unipept 1.4.1 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 773e25ed2369843012700a657100181e84254c905365d534136d0c1a68606ab7
-  data.tar.gz: 78830dfab3cab145cf248b2bc41d65806a3a4274ab1d48b9be1d530c7f2eb865
+  metadata.gz: 2f4204e204b0e0ab2cb65f750fb89e021d4d5cc1d0b845dd1e7d88e40fe221b0
+  data.tar.gz: 0f75df131260799611b2fab1d1be26ac248212bc7430b5845fc7afc97c70abfe
 SHA512:
-  metadata.gz: 55648c665eadd51c119e4091a3e8e110189df97ac70eeceee0a9a508e8b7112cd522dbec23ae55c1172e061b95b192cc383190640ec68dbd316273510640b598
-  data.tar.gz: fdcf5b35931caee76a3acfa6a8be7828f902570e9b6669f4d184b55321dcd0177d9bd363214f4d2b4766b13010eebe15e7e1beed549b4b864a95f71684a178ae
+  metadata.gz: 4d768d34c3b5aa94d507fd927235ba5347a7432c68e6d5a42cbb8efc360180dddccde15f3495c4717942a097ee538abde9cbf5edefefb2de485370f078bd396f
+  data.tar.gz: 61987f9e9f0b3be37801f7ef88ec39e4915207785fe294f781dafcdac7de8dd0041022497ec794b033022c132378f4393f40ed2d9eac55601b61d0398fe307a4

data/VERSION

@@ -1 +1 @@
-1.4.1
+2.0.0

data/lib/commands/unipept.rb

@@ -9,9 +9,13 @@ require_relative '../server_message'
 require_relative '../version'
 
 require_relative 'unipept/config'
+require_relative 'unipept/pept2ec'
+require_relative 'unipept/pept2funct'
+require_relative 'unipept/pept2go'
 require_relative 'unipept/pept2lca'
 require_relative 'unipept/pept2prot'
 require_relative 'unipept/pept2taxa'
+require_relative 'unipept/peptinfo'
 require_relative 'unipept/taxa2lca'
 require_relative 'unipept/taxonomy'
 
@@ -21,7 +25,11 @@ module Unipept
       @root_command = create_root_command
       add_config_command
       add_pept2taxa_command
+      add_pept2ec_command
+      add_pept2funct_command
+      add_pept2go_command
       add_pept2lca_command
+      add_peptinfo_command
       add_taxa2lca_command
       add_pept2prot_command
       add_taxonomy_command
@@ -115,6 +123,78 @@ module Unipept
       end
     end
 
+    def add_pept2ec_command
+      @root_command.define_command('pept2ec') do
+        usage 'pept2ec[options]'
+        summary 'Fetch EC numbers of UniProt entries that match tryptic peptides.'
+        description <<-EOS
+        For each tryptic peptide the unipept pept2ec command retrieves from Unipept the set of EC numbers 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 of the EC numbers. 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::Pept2ec
+      end
+    end
+
+    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.'
+        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
+
+        - 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 of the EC numbers and GO terms. 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
+      end
+    end
+
+    def add_pept2go_command
+      @root_command.define_command('pept2go') do
+        usage 'pept2go [options]'
+        summary 'Fetch GO terms of UniProt entries that match tryptic peptides.'
+        description <<-EOS
+        For each tryptic peptide the unipept pept2go command retrieves from Unipept the set of 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
+
+        - 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 of the GO terms. 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::Pept2go
+      end
+    end
+
     def add_pept2lca_command
       @root_command.define_command('pept2lca') do
         usage 'pept2lca [options]'
@@ -139,6 +219,30 @@ module Unipept
       end
     end
 
+    def add_peptinfo_command
+      @root_command.define_command('peptinfo') do
+        usage 'peptinfo [options]'
+        summary 'Fetch functional information and the taxonomic lowest common ancestor of UniProt entries that match tryptic peptides.'
+        description <<-EOS
+        For each tryptic peptide the unipept peptinfo command retrieves from Unipept the functional information and the lowest common ancestor of the set of taxa from all UniProt entries whose protein sequence contains an exact matches to the tryptic peptide. The lowest common ancestor is based on the topology of the Unipept Taxonomy -- a cleaned up version of the NCBI Taxonomy -- and is itself a record from the NCBI Taxonomy. 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, 'report the names of the functional annotations and all information fields of NCBI Taxonomy records available in Unipept. 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::Peptinfo
+      end
+    end
+
     def add_taxa2lca_command
       @root_command.define_command('taxa2lca') do
         usage 'taxa2lca [options]'

