termium 0.1.4 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5b855f913eef90fa77b61f7d6c0be26b202c4cfc140bc92d9b5ad81bc5b35a15
-  data.tar.gz: 25b5dbfbc1854402fdd909f40692b1170b27626b42eba0fcad9b1486c9d8ca4f
+  metadata.gz: 86c7b7ab35fb63b5d7bc8a085125db1452284327970939abd89808bf6e628097
+  data.tar.gz: 2229305a3891ae26ec4b8ec82ad3ae55735912c7503c19ed5c478c52d3c1db68
 SHA512:
-  metadata.gz: 2e7ffe3b5666f22dcd10884b174257f72c21959858695419e86bb8cfea9965857ceea92474f395264e8e3c2e23a31e0d96b7c476c1412050732c85507f5103df
-  data.tar.gz: a32cd300c99f24a8a88d6f7acfef6f3e4e1b7b892fa7db4ff48fe5d945fb666cf011f77a567ee537349ee09055e6441f2d7906cc90294a5850c0d567efe1c7da
+  metadata.gz: 67d16aa7301c765ec97865d5eff33b98e02de93d8cce9332fb7d8953613fa638719941bc2e5b695a29ae4f0e8e4960864d4ad17a8d1ee8332bc66e877f81c099
+  data.tar.gz: 97d56933d5946f70209e0c3b054583e683e2e510ff235615d324dbb80c3f76dcdc1acea3fb1fd3003b9f1c417b193636df701685c85dfee0f2f33aee4173bc17

data/.github/workflows/rake.yml

@@ -0,0 +1,13 @@
+# Auto-generated by Cimas: Do not edit it manually!
+# See https://github.com/metanorma/cimas
+name: rake
+
+on:
+  push:
+    branches: [ master, main ]
+    tags: [ v* ]
+  pull_request:
+
+jobs:
+  rake:
+    uses: relaton/support/.github/workflows/rake.yml@main

data/.github/workflows/{release.yaml → release.yml}

@@ -1,3 +1,5 @@
+# Auto-generated by Cimas: Do not edit it manually!
+# See https://github.com/metanorma/cimas
 name: release
 
 on:
@@ -5,17 +7,18 @@ on:
     inputs:
       next_version:
         description: |
-          Next release version. Possible values: x.y.z, major, minor, patch or pre|rc|etc
+          Next release version. Possible values: x.y.z, major, minor, patch (or pre|rc|etc).
+          Also, you can pass 'skip' to skip 'git tag' and do 'gem push' for the current version
         required: true
         default: 'skip'
   repository_dispatch:
     types: [ do-release ]
 
+
 jobs:
   release:
-    uses: metanorma/ci/.github/workflows/rubygems-release.yml@main
+    uses: relaton/support/.github/workflows/release.yml@main
     with:
       next_version: ${{ github.event.inputs.next_version }}
     secrets:
       rubygems-api-key: ${{ secrets.GLOSSARIST_CI_RUBYGEMS_API_KEY }}
-      pat_token: ${{ secrets.GLOSSARIST_CI_PAT_TOKEN }}

data/Gemfile

@@ -3,3 +3,11 @@
 source "https://rubygems.org"
 
 gemspec
+
+gem "equivalent-xml"
+gem "pry"
+gem "rake", "~> 13.0"
+gem "rspec", "~> 3.11"
+gem "rubocop", "~> 1.58"
+gem "rubocop-performance", "~> 1.21", ">= 1.21.1"
+gem "xml-c14n"

data/README.adoc

@@ -6,6 +6,39 @@ The Termium Ruby gem parses export data formats from the
 https://www.btb.termiumplus.gc.ca[TERMIUM Plus]
 terminology database service from the Government of Canada.
 
+== WARNING - Termium XML output requires manual correction
+
+The default Termium XML output is invalid where the term domains using angular
+brackets have the "greater than" sign not escaped:
+
+[source,xml]
+----
+<textualSupport order="1" type="DEF">
+  <value>&lt;artificial intelligence> operation that allows the firing of a rule, or the
+    invocation of a program or a subprogram</value>
+  <sourceRef order="1" />
+</textualSupport>
+----
+
+The remedy is to manually escape the "greater than" sign using a find/replace or a regular expression:
+
+[source,ruby]
+----
+string.gsub(/&lt;([^>]+)>/, '<\1>')
+----
+
+Results in:
+
+[source,xml]
+----
+<textualSupport order="1" type="DEF">
+  <value>&lt;artificial intelligence&gt; operation that allows the firing of a rule, or the
+    invocation of a program or a subprogram</value>
+  <sourceRef order="1" />
+</textualSupport>
+----
+
+
 
 == Commands
 
@@ -43,7 +76,7 @@ If not provided, defaults to the basename of the input file, e.g. `foo/bar.xml`
 
 === Usage
 
-This gem makes heavy use of the `Shale` classes for XML serialization.
+This gem makes heavy use of the `lutaml-model` classes for XML serialization.
 
