unipept 2.1.1 → 2.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7d025bbebf7189fd0afe9cc8cd945f8ec2f14c231a3081f0e24284dc2997d988
-  data.tar.gz: 889ea0bbd83199ec503a7736a21d5e216e054af1a691603c413958ee1c5efab1
+  metadata.gz: b8f3ce5867e19149483ab70e24d4cc5da97d8ba9e62e38491b714abc89078ee0
+  data.tar.gz: '09cfcdbc3608ee36a75825e7f7555add797b002e3abccc5caada703efc8d0b86'
 SHA512:
-  metadata.gz: 7394e17376d8e22f57d07046b2a0930a133717b44c3ceeaf50c61ef05011d711ea36a88e0212b8dde1c6efefce5b7d4e9fd4d7461363b8aea83e24bbaea214d8
-  data.tar.gz: 470253cdb07286db4f9e9d37f0323a64f4f9726d007d7b4a8dc73fdc171335ec6e50d91505e94c533b11fb141ec6ccaff4f6437d41d8997108e0ed9df3a537bf
+  metadata.gz: e29aa64adf5a502377c2fdedb63969d29fa6f8d517262d65ec0a99b40746bf0a1d3d5a26552dd25cd425008d0f8a0f4ed7bd69ec6786d77ef2c688ce0e5b7c5d
+  data.tar.gz: 9d6d08532a96fb85423a7843745df5fe75f125e6f7119f6ba2b76f525b75bd057aac7d27a0a5c1e6357df5aed0af42bb00a9fd260aa3be58f623dc75dc3da7b4

data/.github/dependabot.yml

@@ -0,0 +1,9 @@
+version: 2
+updates:
+- package-ecosystem: bundler
+  directory: "/"
+  schedule:
+    interval: monthly
+    time: "13:00"
+    timezone: Europe/Brussels
+  open-pull-requests-limit: 99

data/.github/workflows/ci.yml

@@ -7,27 +7,16 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        ruby: ["2.4", "2.5", "2.6"]
+        ruby: ["2.7", "3.0", "3.1"]
       fail-fast: false
     name: Test Ruby ${{ matrix.ruby }}
     steps:
       - uses: actions/checkout@v1
       - name: Setup Ruby
-        uses: actions/setup-ruby@v1
+        uses: ruby/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
+          bundler-cache: true
       - name: Test
         run: |
           bundle exec rake test
@@ -37,19 +26,9 @@ jobs:
     steps:
       - uses: actions/checkout@v1
       - name: Setup Ruby
-        uses: actions/setup-ruby@v1
-      - name: Cache Ruby dependencies
-        uses: actions/cache@v1
+        uses: ruby/setup-ruby@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
+          bundler-cache: true
       - name: Test
         run: |
-          bundle exec rake test_style
+          bundle exec rake test_style

