normalize_country 0.3.1 → 0.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6124424b5e181869babd13d860f376b0f50f556897a1d9eb2d81137d12a85971
-  data.tar.gz: d0459074064cdcf9e8c9aa6a20f47a15b750be93b11406faccf281990dc2572a
+  metadata.gz: 793c9aaa61fc028c685a5a27681a670e12675e3f2d5db2bb4343167da1d24d94
+  data.tar.gz: 3a8c27532a071ec3baf5114a0ff1c35c643b85944fc45f0cb69586261eb1616c
 SHA512:
-  metadata.gz: eb739c3d7b976793f8b479d4e271a704733d85183c3c55ea3b9e768b03c8992a8c5ed5fdc9d08a03e14757ae4ae5df821f0d5a671bda1a04d7acdf2f419583d9
-  data.tar.gz: bfdedc36669f192bd2c2efc85fd51e7b99340b14f614e2f6ffe8f726a0db29989dcc699c0fd4d41bb0f288487b4325e1474e41212604d6268824597754620743
+  metadata.gz: 800a8cd65c7ed894a1af4950e03e795a86a2debbec7641aba2d676d6ebd99061a28e0672590d6ce087078813ac365ef1eff3f4a19adbb925cc8d7b8022a7fd98
+  data.tar.gz: ec34c172d11d352a1107254f9e9382a8a6e50538cf6cdf4b9a354cb751d6f8cce1b422900182dc530ff051fd6f9e08b2093ba5f5571c43a8c0b1b765222840f6

data/README.rdoc

@@ -2,8 +2,7 @@
 
 Convert country names and codes to a standard.
 
-{<img src="https://travis-ci.org/sshaw/normalize_country.svg?branch=master" alt="Build Status" />}[https://travis-ci.org/sshaw/normalize_country]
-{<img src="https://codeclimate.com/github/sshaw/normalize_country.svg" />}[https://codeclimate.com/github/sshaw/normalize_country]
+{<img src="https://github.com/sshaw/normalize_country/actions/workflows/ci.yml/badge.svg" alt="Build Status" />}[https://github.com/sshaw/normalize_country]
 
 === Overview
 
@@ -59,6 +58,11 @@ will convert to/from the following:
 
 A list of valid formats can be obtained by calling +NormalizeCountry.formats+.
 
+=== Custom Database
+
+Set the +NORMALIZE_COUNTRY_DB+ environment variable.
+See <code>lib/normalize_country/countries/en.yml</code> for the expected format.
+
 === Obtaining an Array or Hash
 
  NormalizeCountry.to_a                              # Defaults to NormalizeCountry.to