 The following code converts the Termium extract into a Glossarist dataset.
 

data/lib/termium/abbreviation.rb

@@ -6,9 +6,9 @@ require_relative "designation_operations"
 
 module Termium
   # For <abbreviation>
-  class Abbreviation < Shale::Mapper
-    attribute :order, Shale::Type::Integer
-    attribute :value, Shale::Type::String
+  class Abbreviation < Lutaml::Model::Serializable
+    attribute :order, :integer
+    attribute :value, :string
     attribute :source_ref, SourceRef
     attribute :parameter, Parameter, collection: true
     include DesignationOperations
@@ -35,7 +35,7 @@ module Termium
       set = {
         "designation" => value,
         "type" => "abbreviation",
-        "normative_status" => deprecated ? "deprecated" : "preferred"
+        "normative_status" => deprecated ? "deprecated" : "preferred",
       }
 
       # if geographical_area

data/lib/termium/cli.rb

@@ -9,6 +9,7 @@ module Termium
 
     option :input_file, aliases: :i, required: true, desc: "Path to TERMIUM Plus XML extract"
     option :output_file, aliases: :o, desc: "Output file path"
+    option :date_accepted, desc: "Date of acceptance for the dataset"
 
     no_commands do
       def input_file_as_path(input_file)
@@ -50,7 +51,11 @@ module Termium
       puts "Size of TERMIUM dataset: #{termium_extract.core.size}"
 
       puts "Converting to Glossarist..."
-      glossarist_col = termium_extract.to_concept
+      convert_options = {}
+      if options[:date_accepted]
+        convert_options[:date_accepted] = Date.parse(options[:date_accepted])
+      end
+      glossarist_col = termium_extract.to_concept(convert_options)
       # pp glossarist_col.first
 
       output_path = output_dir_as_path(options[:output_file], input_path)

data/lib/termium/core.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "uuidtools"
 require_relative "language_module"
 require_relative "subject"
 require_relative "universal_entry"
@@ -7,13 +8,13 @@ require_relative "source"
 
 module Termium
   # For <core>
-  class Core < Shale::Mapper
-    attribute :identification_number, Shale::Type::String
-    attribute :dissemination_level, Shale::Type::String
+  class Core < Lutaml::Model::Serializable
+    attribute :identification_number, :string
+    attribute :dissemination_level, :string
 
     attribute :language_module, LanguageModule, collection: true
     attribute :subject, Subject
-    attribute :universal_entry, UniversalEntry
+    attribute :universal_entry, UniversalEntry, collection: true
     attribute :source, Source, collection: true
 
     xml do
@@ -36,25 +37,41 @@ module Termium
       source.map(&:to_concept_source)
     end
 
+    # Deterministic v4 UUID by using the number string
+    def uuid
+      UUIDTools::UUID.md5_create(UUIDTools::UUID_DNS_NAMESPACE, identification_number).to_s
+    end
+
     # TODO: Utilize "subject" in the Glossarist object:
     # <subject abbreviation="YBB"
     # details="Compartment - ISO/IEC JTC 1 Information Technology Vocabulary" />
-    def to_concept
-      concept = Glossarist::ManagedConcept.new(id: identification_number)
-
-      language_module.map(&:to_concept).each do |localized_concept|
-        # TODO: This is needed to skip the empty french entries of 10031781 and 10031778
-        next if localized_concept.nil?
-
-        localized_concept.id = identification_number
-        # TODO: this should just be localized_concept.notes << universal_entry.value
-        # TODO: Depends on https://github.com/glossarist/glossarist-ruby/issues/82
-        localized_concept.notes << Glossarist::DetailedDefinition.new(universal_entry.value)
-        localized_concept.sources = concept_sources
-        concept.add_localization(localized_concept)
-      end
+    def to_concept(options = {})
+      Glossarist::ManagedConcept.new.tap do |concept|
+        # The way to set the universal concept's identifier: data.identifier
+        concept.id = identification_number
+
+        concept.uuid = uuid
 
