namae 0.11.2 → 1.1.0

data/lib/namae/parser.y

@@ -3,7 +3,7 @@
 
 class Namae::Parser
 
-token COMMA UWORD LWORD PWORD NICK AND APPELLATION TITLE SUFFIX
+token COMMA UWORD LWORD PWORD NICK AND APPELLATION TITLE SUFFIX UPARTICLE
 
 expect 0
 
@@ -20,7 +20,7 @@ rule
        | sort_order
 
   honorific : APPELLATION { result = Name.new(:appellation => val[0]) }
-            | TITLE       { result = Name.new(:title => val[0]) }
+            | titles      { result = Name.new(:title => val[0]) }
 
   display_order : u_words word opt_suffices opt_titles
        {
@@ -43,6 +43,14 @@ rule
          result = Name.new(:given => val[0], :particle => val[1],
           :family => val[2])
        }
+       | u_words UPARTICLE last
+       {
+          result = if include_particle_in_family?
+                    Name.new(:given => val[0], :family => val[1,2].join(' '))
+                   else
+                     Name.new(:given => val[0], :particle => val[1], :family => val[2])
+                   end
+       }
        | von last
        {
          result = Name.new(:particle => val[0], :family => val[1])
@@ -53,6 +61,14 @@ rule
          result = Name.new({ :family => val[0], :suffix => val[2][0],
            :given => val[2][1] }, !!val[2][0])
        }
+       | UPARTICLE last COMMA first
+       {
+         result = if include_particle_in_family?
+                    Name.new({ :family => val[0,2].join(' '), :suffix => val[3][0], :given => val[3][1] }, !!val[3][0])
+                  else
+                    Name.new({ :particle => val[0], :family => val[1], :suffix => val[3][0], :given => val[3][1] }, !!val[3][0])
+                  end
+       }
        | von last COMMA first
        {
          result = Name.new({ :particle => val[0], :family => val[1],
@@ -100,26 +116,35 @@ rule
          | titles TITLE { result = val.join(' ') }
 
 ---- header
-require 'singleton'
 require 'strscan'
 
 ---- inner
 
-  include Singleton
+  @defaults = {
+    :debug => false,
+    :prefer_comma_as_separator => false,
+    :include_particle_in_family => false,
+    :comma => ',',
+    :stops => ',;',
+    :separator => /\s*(\band\b|\&|;)\s*/i,
+    :title => /\s*\b(sir|lord|count(ess)?|(gen|adm|col|maj|capt|cmdr|lt|sgt|cpl|pvt|pastor|pr|reverend|rev|elder|deacon|deaconess|father|fr|rabbi|cantor|vicar|prof|dr|md|ph\.?d)\.?)(\s+|$)/i,
+    :suffix => /\s*\b(JR|Jr|jr|SR|Sr|sr|[IVX]{2,})(\.|\b)/,
+    :appellation => /\s*\b((mrs?|ms|fr|hr)\.?|miss|herr|frau)(\s+|$)/i,
+    :uppercase_particle => /\s*\b((Da|De|Di|De\sLa|Du|Der|Des|Da|St|Saint|Les|Van)\.?)(\s+|$)/
+  }
+
+  class << self
+    attr_reader :defaults
+
+    def instance
+      Thread.current[:namae] ||= new
+    end
+  end
 
   attr_reader :options, :input
 
-  def initialize
-    @input, @options = StringScanner.new(''), {
-      :debug => false,
-      :prefer_comma_as_separator => false,
-      :comma => ',',
-      :stops => ',;',
-      :separator => /\s*(\band\b|\&|;)\s*/i,
-      :title => /\s*\b(sir|lord|count(ess)?|(gen|adm|col|maj|capt|cmdr|lt|sgt|cpl|pvt|pastor|pr|reverend|rev|elder|deacon|deaconess|father|fr|vicar|prof|dr|md|ph\.?d)\.?)(\s+|$)/i,
-      :suffix => /\s*\b(JR|Jr|jr|SR|Sr|sr|[IVX]{2,})(\.|\b)/,
-      :appellation => /\s*\b((mrs?|ms|fr|hr)\.?|miss|herr|frau)(\s+|$)/i
-    }
+  def initialize(options = {})
+    @options = self.class.defaults.merge(options)
   end
 
   def debug?
@@ -134,6 +159,10 @@ require 'strscan'
     options[:comma]
   end
 
+  def include_particle_in_family?
+    options[:include_particle_in_family]
+  end
+
   def stops
     options[:stops]
   end
@@ -150,26 +179,29 @@ require 'strscan'
     options[:appellation]
   end
 