data/.rakeTasks

@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<Settings><!--This file was automatically generated by Ruby plugin.
+You are allowed to: 
+1. Remove rake task
+2. Add existing rake tasks
+To add existing rake tasks automatically delete this file and reload the project.
+--><RakeGroup description="" fullCmd="" taksId="rake"><RakeTask description="Build gem into pkg/" fullCmd="build" taksId="build" /><RakeTask description="Remove any temporary products" fullCmd="clean" taksId="clean" /><RakeTask description="Remove any generated files" fullCmd="clobber" taksId="clobber" /><RakeTask description="Remove RDoc HTML files" fullCmd="clobber_rdoc" taksId="clobber_rdoc" /><RakeTask description="Start IRB with all runtime dependencies loaded" fullCmd="console[script]" taksId="console[script]" /><RakeGroup description="" fullCmd="" taksId="gemcutter"><RakeTask description="Release gem to Gemcutter" fullCmd="gemcutter:release" taksId="release" /></RakeGroup><RakeTask description="Generate and validate gemspec" fullCmd="gemspec" taksId="gemspec" /><RakeGroup description="" fullCmd="" taksId="gemspec"><RakeTask description="Display the gemspec for debugging purposes, as juwelier knows it (not from the filesystem)" fullCmd="gemspec:debug" taksId="debug" /><RakeTask description="Regenerate the gemspec on the filesystem" fullCmd="gemspec:generate" taksId="generate" /><RakeTask description="Regenerate and validate gemspec, and then commits and pushes to git" fullCmd="gemspec:release" taksId="release" /><RakeTask description="Validates the gemspec on the filesystem" fullCmd="gemspec:validate" taksId="validate" /></RakeGroup><RakeGroup description="" fullCmd="" taksId="git"><RakeTask description="Tag and push release to git" fullCmd="git:release" taksId="release" /></RakeGroup><RakeTask description="Build and install gem using `gem install`" fullCmd="install" taksId="install" /><RakeTask description="Build RDoc HTML files" fullCmd="rdoc" taksId="rdoc" /><RakeTask description="Release gem" fullCmd="release" taksId="release" /><RakeTask description="Rebuild RDoc HTML files" fullCmd="rerdoc" taksId="rerdoc" /><RakeTask description="Run RuboCop" fullCmd="test_style" taksId="test_style" /><RakeGroup description="" fullCmd="" taksId="test_style"><RakeTask description="Auto-correct RuboCop offenses" fullCmd="test_style:auto_correct" taksId="auto_correct" /></RakeGroup><RakeTask description="Displays the current version" fullCmd="version" taksId="version" /><RakeGroup description="" fullCmd="" taksId="version"><RakeGroup description="" fullCmd="" taksId="bump"><RakeTask description="Bump the major version by 1" fullCmd="version:bump:major" taksId="major" /><RakeTask description="Bump the a minor version by 1" fullCmd="version:bump:minor" taksId="minor" /><RakeTask description="Bump the patch version by 1" fullCmd="version:bump:patch" taksId="patch" /></RakeGroup><RakeTask description="Writes out an explicit version" fullCmd="version:write" taksId="write" /></RakeGroup><RakeTask description="" fullCmd="console" taksId="console" /><RakeTask description="" fullCmd="default" taksId="default" /><RakeTask description="" fullCmd="gemspec_required" taksId="gemspec_required" /><RakeTask description="" fullCmd="rdoc/created.rid" taksId="rdoc/created.rid" /><RakeTask description="" fullCmd="test" taksId="test" /><RakeTask description="" fullCmd="test_unit" taksId="test_unit" /><RakeTask description="" fullCmd="version_required" taksId="version_required" /></RakeGroup></Settings>

data/.rubocop.yml

@@ -1,4 +1,6 @@
 AllCops:
+  TargetRubyVersion: 2.7
+  NewCops: enable
   Exclude:
     - 'unipept.gemspec'
     - 'vendor/**/*'

data/.ruby-version

@@ -1 +1 @@
-2.6.6
+3.0.0

data/CITATION.cff

@@ -0,0 +1,30 @@
+cff-version: 1.2.0
+message: "If you use this software, please cite it as below."
+preferred-citation:
+  type: article
+  authors:
+  - family-names: "Verschaffelt"
+    given-names: "Pieter"
+  - family-names: "Van Thienen"
+    given-names: "Philippe"
+  - family-names: "Van Den Bossche"
+    given-names: "Tim"
+  - family-names: "Van der Jeugt"
+    given-names: "Felix"
+  - family-names: "De Tender"
+    given-names: "Caroline"
+  - family-names: "Martens"
+    given-names: "Lennart"
+  - family-names: "Dawyndt"
+    given-names: "Peter"
+  - family-names: "Mesuere"
+    given-names: "Bart"
+    orcid: "https://orcid.org/0000-0003-0610-3441"
+  doi: "10.1093/bioinformatics/btaa553"
+  journal: "Bioinformatics"
+  start: 4220
+  end: 4221
+  title: "Unipept CLI 2.0: adding support for visualizations and functional annotations"
+  issue: 14
+  volume: 36
+  year: 2020

data/Gemfile

@@ -7,9 +7,9 @@ gem 'typhoeus', '~> 1.3.1'
 group :development do
   gem 'minitest', '~> 5.14'
   gem 'rake', '~> 13.0.1'
-  gem 'rubocop', '~>0.79.0'
+  gem 'rubocop', '~>1.11'
 end
 
