gn_crossmap 3.0.3 → 3.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: fe8f372dc41915a9e7b577a5ec01453be24b5edb
-  data.tar.gz: a260cec5476a37645aaa68443204aa262ab4f3f4
+  metadata.gz: 540a1ea4dd47918c8d98c6de7e38f0f6cc116798
+  data.tar.gz: eca75610cb8974aced0dbed71a51c38c865c4a27
 SHA512:
-  metadata.gz: a86e7b8931b712bcb01d85c50e0ef41512e447bfec3884ce21b9193a23cbfa4c002032f29aafe6839f2f2c306e88f8f51a29458317629eec2b090f18f71876df
-  data.tar.gz: 4067fade993942853c1a80c58411f250798fe8aad17d299edf8a25c301cd4d794ff56e845cac74c25facf8928ade73f68bfa61fd73cd19537bb9991012b995bb
+  metadata.gz: 4f008fac59680bac5c8160f8a5e3f8caf22b801b67be088145b0ef1d8ed3657fa50f72d36960efac54410f11f2853cdd6354a1bd2bc7535d87e30d84558ad45e
+  data.tar.gz: ba229e5d5f0d954e9884a7bb31ace418eeb3c5131e77ebc70df945706d66990a042a8e73865ca925da5c16eb4dfd1ca021aba82a9e2ebed35447ce159990ff00

data/.rubocop.yml

@@ -3,6 +3,7 @@ AllCops:
     - db/**/*
     - bundle_bin/**/*
     - spec/**/*
+    - bin/**/*
   Include:
     - exe/crossmap
     - "**/Gemfile"

data/.ruby-version

@@ -0,0 +1 @@
+2.4.1

data/.travis.yml

@@ -1,7 +1,6 @@
 language: ruby
 rvm:
-  - 2.1
-  - 2.2
+  - 2.4
 script:
   - bundle exec rake
 branches:

data/CHANGELOG.md

@@ -1,5 +1,11 @@
 # ``gn_crossmap`` CHANGELOG
 
+## 3.1.0
+
+* @dimus - Fixes #34 add canonical form input
+
+* @dimus - Fixes #35 optionally returns classification path
+
 ## 3.0.3
 
 * @dimus - Fixes #33 infraspecies rank is given for all 'unknown' infra-specific

data/Gemfile

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 source "https://rubygems.org"
 