+  def uppercase_particle
+    options[:uppercase_particle]
+  end
+
   def prefer_comma_as_separator?
     options[:prefer_comma_as_separator]
   end
 
-  def parse(input)
-    parse!(input)
+  def parse(string)
+    parse!(string)
   rescue => e
     warn e.message if debug?
     []
   end
 
   def parse!(string)
-    input.string = normalize(string)
+    @input = StringScanner.new(normalize(string))
     reset
     do_parse
   end
 
   def normalize(string)
-    string = string.strip
-    string
+    string.scrub.strip
   end
 
   def reset
@@ -220,11 +252,11 @@ require 'strscan'
   end
 
   def will_see_suffix?
-    input.peek(8).to_s.strip.split(/\s+/)[0] =~ suffix
+    input.rest.strip.split(/\s+/)[0] =~ suffix
   end
 
   def will_see_initial?
-    input.peek(6).to_s.strip.split(/\s+/)[0] =~ /^[[:upper:]]+\b/
+    input.rest.strip.split(/\s+/)[0] =~ /^[[:upper:]]+\b/
   end
 
   def seen_full_name?
@@ -256,6 +288,8 @@ require 'strscan'
       else
         consume_word(:UWORD, input.matched)
       end
+    when input.scan(uppercase_particle)
+      consume_word(:UPARTICLE, input.matched.strip)
     when input.scan(/((\\\w+)?\{[^\}]*\})*[[:upper:]][^\s#{stops}]*/)
       consume_word(:UWORD, input.matched)
     when input.scan(/((\\\w+)?\{[^\}]*\})*[[:lower:]][^\s#{stops}]*/)

data/lib/namae/utility.rb

@@ -44,4 +44,8 @@ module Namae
     Parser.instance.options
   end
 
+  # @yield [Hash] the parser's default configuration.
+  def configure
+    yield Parser.defaults
+  end
 end

data/lib/namae/version.rb

@@ -1,8 +1,8 @@
 module Namae
   module Version
-    MAJOR = 0
-    MINOR = 11
-    PATCH = 2
+    MAJOR = 1
+    MINOR = 1
+    PATCH = 0
     BUILD = nil
 
     STRING = [MAJOR, MINOR, PATCH, BUILD].compact.join('.').freeze

data/namae.gemspec

@@ -2,23 +2,22 @@
 # DO NOT EDIT THIS FILE DIRECTLY
 # Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
 # -*- encoding: utf-8 -*-
-# stub: namae 0.11.2 ruby lib
+# stub: namae 1.1.0 ruby lib
 
 Gem::Specification.new do |s|
   s.name = "namae".freeze
-  s.version = "0.11.2"
+  s.version = "1.1.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0".freeze) if s.respond_to? :required_rubygems_version=
   s.require_paths = ["lib".freeze]
   s.authors = ["Sylvester Keil".freeze, "Dan Collis-Puro".freeze]
-  s.date = "2016-11-12"
-  s.description = " Namae (\u{540d}\u{524d}) is a parser for human names. It recognizes personal names of various cultural backgrounds and tries to split them into their component parts (e.g., given and family names, honorifics etc.). ".freeze
+  s.date = "2021-03-12"
+  s.description = " Namae (\u540D\u524D) is a parser for human names. It recognizes personal names of various cultural backgrounds and tries to split them into their component parts (e.g., given and family names, honorifics etc.). ".freeze
   s.email = ["sylvester@keil.or.at".freeze, "dan@collispuro.com".freeze]
   s.extra_rdoc_files = [
     "README.md"
   ]
   s.files = [
-    ".autotest",
     ".codeclimate.yml",
     ".coveralls.yml",
     ".document",
@@ -50,23 +49,22 @@ Gem::Specification.new do |s|
     "spec/namae/name_spec.rb",
     "spec/namae/parser_spec.rb",
     "spec/namae/utility_spec.rb",
-    "spec/spec_helper.rb"
+    "spec/spec_helper.rb",
+    "spec/thread_safety_spec.rb"
   ]
   s.homepage = "https://github.com/berkmancenter/namae".freeze
   s.licenses = ["AGPL-3.0".freeze]
-  s.rubygems_version = "2.6.3".freeze
-  s.summary = "Namae (\u{540d}\u{524d}) parses personal names and splits them into their component parts.".freeze
+  s.rubygems_version = "3.2.3".freeze
+  s.summary = "Namae (\u540D\u524D) parses personal names and splits them into their component parts.".freeze
 
   if s.respond_to? :specification_version then
     s.specification_version = 4
+  end
 