 group :release do
-  gem 'jeweler'
+  gem 'juwelier'
 end

data/Gemfile.lock

@@ -1,85 +1,104 @@
 GEM
   remote: http://rubygems.org/
   specs:
-    addressable (2.4.0)
-    ast (2.4.0)
+    addressable (2.8.0)
+      public_suffix (>= 2.0.2, < 5.0)
+    ast (2.4.2)
     builder (3.2.4)
-    cri (2.15.10)
+    cri (2.15.11)
     descendants_tracker (0.0.4)
       thread_safe (~> 0.3, >= 0.3.1)
     ethon (0.12.0)
       ffi (>= 1.3.0)
-    faraday (0.9.2)
+    faraday (1.3.0)
+      faraday-net_http (~> 1.0)
       multipart-post (>= 1.2, < 3)
+      ruby2_keywords
+    faraday-net_http (1.0.1)
     ffi (1.12.1)
-    git (1.5.0)
-    github_api (0.16.0)
-      addressable (~> 2.4.0)
+    git (1.8.1)
+      rchardet (~> 1.8)
+    github_api (0.19.0)
+      addressable (~> 2.4)
       descendants_tracker (~> 0.0.4)
-      faraday (~> 0.8, < 0.10)
-      hashie (>= 3.4)
-      mime-types (>= 1.16, < 3.0)
+      faraday (>= 0.8, < 2)
+      hashie (~> 3.5, >= 3.5.2)
       oauth2 (~> 1.0)
-    hashie (4.0.0)
+    hashie (3.6.0)
     highline (2.0.3)
-    jaro_winkler (1.5.4)
-    jeweler (2.3.9)
+    juwelier (2.4.9)
       builder
       bundler
-      git (>= 1.2.5)
-      github_api (~> 0.16.0)
-      highline (>= 1.6.15)
-      nokogiri (>= 1.5.10)
+      git
+      github_api
+      highline
+      kamelcase (~> 0)
+      nokogiri
       psych
       rake
       rdoc
       semver2
-    jwt (2.2.1)
-    mime-types (2.99.3)
-    mini_portile2 (2.4.0)
-    minitest (5.14.0)
-    multi_json (1.14.1)
+    jwt (2.2.2)
+    kamelcase (0.0.2)
+      semver2 (~> 3)
+    mini_portile2 (2.8.0)
+    minitest (5.15.0)
+    multi_json (1.15.0)
     multi_xml (0.6.0)
     multipart-post (2.1.1)
-    nokogiri (1.10.7)
-      mini_portile2 (~> 2.4.0)
-    oauth2 (1.4.2)
+    nokogiri (1.13.4)
+      mini_portile2 (~> 2.8.0)
+      racc (~> 1.4)
+    oauth2 (1.4.4)
       faraday (>= 0.8, < 2.0)
       jwt (>= 1.0, < 3.0)
       multi_json (~> 1.3)
       multi_xml (~> 0.5)
       rack (>= 1.2, < 3)
-    parallel (1.19.1)
-    parser (2.7.0.2)
-      ast (~> 2.4.0)
-    psych (3.1.0)
-    rack (2.1.1)
-    rainbow (3.0.0)
-    rake (13.0.1)
-    rdoc (6.2.1)
-    rubocop (0.79.0)
-      jaro_winkler (~> 1.5.1)
+    parallel (1.22.1)
+    parser (3.1.1.0)
+      ast (~> 2.4.1)
+    psych (4.0.3)
+      stringio
+    public_suffix (4.0.6)
+    racc (1.6.0)
+    rack (2.2.3)
+    rainbow (3.1.1)
+    rake (13.0.6)
+    rchardet (1.8.0)
+    rdoc (6.4.0)
+      psych (>= 4.0.0)
+    regexp_parser (2.3.0)
+    rexml (3.2.5)
+    rubocop (1.27.0)
       parallel (~> 1.10)
-      parser (>= 2.7.0.1)
+      parser (>= 3.1.0.0)
       rainbow (>= 2.2.2, < 4.0)
+      regexp_parser (>= 1.8, < 3.0)
+      rexml
+      rubocop-ast (>= 1.16.0, < 2.0)
       ruby-progressbar (~> 1.7)
-      unicode-display_width (>= 1.4.0, < 1.7)
-    ruby-progressbar (1.10.1)
+      unicode-display_width (>= 1.4.0, < 3.0)
+    rubocop-ast (1.17.0)
+      parser (>= 3.1.1.0)
+    ruby-progressbar (1.11.0)
+    ruby2_keywords (0.0.4)
     semver2 (3.4.2)
+    stringio (3.0.1)
     thread_safe (0.3.6)
     typhoeus (1.3.1)
       ethon (>= 0.9.0)
-    unicode-display_width (1.6.1)
+    unicode-display_width (2.1.0)
 
 PLATFORMS
   ruby
 
 DEPENDENCIES
   cri (~> 2.15.10)
-  jeweler
+  juwelier
   minitest (~> 5.14)
   rake (~> 13.0.1)
-  rubocop (~> 0.79.0)
+  rubocop (~> 1.11)
   typhoeus (~> 1.3.1)
 