 # Specify your gem's dependencies in gn_crossmap.gemspec

data/README.md

@@ -59,6 +59,9 @@ cat my_list.csv | crossmap -i - -o - > output
 # to keep only taxonID (if given) from original input
 # no original fields will be kept without taxonID
 cat my_list.csv | crossmap -i my_list.csv -s
+
+# to show classification from the source
+cat my_list.csv | crossmap -i my_list.csv -w
 ```
 
 ### Usage as Ruby Library (API description)
@@ -97,6 +100,9 @@ instead of the headers supplied with the file
 ``resolver_url``
 : URL to globalnames' resolver. Default is ``http://resolver.globalnames.org``
 
+``with_classification``
+: (boolean) if true, adds classification path to the output
+
 #### `GnCrossmap.logger=`
 
 Allows to set logger to a custom logger (default is `STDERR`)

data/Rakefile

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require "bundler/gem_tasks"
 require "rspec/core/rake_task"
 require "rubocop/rake_task"

data/exe/crossmap

@@ -1,11 +1,13 @@
 #!/usr/bin/env ruby
+# frozen_string_literal: true
+
 require "trollop"
 require "gn_crossmap"
 
-puts "This program requires Ruby >= v. 2.1.0" if RUBY_VERSION < "2.1.0"
+puts "This program requires Ruby >= v. 2.4.1" if RUBY_VERSION < "2.4.1"
 
 CATALOGUE_OF_LIFE = 1
-OUTPUT = "output.csv".freeze
+OUTPUT = "output.csv"
 opts = Trollop.options do
   banner "Compares a list of scientific names to scientific names from a " \
     "data source from Global Names Resolver\n\n " \
@@ -18,6 +20,8 @@ opts = Trollop.options do
       default: CATALOGUE_OF_LIFE)
   opt(:skip_original, "If given, only 'taxonID' is shown " \
       "from the original input", type: :boolean)
+  opt(:with_classification, "If given, returns classification path of " \
+                            "matched names", type: :boolean)
 end
 
 Trollop.die :input, "must be set" if opts[:input].nil?

data/gn_crossmap.gemspec

@@ -1,4 +1,4 @@
-# coding: utf-8
+# frozen_string_literal: true
 
 lib = File.expand_path("../lib", __FILE__)
 
@@ -30,18 +30,18 @@ Gem::Specification.new do |gem|
   gem.executables   = gem.files.grep(%r{^exe/}) { |f| File.basename(f) }
   gem.require_paths = ["lib"]
 
-  gem.add_dependency "trollop", "~> 2.1"
   gem.add_dependency "biodiversity", "~> 3.1"
-  gem.add_dependency "rest-client", "~> 2.0"
-  gem.add_dependency "logger-colors", "~> 1.0"
   gem.add_dependency "gn_uuid", "~> 0.5"
+  gem.add_dependency "logger-colors", "~> 1.0"
+  gem.add_dependency "rest-client", "~> 2.0"
+  gem.add_dependency "trollop", "~> 2.1"
 
   gem.add_development_dependency "bundler", "~> 1.7"
-  gem.add_development_dependency "rake", "~> 11.0"
-  gem.add_development_dependency "rspec", "~> 3.2"
-  gem.add_development_dependency "rubocop", "~> 0.31"
-  gem.add_development_dependency "coveralls", "~> 0.8"
   gem.add_development_dependency "byebug", "~> 9.0"
+  gem.add_development_dependency "coveralls", "~> 0.8"
+  gem.add_development_dependency "rake", "~> 12.0"
+  gem.add_development_dependency "rspec", "~> 3.2"
+  gem.add_development_dependency "rubocop", "~> 0.49"
 end
 
 # rubocop:enable Metrics/BlockLength:

data/lib/gn_crossmap.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require "csv"
 require "ostruct"
 require "rest_client"
@@ -19,8 +21,8 @@ require "gn_crossmap/stats"
 
 # Namespace module for crossmapping checklists wth GN sources
 module GnCrossmap
-  INPUT_MODE = "r:utf-8".freeze
-  OUTPUT_MODE = "w:utf-8".freeze
+  INPUT_MODE = "r:utf-8"
+  OUTPUT_MODE = "w:utf-8"
   MATCH_TYPES = {
     0 => "No match",
     1 => "Exact string match",
@@ -65,11 +67,13 @@ module GnCrossmap
     private
 
     def create_resolver(writer, opts)
-      Resolver.new(writer, opts.data_source_id, opts.resolver_url, opts.stats)
+      Resolver.new(writer, opts.data_source_id, opts.resolver_url,
+                   opts.stats, opts.with_classification)
     end
 
     def create_writer(reader, output_io, opts)
-      Writer.new(output_io, reader.original_fields, output_name(opts.output))
+      Writer.new(output_io, reader.original_fields,
+                 output_name(opts.output), opts.with_classification)
     end
 
     def create_reader(input_io, opts)

data/lib/gn_crossmap/collector.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Assemble data from CSV reader by checking column fields
   class Collector

data/lib/gn_crossmap/column_collector.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Assemble data from CSV reader by checking column fields
   class ColumnCollector

data/lib/gn_crossmap/errors.rb