data/lib/commands/unipept/pept2ec.rb

@@ -0,0 +1,16 @@
+require_relative 'api_runner'
+module Unipept::Commands
+  class Pept2ec < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
+    def default_batch_size
+      if options[:all]
+        100
+      else
+        1000
+      end
+    end
+  end
+end

data/lib/commands/unipept/pept2funct.rb

@@ -0,0 +1,16 @@
+require_relative 'api_runner'
+module Unipept::Commands
+  class Pept2funct < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
+    def default_batch_size
+      if options[:all]
+        100
+      else
+        1000
+      end
+    end
+  end
+end

data/lib/commands/unipept/pept2go.rb

@@ -0,0 +1,16 @@
+require_relative 'api_runner'
+module Unipept::Commands
+  class Pept2go < ApiRunner
+    def required_fields
+      ['peptide']
+    end
+
+    def default_batch_size
+      if options[:all]
+        100
+      else
+        1000
+      end
+    end
+  end
+end

data/lib/commands/unipept/peptinfo.rb

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

data/lib/formatters.rb

@@ -186,7 +186,18 @@ module Unipept
       CSV.generate do |csv|
         first = data.first
         keys = fasta_mapper ? ['fasta_header'] : []
-        csv << (keys + first.keys).map(&:to_s) if first
+
+        keys += first.keys
+
+        %w[ec go].each do |annotation|
+          next unless keys.include?(annotation)
+
+          idx = keys.index(annotation)
+          keys.delete_at(idx)
+          keys.insert(idx, *first[annotation].first.keys.map { |el| el == 'protein_count' ? annotation + '_protein_count' : el })
+        end
+
+        csv << keys.map(&:to_s) if first
       end
     end
 
@@ -204,7 +215,17 @@ module Unipept
     def convert(data, _first)
       CSV.generate do |csv|
         data.each do |o|
-          csv << o.values.map { |v| v == '' ? nil : v }
+          row = []
+          o.each do |k, v|
+            if %w[ec go].include? k
+              v.first.keys.each do |key|
+                row << (v.map { |el| el[key] }).join(' ')
+              end
+            else
+              row << (v == '' ? nil : v)
+            end
+          end
+          csv << row
         end
       end
     end

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 1.4.1 ruby lib
+# stub: unipept 2.0.0 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "unipept".freeze
-  s.version = "1.4.1"
+  s.version = "2.0.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0".freeze) if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib".freeze]
   s.authors = ["Toon Willems".freeze, "Bart Mesuere".freeze, "Tom Naessens".freeze]
-  s.date = "2019-01-19"
+  s.date = "2019-06-19"
   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".freeze
   s.email = "unipept@ugent.be".freeze
   s.executables = ["unipept".freeze, "prot2pept".freeze, "peptfilter".freeze, "uniprot".freeze]
@@ -42,9 +42,13 @@ Gem::Specification.new do |s|
     "lib/commands/unipept.rb",
     "lib/commands/unipept/api_runner.rb",
     "lib/commands/unipept/config.rb",
+    "lib/commands/unipept/pept2ec.rb",
+    "lib/commands/unipept/pept2funct.rb",
+    "lib/commands/unipept/pept2go.rb",
     "lib/commands/unipept/pept2lca.rb",
     "lib/commands/unipept/pept2prot.rb",
     "lib/commands/unipept/pept2taxa.rb",
+    "lib/commands/unipept/peptinfo.rb",
     "lib/commands/unipept/taxa2lca.rb",
     "lib/commands/unipept/taxonomy.rb",
     "lib/commands/uniprot.rb",

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: unipept
 version: !ruby/object:Gem::Version
-  version: 1.4.1
+  version: 2.0.0
 platform: ruby
 authors:
 - Toon Willems
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-01-19 00:00:00.000000000 Z
+date: 2019-06-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cri
@@ -133,9 +133,13 @@ files:
 - lib/commands/unipept.rb
 - lib/commands/unipept/api_runner.rb
 - lib/commands/unipept/config.rb
+- lib/commands/unipept/pept2ec.rb
+- lib/commands/unipept/pept2funct.rb
+- lib/commands/unipept/pept2go.rb
 - lib/commands/unipept/pept2lca.rb
 - lib/commands/unipept/pept2prot.rb
 - lib/commands/unipept/pept2taxa.rb
+- lib/commands/unipept/peptinfo.rb
 - lib/commands/unipept/taxa2lca.rb
 - lib/commands/unipept/taxonomy.rb
 - lib/commands/uniprot.rb