 BUNDLED WITH

data/README.md

@@ -7,11 +7,11 @@ Documentation about the web service can be found at [http://unipept.ugent.be/api
 
 ## Installation
 
-To use the Unipept CLI, Ruby version 1.9.3 or higher needs to be installed. You can check this by running `ruby -v` on the commandline:
+To use the Unipept CLI, Ruby version 2.6 or higher needs to be installed. You can check this by running `ruby -v` on the commandline:
 
 ```
 $ ruby -v
-ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-darwin14]
+ruby 3.0.0p0 (2020-12-25 revision 95aff21468) [arm64-darwin21]
 ```
 
 More information on installing Ruby can be found at https://www.ruby-lang.org/en/installation/

data/Rakefile

@@ -11,8 +11,8 @@ require 'rake'
 require 'rake/testtask'
 require 'rubocop/rake_task'
 begin
-  require 'jeweler'
-  Jeweler::Tasks.new do |gem|
+  require 'juwelier'
+  Juwelier::Tasks.new do |gem|
     # gem is a Gem::Specification... see http://guides.rubygems.org/specification-reference/ for more options
     gem.name = 'unipept'
     gem.executables = %w[unipept prot2pept peptfilter uniprot]
@@ -26,9 +26,9 @@ begin
     EOS
     gem.email = 'unipept@ugent.be'
     gem.authors = ['Bart Mesuere', 'Pieter Verschaffelt', 'Toon Willems', 'Tom Naessens']
-    gem.required_ruby_version = '>= 2.0.0'
+    gem.required_ruby_version = '>= 2.6.0'
   end
-  Jeweler::RubygemsDotOrgTasks.new
+  Juwelier::RubygemsDotOrgTasks.new
 rescue LoadError
   # do nothing
 end

data/VERSION

@@ -1 +1 @@
-2.1.1
+2.2.2

data/lib/batch_iterator.rb

@@ -20,6 +20,8 @@ module Unipept
       first_line = lines.next rescue return
       if fasta? first_line
         fasta_iterator(first_line, lines, &block)
+      elsif csv_taxa2tree? first_line
+        csv_taxa_iterator(first_line, lines, &block)
       else
         normal_iterator(first_line, lines, &block)
       end
@@ -34,6 +36,10 @@ module Unipept
       line.start_with? '>'
     end
 
+    def csv_taxa2tree?(line)
+      line.include? 'taxon_id'
+    end
+
     private
 
     # Splits the input lines in fasta format into slices, based on the
@@ -69,5 +75,15 @@ module Unipept
         end
       end.each_slice(batch_size).with_index(&block)
     end
+
+    def csv_taxa_iterator(first_line, next_lines, &block)
+      # Find index of taxon_id in the first_line and only parse this part from the next lines
+      taxon_idx = first_line.rstrip.split(',').find_index('taxon_id')
+      Enumerator.new do |y|
+        loop do
+          y << next_lines.next.rstrip.split(',')[taxon_idx]
+        end
+      end.each_slice(batch_size).with_index(&block)
+    end
   end
 end

data/lib/commands/prot2pept.rb

@@ -1,7 +1,6 @@
 module Unipept::Commands
   class Prot2pept
-    attr_reader :root_command
-    attr_reader :valid_formats
+    attr_reader :root_command, :valid_formats
 
     @root_command = Cri::Command.define do
       name 'prot2pept'

data/lib/commands/unipept/api_runner.rb

@@ -2,18 +2,14 @@ require_relative '../../retryable_typhoeus'
 
 module Unipept
   class Commands::ApiRunner < Cri::CommandRunner
-    attr_reader :configuration
-
-    attr_reader :url
-
-    attr_reader :user_agent
+    attr_reader :configuration, :url, :user_agent
 
     def initialize(args, opts, cmd)
       super
       @configuration = Unipept::Configuration.new
 
       @host = host
-      @user_agent = 'Unipept CLI - unipept ' + Unipept::VERSION
+      @user_agent = "Unipept CLI - unipept #{Unipept::VERSION}"
       @url = "#{@host}/api/v1/#{cmd.name}.json"
       @fasta = false
     end
@@ -39,7 +35,7 @@ module Unipept
     # - if none of the previous are given, uses stdin
     def input_iterator
       return arguments.each unless arguments.empty?
-      return IO.foreach(options[:input]) if options[:input]
+      return File.foreach(options[:input]) if options[:input]
 