-    if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
-      s.add_development_dependency(%q<racc>.freeze, ["= 1.4.9"])
-    else
-      s.add_dependency(%q<racc>.freeze, ["= 1.4.9"])
-    end
+  if s.respond_to? :add_runtime_dependency then
+    s.add_development_dependency(%q<racc>.freeze, ["~> 1.4"])
   else
-    s.add_dependency(%q<racc>.freeze, ["= 1.4.9"])
+    s.add_dependency(%q<racc>.freeze, ["~> 1.4"])
   end
 end
 

data/spec/namae/parser_spec.rb

@@ -1,10 +1,5 @@
 module Namae
   describe 'Parser' do
-
-    it 'does not respond to .new' do
-      expect(Parser).not_to respond_to(:new)
-    end
-
     describe '.instance' do
       let(:parser) { Parser.instance }
 
@@ -120,7 +115,7 @@ module Namae
           end
         end
 
-        %w{Pastor Pr. Reverend Rev. Elder Deacon Deaconess Father Fr. Vicar}.each do |title|
+        %w{Pastor Pr. Reverend Rev. Elder Deacon Deaconess Father Fr. Vicar Rabbi Cantor}.each do |title|
           describe "the next token is #{title.inspect}" do
             before { parser.send(:input).string = title }
             it 'returns a TITLE token' do
@@ -149,10 +144,6 @@ module Namae
             expect(parser.parse!('Ichiro')[0].given).to eq('Ichiro')
           end
 
-          it 'removes numbers' do
-            expect(parser.parse!('Ichiro 20156')[0].given).to eq('Ichiro')
-          end
-
           it 'treats "Lord Byron" as a title and family name' do
             expect(parser.parse!('Lord Byron')[0].values_at(:family, :title)).to eq(['Byron', 'Lord'])
           end
@@ -200,6 +191,66 @@ module Namae
             expect(parser.parse!('Bernado Franecki Ph.D.')[0].values_at(:given, :family, :title)).to eq(['Bernado', 'Franecki', 'Ph.D.'])
             #expect(parser.parse!('Bernado Franecki, Ph.D.')[0].values_at(:given, :family, :title)).to eq(['Bernado', 'Franecki', 'Ph.D.'])
           end
+
+          it 'parses consecutive titles in display order' do
+            expect(parser.parse!('Lt. Col. Bernado Franecki')[0].values_at(:given, :family, :title)).to eq(['Bernado', 'Franecki', 'Lt. Col.'])
+          end
+
+          context 'when include_particle_in_family is false' do
+            let(:parser) { Parser.new(include_particle_in_family: false) }
+
+            it 'parses common capitalized particles as the family name in display order' do
+              expect(parser.parse!('Carlos De Silva')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'De'])
+            end
+
+            it 'parses common capitalized particles with punctuation as the family name in display order' do
+              expect(parser.parse!('Matt St. Hilaire')[0].values_at(:given, :family, :particle)).to eq(['Matt', 'Hilaire', 'St.'])
+            end
+
+            it 'parses common lowercase particles as a particle, not family name in display order' do
+              expect(parser.parse!('Carlos de Silva')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'de'])
+            end
+
+            it 'parses common capitalized particles as the family name in sort order' do
+              expect(parser.parse!('De Silva, Carlos')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'De'])
+            end
+
+            it 'parses common lowercase particles as a particle, not family name in sort order' do
+              expect(parser.parse!('de Silva, Carlos')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'de'])
+            end
+
+            it 'parses common capitalized particles with punctuation as the family name in display order' do
+              expect(parser.parse!('St. Hilaire, Matt')[0].values_at(:given, :family, :particle)).to eq(['Matt', 'Hilaire', 'St.'])
+            end
+          end
+
+          context 'when include_particle_in_family is true' do
+            let(:parser) { Parser.new(include_particle_in_family: true) }
+
+            it 'parses common capitalized particles as the family name in display order' do
+              expect(parser.parse!('Carlos De Silva')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'De Silva', nil])
+            end
+
+            it 'parses common capitalized particles with punctuation as the family name in display order' do
+              expect(parser.parse!('Matt St. Hilaire')[0].values_at(:given, :family, :particle)).to eq(['Matt', 'St. Hilaire', nil])
+            end
+
+            it 'parses common lowercase particles as a particle, not family name in display order' do
+              expect(parser.parse!('Carlos de Silva')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'de'])
+            end
+
+            it 'parses common capitalized particles as the family name in sort order' do
+              expect(parser.parse!('De Silva, Carlos')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'De Silva', nil])
+            end
+
+            it 'parses common lowercase particles as a particle, not family name in sort order' do
+              expect(parser.parse!('de Silva, Carlos')[0].values_at(:given, :family, :particle)).to eq(['Carlos', 'Silva', 'de'])
+            end
+
+            it 'parses common capitalized particles with punctuation as the family name in display order' do
+              expect(parser.parse!('St. Hilaire, Matt')[0].values_at(:given, :family, :particle)).to eq(['Matt', 'St. Hilaire', nil])
+            end
+          end
         end
       end
 