-      concept
+        # Assume no related concepts
+        concept.related = []
+        concept.status = "valid"
+
+        if options[:date_accepted]
+          concept.date_accepted = options[:date_accepted]
+        end
+
+        language_module.map(&:to_concept).each do |localized_concept|
+          # TODO: This is needed to skip the empty french entries of 10031781 and 10031778
+          next if localized_concept.nil?
+
+          localized_concept.id = identification_number
+          universal_entry.each do |entry|
+            localized_concept.notes << entry.value
+          end
+          localized_concept.sources = concept_sources
+          concept.add_localization(localized_concept)
+        end
+      end
     end
   end
 end

data/lib/termium/entry_term.rb

@@ -7,9 +7,9 @@ require_relative "designation_operations"
 
 module Termium
   # For <entryTerm>
-  class EntryTerm < Shale::Mapper
-    attribute :order, Shale::Type::Integer
-    attribute :value, Shale::Type::String
+  class EntryTerm < Lutaml::Model::Serializable
+    attribute :order, :integer
+    attribute :value, :string
     attribute :source_ref, SourceRef
     attribute :abbreviation, Abbreviation, collection: true
     attribute :parameter, Parameter, collection: true
@@ -19,9 +19,9 @@ module Termium
       root "entryTerm"
       map_attribute "order", to: :order
       map_attribute "value", to: :value
-      map_element "abbreviation", to: :abbreviation
       map_element "sourceRef", to: :source_ref
       map_element "parameter", to: :parameter
+      map_element "abbreviation", to: :abbreviation
     end
 
     # attr_accessor :geographical_area,
@@ -35,8 +35,9 @@ module Termium
       "CAN" => "CA",
       "GB" => "GB",
       "AUS" => "AU",
-      "EUR" => "EU"
+      "EUR" => "EU",
     }.freeze
+
     def geographical_area
       keys = GEOGRAPHICAL_CODE_MAPPING.keys
       usage = parameter.select do |x|
@@ -72,7 +73,7 @@ module Termium
       set = {
         "designation" => value,
         "type" => "expression",
-        "normative_status" => normative_status
+        "normative_status" => normative_status,
       }
 
       set["geographical_area"] = geographical_area if geographical_area

data/lib/termium/extract.rb

@@ -5,23 +5,34 @@ require_relative "core"
 
 module Termium
   # For <extract>
-  class Extract < Shale::Mapper
-    attribute :language, Shale::Type::String
+  class Extract < Lutaml::Model::Serializable
+    attribute :language, :string
     attribute :extract_language, ExtractLanguage, collection: true
     attribute :core, Core, collection: true
 
     xml do
       root "termium_extract"
-      # namespace 'http://termium.tpsgc-pwgsc.gc.ca/schemas/2012/06/Termium', 'ns2'
+      # namespace "http://termium.tpsgc-pwgsc.gc.ca/schemas/2012/06/Termium", "ns2"
 
       map_attribute "language", to: :language
       map_element "extractLanguage", to: :extract_language
       map_element "core", to: :core
     end
 
-    def to_concept
+    # xml do
+    #   root "termium_extract"
+    #   namespace "http://termium.tpsgc-pwgsc.gc.ca/schemas/2012/06/Termium", "ns2"
+
+    #   map_attribute "language", to: :language, namespace: nil
+    #   map_element "extractLanguage", to: :extract_language, namespace: nil
+    #   map_element "core", to: :core, namespace: nil
+    # end
+
+    def to_concept(options = {})
       coll = Glossarist::ManagedConceptCollection.new
-      coll.managed_concepts = core.map(&:to_concept)
+      coll.managed_concepts = core.map do |managed_concept|
+        managed_concept.to_concept(options)
+      end
       coll
     end
   end

data/lib/termium/extract_language.rb

@@ -2,9 +2,9 @@
 
 module Termium
   # For <extractLanguage>
-  class ExtractLanguage < Shale::Mapper
-    attribute :language, Shale::Type::String
-    attribute :order, Shale::Type::Integer
+  class ExtractLanguage < Lutaml::Model::Serializable
+    attribute :language, :string
+    attribute :order, :integer
     xml do
       root "extractLanguage"
       map_attribute "language", to: :language

data/lib/termium/language_module.rb

@@ -5,8 +5,8 @@ require_relative "textual_support"
 
 module Termium
   # For <languageModule>
-  class LanguageModule < Shale::Mapper
-    attribute :language, Shale::Type::String
+  class LanguageModule < Lutaml::Model::Serializable
+    attribute :language, :string
     attribute :entry_term, EntryTerm, collection: true
     attribute :textual_support, TextualSupport, collection: true
     xml do
@@ -42,7 +42,7 @@ module Termium
 