       $stdin.each_line
     end
@@ -115,6 +111,8 @@ module Unipept
       batch_order = Unipept::BatchOrder.new
       last_id = 0
 
+      # puts input_iterator.inspect
+
       batch_iterator.iterate(input_iterator) do |input_slice, batch_id, fasta_mapper|
         last_id = batch_id
         @fasta = !fasta_mapper.nil?
@@ -146,11 +144,11 @@ module Unipept
     def save_error(message)
       path = error_file_path
       FileUtils.mkdir_p File.dirname(path)
-      File.open(path, 'w') { |f| f.write message }
+      File.write(path, message)
       warn "API request failed! log can be found in #{path}"
     end
 
-    private
+    protected
 
     def error_file_path
       File.expand_path(File.join(Dir.home, '.unipept', "unipept-#{Time.now.strftime('%F-%T')}.log"))
@@ -171,7 +169,7 @@ module Unipept
 
       lambda do
         unless result.empty?
-          output_writer.write_line formatter.header(result, fasta_mapper) if batch_id.zero? && !options[:"no-header"]
+          output_writer.write_line formatter.header(result, fasta_mapper) if batch_id.zero? && !options[:'no-header']
           output_writer.write_line formatter.format(result, fasta_mapper, batch_id.zero?)
         end
       end
@@ -181,7 +179,7 @@ module Unipept
       if response.timed_out?
         -> { save_error('request timed out, continuing anyway, but results might be incomplete') }
       elsif response.code.zero?
-        -> { save_error('could not get an http response, continuing anyway, but results might be incomplete' + response.return_message) }
+        -> { save_error("could not get an http response, continuing anyway, but results might be incomplete#{response.return_message}") }
       else
         -> { save_error("Got #{response.code}: #{response.response_body}\nRequest headers: #{response.request.options}\nRequest body:\n#{response.request.encoded_body}\n\n") }
       end
@@ -210,7 +208,7 @@ module Unipept
           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
+                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

data/lib/commands/unipept/config.rb

@@ -7,7 +7,7 @@ module Unipept
 
       if arguments.size == 2
         set_config(key, value)
-        puts key + ' was set to ' + value
+        puts "#{key} was set to #{value}"
       else
         puts get_config(key)
       end

data/lib/commands/unipept/pept2taxa.rb

@@ -6,11 +6,7 @@ module Unipept::Commands
     end
 
     def default_batch_size
-      if options[:all]
-        5
-      else
-        10
-      end
+      5
     end
   end
 end

data/lib/commands/unipept/taxa2tree.rb

@@ -0,0 +1,74 @@
+require_relative 'api_runner'
+module Unipept::Commands
+  class Taxa2Tree < ApiRunner
+    def initialize(args, opts, cmd)
+      super
+
+      # JSON is the default format for this command
+      args[:format] = 'json' unless args[:format]
+
+      unless %w[url html json].include? args[:format]
+        warn "Format #{args[:format]} is not supported by taxa2tree. Use html, url or json (default)."
+        exit 1
+      end
+
+      if options[:format] == 'html'
+        # Overwrite the URL for this command, since it's possible that it uses HTML generated by the server.
+        @url = "#{@host}/api/v1/#{cmd.name}.html"
+      elsif args[:format] == 'url'
+        @link = true
+      end
+    end
+
+    def required_fields
+      ['taxon_id']
+    end
+
+    def batch_size
+      return arguments.length unless arguments.empty?
+      return File.foreach(options[:input]).inject(0) { |c, _| c + 1 } if options[:input]
+
+      @stdin_contents = $stdin.readlines
+      @stdin_contents.length
+    end
+
+    def input_iterator
+      return arguments.each unless arguments.empty?
+      return File.foreach(options[:input]) if options[:input]
+
+      @stdin_contents.each
+    end
+
+    protected
+
+    def filter_result(response)
+      return response if response.start_with?('<!DOCTYPE')
+
+      # We do not filter here, since select is not supported by the taxa2tree-command
+      [JSON[response]] rescue []
+    end
+
+    def construct_request_body(input)
+      data = nil
+
+      if input[0].include?(',')
+        data = input.map do |item|
+          splitted = item.rstrip.split ','
+          splitted[1] = splitted[1].to_i
+          splitted
+        end
+        data = data.to_h
+      else
+        data = Hash.new 0
+        input.each do |i|
+          data[i.rstrip] += 1
+        end
+      end
+
+      {
+        counts: data,
+        link: @link
+      }
+    end
+  end
+end