data/spec/thread_safety_spec.rb

@@ -0,0 +1,25 @@
+module Namae
+  describe 'Parser using threads' do
+    let(:name_1_str) { "Foo Bar" }
+    let(:name_2_str) { "Baz" }
+    let(:name_1) { Namae.parse(name_1_str).first }
+    let(:name_2) { Namae.parse(name_2_str).first }
+
+    def compare(string, expectation)
+      name = Namae.parse(string).first
+      given_name_match = expectation.given == name.given
+      family_name_match = expectation.family == name.family
+      raise unless given_name_match && family_name_match
+    end
+
+    it 'has no conflicts' do
+      [[name_1_str, name_1], [name_2_str, name_2]].map do |string, expectation|
+        Thread.new do
+          1000.times do
+            compare(string, expectation)
+          end
+        end
+      end.each(&:join)
+    end
+  end
+end

metadata

@@ -1,30 +1,30 @@
 --- !ruby/object:Gem::Specification
 name: namae
 version: !ruby/object:Gem::Version
-  version: 0.11.2
+  version: 1.1.0
 platform: ruby
 authors:
 - Sylvester Keil
 - Dan Collis-Puro
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2016-11-12 00:00:00.000000000 Z
+date: 2021-03-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: racc
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '='
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.4.9
+        version: '1.4'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '='
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.4.9
+        version: '1.4'
 description: " Namae (名前) is a parser for human names. It recognizes personal names
   of various cultural backgrounds and tries to split them into their component parts
   (e.g., given and family names, honorifics etc.). "
@@ -36,7 +36,6 @@ extensions: []
 extra_rdoc_files:
 - README.md
 files:
-- ".autotest"
 - ".codeclimate.yml"
 - ".coveralls.yml"
 - ".document"
@@ -69,11 +68,12 @@ files:
 - spec/namae/parser_spec.rb
 - spec/namae/utility_spec.rb
 - spec/spec_helper.rb
+- spec/thread_safety_spec.rb
 homepage: https://github.com/berkmancenter/namae
 licenses:
 - AGPL-3.0
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -88,9 +88,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.3
-signing_key: 
+rubygems_version: 3.2.3
+signing_key:
 specification_version: 4
 summary: Namae (名前) parses personal names and splits them into their component parts.
 test_files: []

data/.autotest

@@ -1,21 +0,0 @@
-require 'bundler'
-begin
-  if RUBY_PLATFORM =~ /darwin/
-    Bundler.setup(:default, :development, :debug, :test, :osx)
-    require 'autotest/fsevent'  
-  else
-    Bundler.setup(:default, :development, :debug, :test)
-  end
-rescue Bundler::BundlerError => e
-  $stderr.puts e.message
-  $stderr.puts "Run `bundle install` to install missing gems"
-  exit e.status_code
-end
-
-
-Autotest.add_hook :initialize do |at|
-  at.add_mapping(/.+\.y$/) do |f,_|
-    system 'rake clean racc'
-  end
-end
-