@@ -1,2 +1,4 @@
+# frozen_string_literal: true
+
 # Error to raise in case of problems
 class GnCrossmapError < RuntimeError; end

data/lib/gn_crossmap/reader.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Reads supplied csv file and creates ruby structure to compare
   # with a Global Names Resolver source
@@ -99,7 +101,7 @@ module GnCrossmap
 
     def taxon_id_header(hdrs)
       hdrs.each do |h|
-        return [h] if h =~ /taxonid\s*$/i
+        return [h] if h && h.match?(/taxonid\s*$/i)
       end
       []
     end

data/lib/gn_crossmap/resolver.rb

@@ -1,12 +1,16 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Sends data to GN Resolver and collects results
   class Resolver
     attr_reader :stats
 
-    def initialize(writer, data_source_id, resolver_url, stats)
+    def initialize(writer, data_source_id,
+                   resolver_url, stats, with_classification = false)
       @stats = stats
       @resolver_url = resolver_url
-      @processor = GnCrossmap::ResultProcessor.new(writer, @stats)
+      @processor = GnCrossmap::ResultProcessor.
+                   new(writer, @stats, with_classification)
       @ds_id = data_source_id
       @count = 0
       @current_data = {}
@@ -57,12 +61,12 @@ module GnCrossmap
 
     def collect_names(slice)
       @current_data = {}
-      slice.each_with_object("") do |row, str|
+      slice.each_with_object([]) do |row, str|
         id = row[:id].strip
         @current_data[id] = row[:original]
         @processor.input[id] = { rank: row[:rank] }
-        str << "#{id}|#{row[:name]}\n"
-      end
+        str << "#{id}|#{row[:name]}"
+      end.join("\n")
     end
 
     def remote_resolve(names)

data/lib/gn_crossmap/result_processor.rb

@@ -1,9 +1,13 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Processes data received from the GN Resolver
   class ResultProcessor
     attr_reader :input, :writer
 
-    def initialize(writer, stats)
+    def initialize(writer, stats, with_classification = false)
+      @with_classification = with_classification
+      @parser = ScientificNameParser.new
       @stats = stats
       @writer = writer
       @input = {}
@@ -28,8 +32,8 @@ module GnCrossmap
       @stats.stats[:resolved_records] += 1
       res = @original_data[datum[:supplied_id]]
       res += [GnCrossmap::MATCH_TYPES[0], datum[:supplied_name_string], nil,
-              nil, @input[datum[:supplied_id]][:rank], nil,
-              nil, nil, nil]
+              datum[:supplied_canonical_form], nil,
+              @input[datum[:supplied_id]][:rank], nil, nil, nil, nil]
       @writer.write(res)
     end
 
@@ -50,13 +54,28 @@ module GnCrossmap
       @original_data[datum[:supplied_id]] + new_data(datum, result)
     end
 
+    # rubocop:disable Metrics/AbcSize
+
     def new_data(datum, result)
       synonym = result[:current_name_string] ? "synonym" : nil
-      [matched_type(result), datum[:supplied_name_string],
-       result[:name_string], result[:canonical_form],
-       @input[datum[:supplied_id]][:rank], matched_rank(result),
-       synonym, result[:current_name_string] || result[:name_string],
-       result[:edit_distance], result[:score], result[:taxon_id]]
+      res = [matched_type(result), datum[:supplied_name_string],
+             result[:name_string], canonical(datum[:supplied_name_string]),
+             result[:canonical_form], @input[datum[:supplied_id]][:rank],
+             matched_rank(result), synonym,
+             result[:current_name_string] || result[:name_string],
+             result[:edit_distance], result[:score], result[:taxon_id]]
+      res << classification(result) if @with_classification
+      res
+    end
+
+    # rubocop:enable all
+
+    def canonical(name_string)
+      parsed = @parser.parse(name_string)[:scientificName]
+      parsed[:canonical].nil? || parsed[:hybrid] ? nil : parsed[:canonical]
+    rescue StandardError
+      @parser = ScientificNameParser.new
+      nil
     end
 
     def matched_rank(record)
@@ -66,5 +85,19 @@ module GnCrossmap
     def matched_type(record)
       GnCrossmap::MATCH_TYPES[record[:match_type]]
     end
+
+    # rubocop:disable Metrics/AbcSize
+
+    def classification(result)
+      return nil if result[:classification_path].to_s.strip == ""
+      path = result[:classification_path].split("|")
+      ranks = result[:classification_path_ranks].split("|")
+      if path.size == ranks.size
+        path = path.zip(ranks).map { |e| "#{e[0]}(#{e[1]})" }
+      end