data/lib/commands/unipept.rb

@@ -19,6 +19,7 @@ require_relative 'unipept/pept2taxa'
 require_relative 'unipept/peptinfo'
 require_relative 'unipept/taxa2lca'
 require_relative 'unipept/taxonomy'
+require_relative 'unipept/taxa2tree'
 
 module Unipept
   class Commands::Unipept
@@ -35,6 +36,7 @@ module Unipept
       add_taxa2lca_command
       add_pept2prot_command
       add_taxonomy_command
+      add_taxa2tree_command
     end
 
     def run(args)
@@ -66,7 +68,7 @@ module Unipept
         option nil, :batch, 'specify the batch size', argument: :required, hidden: true
         option nil, :parallel, 'specify the number of parallel requests', argument: :required, hidden: true
         option :o, :output, 'write output to file', argument: :required
-        option :f, :format, "define the output format (available: #{Unipept::Formatter.available.join ', '}) (default: #{Unipept::Formatter.default})", argument: :required
+        option :f, :format, "define the output format (available: #{Unipept::Formatter.available.select { |f| f != 'html' && f != 'url' }.join(', ')}) (default: #{Unipept::Formatter.default}).", argument: :required
 
         # Configuration options
         option nil, 'host', 'specify the server running the Unipept web service', argument: :required
@@ -292,6 +294,28 @@ module Unipept
       end
     end
 
+    def add_taxa2tree_command
+      @root_command.define_command('taxa2tree') do
+        usage 'taxa2tree [options]'
+        summary 'Compute lineage tree for given list of taxa'
+        description <<-EOS
+        The unipept taxa2tree command computes a lineage tree of a given list of NCBI Taxonomy Identifiers. A frequency table is computed for the given list of taxa. Secondly, the lineages for all taxa are looked up. These are then used to build a lineage tree with all counts set. The command expects a list of NCBI Taxonomy Identifiers 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 NCBI Taxonomy Identifiers are passed, in the order as listed above. Text files and standard input should have one NCBI Taxonomy Identifier per line.
+        EOS
+
+        option :f, :format, "define the output format (available: json, url, html) (default: 'json'). Note that xml and csv are not available for taxa2tree. html and url are used as an output format for visualizations.", argument: :required
+
+        runner Commands::Taxa2Tree
+      end
+    end
+
     def add_pept2prot_command
       @root_command.define_command('pept2prot') do
         usage 'pept2prot [options]'

data/lib/commands/uniprot.rb

@@ -2,8 +2,7 @@ require 'typhoeus'
 
 module Unipept::Commands
   class Uniprot
-    attr_reader :root_command
-    attr_reader :valid_formats
+    attr_reader :root_command, :valid_formats
 
     valid_formats = Set.new %w[fasta txt xml rdf gff sequence]
     @root_command = Cri::Command.define do
@@ -21,7 +20,7 @@ module Unipept::Commands
 
       The uniprot command yields just the protein sequences as a default, but can return several formats.
       EOS
-      required :f, :format, 'specify output format (available: ' + valid_formats.to_a.join(', ') + ') (default: sequence)'
+      required :f, :format, "specify output format (available: #{valid_formats.to_a.join(', ')}) (default: sequence)"
       flag :h, :help, 'show help for this command' do |_value, cmd|
         puts cmd.help
         exit 0
@@ -29,7 +28,7 @@ module Unipept::Commands
       run do |opts, args, _cmd|
         format = opts.fetch(:format, 'sequence')
         unless valid_formats.include? format
-          warn format + ' is not a valid output format. Available formats are: ' + valid_formats.to_a.join(', ')
+          warn "#{format} is not a valid output format. Available formats are: #{valid_formats.to_a.join(', ')}"
           exit 1
         end
         iterator = args.empty? ? $stdin.each_line : args
@@ -58,7 +57,7 @@ module Unipept::Commands
     # @return [String] The requested UniProt entry in the requested format
     def self.get_uniprot_entry(accession, format)
       if format == 'sequence'
-        get_uniprot_entry(accession, 'fasta').lines.map(&:chomp)[1..-1].join('')
+        get_uniprot_entry(accession, 'fasta').lines.map(&:chomp)[1..].join
       else
         # other format has been specified, just download and output
         resp = Typhoeus.get("https://www.uniprot.org/uniprot/#{accession}.#{format}")