     LANGUAGE_CODE_MAPPING = {
       "en" => "eng",
-      "fr" => "fre"
+      "fr" => "fre",
     }.freeze
 
     def designations
@@ -59,7 +59,7 @@ module Termium
         "terms" => designations.map(&:to_h),
         "definition" => [{ content: definition }],
         "notes" => notes,
-        "examples" => examples
+        "examples" => examples,
       }
 
       src["domain"] = domain if domain

data/lib/termium/parameter.rb

@@ -2,9 +2,9 @@
 
 module Termium
   # For <parameter>
-  class Parameter < Shale::Mapper
+  class Parameter < Lutaml::Model::Serializable
     # <parameter abbreviation="NORM"/>
-    attribute :abbreviation, Shale::Type::String
+    attribute :abbreviation, :string
     xml do
       root "parameter"
       map_attribute "abbreviation", to: :abbreviation

data/lib/termium/source.rb

@@ -2,12 +2,12 @@
 
 module Termium
   # For <source>
-  class Source < Shale::Mapper
+  class Source < Lutaml::Model::Serializable
     ISO_BIB_REGEX = /\AISO-([\d-]+)\s+\*\s+(\d{4})\s+.*/.freeze
     ISOIEC_BIB_REGEX = /\AISO-IEC-([\d-]+)\s+\*\s+(\d{4})\s+.*/.freeze
 
-    attribute :order, Shale::Type::Integer
-    attribute :details, Shale::Type::String
+    attribute :order, :integer
+    attribute :details, :string
     xml do
       root "source"
       map_attribute "order", to: :order
@@ -28,7 +28,7 @@ module Termium
       Glossarist::ConceptSource.new({
                                       "type" => "lineage",
                                       "ref" => content,
-                                      "status" => "identical"
+                                      "status" => "identical",
                                     })
     end
   end

data/lib/termium/source_ref.rb

@@ -2,8 +2,8 @@
 
 module Termium
   # For <sourceRef>
-  class SourceRef < Shale::Mapper
-    attribute :order, Shale::Type::Integer
+  class SourceRef < Lutaml::Model::Serializable
+    attribute :order, :integer
     xml do
       root "sourceRef"
       map_attribute "order", to: :order

data/lib/termium/subject.rb

@@ -2,9 +2,9 @@
 
 module Termium
   # For <subject>
-  class Subject < Shale::Mapper
-    attribute :abbreviation, Shale::Type::String
-    attribute :details, Shale::Type::String
+  class Subject < Lutaml::Model::Serializable
+    attribute :abbreviation, :string
+    attribute :details, :string
 
     # <subject abbreviation="YBB" details="Compartment - ISO/IEC JTC 1 Information Technology Vocabulary"/>
     xml do

data/lib/termium/textual_support.rb

@@ -4,10 +4,10 @@ require_relative "source_ref"
 
 module Termium
   # For <textualSupport>
-  class TextualSupport < Shale::Mapper
-    attribute :order, Shale::Type::Integer
-    attribute :type, Shale::Type::String
-    attribute :value, Shale::Type::String
+  class TextualSupport < Lutaml::Model::Serializable
+    attribute :order, :integer
+    attribute :type, :string
+    attribute :value, :string
     attribute :source_ref, SourceRef
     xml do
       root "textualSupport"
@@ -32,6 +32,7 @@ module Termium
     end
 
     EXAMPLE_REGEX = /\AEx[ea]mples?\s*:\s*/.freeze
+
     def is_example?
       value_cleaned.match(EXAMPLE_REGEX)
     end
@@ -53,6 +54,7 @@ module Termium
     end
 
     DEFINITION_REGEX = /\A<(.+?)>\s*/.freeze
+
     def value_definition
       value_cleaned.gsub(DEFINITION_REGEX, "")
     end

data/lib/termium/universal_entry.rb

@@ -5,9 +5,9 @@ require_relative "parameter"
 
 module Termium
   # For <universalEntry>
-  class UniversalEntry < Shale::Mapper
-    attribute :order, Shale::Type::Integer
-    attribute :value, Shale::Type::String
+  class UniversalEntry < Lutaml::Model::Serializable
+    attribute :order, :integer
+    attribute :value, :string
     attribute :source_ref, SourceRef
     attribute :parameter, Parameter
 

data/lib/termium/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Termium
-  VERSION = "0.1.4"
+  VERSION = "0.3.1"
 end

data/lib/termium.rb

@@ -1,9 +1,13 @@
 # frozen_string_literal: true
 
 require "glossarist"
-require "shale"
-require "shale/adapter/nokogiri"
-Shale.xml_adapter = Shale::Adapter::Nokogiri
+
+require "lutaml/model"
+require "lutaml/model/xml_adapter/nokogiri_adapter"
+
+Lutaml::Model::Config.configure do |config|
+  config.xml_adapter = Lutaml::Model::XmlAdapter::NokogiriAdapter
+end
 
 module Termium
   class Error < StandardError; end

data/termium.gemspec

@@ -12,10 +12,10 @@ Gem::Specification.new do |spec|
   spec.authors = ["Ribose"]
   spec.email = ["open.source@ribose.com"]
 