+      path.join(", ")
+    end
+
+    # rubocop:enable all
   end
 end

data/lib/gn_crossmap/sci_name_collector.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Assemble data from CSV reader by parsing scientificName field
   class SciNameCollector

data/lib/gn_crossmap/version.rb

@@ -1,6 +1,8 @@
+# frozen_string_literal: true
+
 # Namespace module for crossmapping checklists to GN sources
 module GnCrossmap
-  VERSION = "3.0.3".freeze
+  VERSION = "3.1.0"
 
   def self.version
     VERSION

data/lib/gn_crossmap/writer.rb

@@ -1,9 +1,13 @@
+# frozen_string_literal: true
+
 module GnCrossmap
   # Saves output from GN Resolver to disk
   class Writer
-    def initialize(output_io, original_fields, output_name)
+    def initialize(output_io, original_fields, output_name,
+                   with_classification = false)
       @output_io = output_io
       @output_fields = output_fields(original_fields)
+      @output_fields << :classification if with_classification
       @output = CSV.new(@output_io, col_sep: "\t")
       @output << @output_fields
       @output_name = output_name
@@ -22,7 +26,7 @@ module GnCrossmap
     private
 
     def output_fields(original_fields)
-      original_fields + %i[matchedType inputName matchedName
+      original_fields + %i[matchedType inputName matchedName inputCanonicalForm
                            matchedCanonicalForm inputRank matchedRank
                            synonymStatus acceptedName matchedEditDistance
                            matchedScore matchTaxonID]

metadata

@@ -1,85 +1,85 @@
 --- !ruby/object:Gem::Specification
 name: gn_crossmap
 version: !ruby/object:Gem::Version
-  version: 3.0.3
+  version: 3.1.0
 platform: ruby
 authors:
 - Dmitry Mozzherin
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-07-26 00:00:00.000000000 Z
+date: 2017-08-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: trollop
+  name: biodiversity
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.1'
+        version: '3.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.1'
+        version: '3.1'
 - !ruby/object:Gem::Dependency
-  name: biodiversity
+  name: gn_uuid
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.1'
+        version: '0.5'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.1'
+        version: '0.5'
 - !ruby/object:Gem::Dependency
-  name: rest-client
+  name: logger-colors
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '1.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '1.0'
 - !ruby/object:Gem::Dependency
-  name: logger-colors
+  name: rest-client
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '2.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.0'
+        version: '2.0'
 - !ruby/object:Gem::Dependency
-  name: gn_uuid
+  name: trollop
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.5'
+        version: '2.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.5'
+        version: '2.1'
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -95,75 +95,75 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '1.7'
 - !ruby/object:Gem::Dependency
-  name: rake
+  name: byebug
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '11.0'
+        version: '9.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '11.0'
+        version: '9.0'
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: coveralls
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.2'
+        version: '0.8'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.2'
+        version: '0.8'
 - !ruby/object:Gem::Dependency
-  name: rubocop
+  name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.31'
+        version: '12.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.31'
+        version: '12.0'
 - !ruby/object:Gem::Dependency
-  name: coveralls
+  name: rspec
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '3.2'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '3.2'
 - !ruby/object:Gem::Dependency
-  name: byebug
+  name: rubocop
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '9.0'
+        version: '0.49'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '9.0'
+        version: '0.49'
 description: Gem uses a checklist in a comma-separated format as an input, and returns
   back a new comma-separated list crossmapping the scientific names to one of the
   data sources from http://resolver.globalnames.org
@@ -177,6 +177,7 @@ files:
 - ".gitignore"
 - ".rspec"
 - ".rubocop.yml"
+- ".ruby-version"
 - ".travis.yml"
 - CHANGELOG.md
 - Gemfile