation_key.length]] = row[annotation_key][i]
+                end
+                reconstructed_objects << reconstructed_object
+              end
+              output_row[original_key] = reconstructed_objects
+            end
+          elsif row.key? original_key
+            output_row[original_key] = row[original_key]
+          end
+        end
+
+        output << output_row
+      end
+      output
+    end
+
     def glob_to_regex(string)
       /^#{string.gsub('*', '.*')}$/
     end

data/lib/commands/unipept/pept2interpro.rb

@@ -0,0 +1,16 @@
+require_relative 'api_runner'
+module Unipept::Commands
+  class Pept2interpro < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
+    def default_batch_size
+      if options[:all]
+        100
+      else
+        1000
+      end
+    end
+  end
+end