-  spec.summary       =
+  spec.summary =
     "Parser for the TERMIUM Plus terminology database of the Government of Canada"
-  spec.homepage      = "https://github.com/glossarist/termium"
-  spec.license       = "BSD-2-Clause"
+  spec.homepage = "https://github.com/glossarist/termium"
+  spec.license = "BSD-2-Clause"
   spec.required_ruby_version = Gem::Requirement.new(">= 2.6.0")
 
   spec.metadata["homepage_uri"] = spec.homepage
@@ -23,19 +23,15 @@ Gem::Specification.new do |spec|
   spec.metadata["bug_tracker_uri"] = "#{spec.homepage}/issues"
 
   # Specify which files should be added to the gem when it is released.
-  spec.files         = all_files_in_git
-                       .reject { |f| f.match(%r{\A(?:test|spec|features|bin|\.)/}) }
+  spec.files = all_files_in_git
+    .reject { |f| f.match(%r{\A(?:test|spec|features|bin|\.)/}) }
 
-  spec.bindir        = "exe"
-  spec.executables   = spec.files.grep(%r{\Aexe/}) { |f| File.basename(f) }
+  spec.bindir = "exe"
+  spec.executables = spec.files.grep(%r{\Aexe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]
 
-  spec.add_dependency "glossarist", "~> 2.0"
-  spec.add_dependency "shale"
+  spec.add_dependency "glossarist", "~> 2.2"
+  spec.add_dependency "lutaml-model"
   spec.add_dependency "thor"
-
-  spec.add_development_dependency "pry", "~> 0.14.0"
-  spec.add_development_dependency "rake", "~> 13.0"
-  spec.add_development_dependency "rspec", "~> 3.10"
-  spec.add_development_dependency "rubocop"
+  spec.add_dependency "uuidtools"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: termium
 version: !ruby/object:Gem::Version
-  version: 0.1.4
+  version: 0.3.1
 platform: ruby
 authors:
 - Ribose
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2023-12-12 00:00:00.000000000 Z
+date: 2024-11-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: glossarist
@@ -16,16 +16,16 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '2.2'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.0'
+        version: '2.2'
 - !ruby/object:Gem::Dependency
-  name: shale
+  name: lutaml-model
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -53,55 +53,13 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: pry
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: 0.14.0
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: 0.14.0
-- !ruby/object:Gem::Dependency
-  name: rake
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '13.0'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '13.0'
-- !ruby/object:Gem::Dependency
-  name: rspec
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.10'
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '3.10'
-- !ruby/object:Gem::Dependency
-  name: rubocop
+  name: uuidtools
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-  type: :development
+  type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
@@ -116,8 +74,8 @@ executables:
 extensions: []
 extra_rdoc_files: []
 files:
-- ".github/workflows/main.yml"
-- ".github/workflows/release.yaml"
+- ".github/workflows/rake.yml"
+- ".github/workflows/release.yml"
 - ".gitignore"
 - ".rspec"
 - CODE_OF_CONDUCT.md
@@ -165,7 +123,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.26
+rubygems_version: 3.3.27
 signing_key: 
 specification_version: 4
 summary: Parser for the TERMIUM Plus terminology database of the Government of Canada

data/.github/workflows/main.yml

@@ -1,27 +0,0 @@
-name: Ruby
-
-on:
-  push:
-    branches:
-      - main
-
-  pull_request:
-
-jobs:
-  build:
-    runs-on: ubuntu-latest
-    name: Ruby ${{ matrix.ruby }}
-    strategy:
-      matrix:
-        ruby:
-          - '3.1.2'
-
-    steps:
-    - uses: actions/checkout@v3
-    - name: Set up Ruby
-      uses: ruby/setup-ruby@v1
-      with:
-        ruby-version: ${{ matrix.ruby }}
-        bundler-cache: true
-    - name: Run the default task
-      run: bundle exec rake