data/lib/normalize_country/countries/en.yml

@@ -172,6 +172,8 @@ AW:
   emoji: "\U0001F1E6\U0001F1FC"
   shortcode: ":flag-aw:"
 AX:
+  aliases:
+  - Aland Islands
   alpha2: AX
   alpha3: ALA
   fifa:
@@ -196,6 +198,7 @@ AZ:
 BA:
   aliases:
   - Bosnia-Herzegovina
+  - Bosnia Herzegovina
   alpha2: BA
   alpha3: BIH
   fifa: BIH
@@ -299,6 +302,9 @@ BL:
   - Saint Barts
   - St. Barts
   - St. Barths
+  - Saint Barthelemy
+  - Saint Barthélemy
+  - St. Barthélemy
   alpha2: BL
   alpha3: BLM
   fifa:
@@ -451,10 +457,15 @@ CC:
   shortcode: ":flag-cc:"
 CD:
   aliases:
+  - Congo Kinshasa
   - Congo-Kinshasa
-  - DRC
-  - DR Congo
   - Congo, The Democratic Republic Of The
+  - Congo, Democratic Republic of
+  - Democratic Republic of Congo
+  - DR Congo
+  - RD Congo
+  - DROC
+  - DRC
   alpha2: CD
   alpha3: COD
   fifa: COD
@@ -479,6 +490,7 @@ CF:
 CG:
   aliases:
   - Congo-Brazzaville
+  - Congo Brazzaville
   - Congo Republic
   - Congo, Republic of the
   alpha2: CG
@@ -503,6 +515,11 @@ CH:
   emoji: "\U0001F1E8\U0001F1ED"
   shortcode: ":flag-ch:"
 CI:
+  aliases:
+  - Côte D'Ivoire
+  - Côte D’Ivoire
+  - Cote D'Ivoire
+  - Cote D’Ivoire
   alpha2: CI
   alpha3: CIV
   fifa: CIV
@@ -536,6 +553,9 @@ CL:
   emoji: "\U0001F1E8\U0001F1F1"
   shortcode: ":flag-cl:"
 CM:
+  aliases:
+  - Cameroun
+  - Cameroon, Republic of
   alpha2: CM
   alpha3: CMR
   fifa: CMR
@@ -547,6 +567,8 @@ CM:
   emoji: "\U0001F1E8\U0001F1F2"
   shortcode: ":flag-cm:"
 CN:
+  aliases:
+  - PRC
   alpha2: CN
   alpha3: CHN
   fifa: CHN
@@ -837,6 +859,9 @@ FM:
   emoji: "\U0001F1EB\U0001F1F2"
   shortcode: ":flag-fm:"
 FO:
+  aliases:
+  - Faroes
+  - Faeroes
   alpha2: FO
   alpha3: FRO
   fifa: FRO
@@ -1050,6 +1075,9 @@ GU:
   emoji: "\U0001F1EC\U0001F1FA"
   shortcode: ":flag-gu:"
 GW:
+  aliases:
+  - Guinea Bissau
+  - Republic of Guinea Bissau
   alpha2: GW
   alpha3: GNB
   fifa: GNB
@@ -2147,6 +2175,8 @@ QA:
   emoji: "\U0001F1F6\U0001F1E6"
   shortcode: ":flag-qa:"
 RE:
+  aliases:
+  - Reunion
   alpha2: RE
   alpha3: REU
   fifa:
@@ -2235,6 +2265,12 @@ SC:
   emoji: "\U0001F1F8\U0001F1E8"
   shortcode: ":flag-sc:"
 SD:
+  aliases:
+  - The Sudan
+  - The Sudan, Republic of
+  - Republic of Sudan
+  - Sudan, Republic of
+  - Sudan, Republic of the
   alpha2: SD
   alpha3: SDN
   fifa: SDN
@@ -2372,7 +2408,10 @@ SR:
   shortcode: ":flag-sr:"
 SS:
   aliases:
+  - South Sudan, Republic of
   - S. Sudan
+  - Republic of S. Sudan
+  - S. Sudan, Republic of
   alpha2: SS
   alpha3: SSD
   fifa:
@@ -2387,6 +2426,11 @@ ST:
   aliases:
   - Democratic Republic of Sao Tome and Principe
   - São Tomé and Príncipe
+  - Sao Tomé and Príncipe
+  - Sao Tome and Príncipe
+  - São Tomé and Principe
+  - São Tome and Principe
+  - São Tome and Príncipe
   alpha2: ST
   alpha3: STP
   fifa: STP
@@ -2456,7 +2500,7 @@ TC:
   iso_name: Turks and Caicos Islands
   numeric: "796"
   official: Turks and Caicos Islands
-  short: Turks and Caicos Islands
+  short: Turks and Caicos
   emoji: "\U0001F1F9\U0001F1E8"
   shortcode: ":flag-tc:"
 TD:

data/lib/normalize_country.rb

@@ -3,7 +3,7 @@
 require "yaml"
 
 module NormalizeCountry
-  VERSION = "0.3.1"
+  VERSION = "0.3.3"
   Countries = {}
 
   class << self
@@ -46,7 +46,7 @@ module NormalizeCountry
     end
   end
 
-  class Country
+  class Country   # :nodoc: all
     def initialize(config)
       raise ArgumentError, "country config must be a hash" unless Hash === config
 
@@ -78,7 +78,7 @@ module NormalizeCountry
     end
   end
 
-  path = File.join(File.dirname(__FILE__), "normalize_country", "countries", "en.yml")
+  path = ENV["NORMALIZE_COUNTRY_DB"] || File.join(File.dirname(__FILE__), "normalize_country", "countries", "en.yml")
   data = YAML.load_file(path)
   data.values.each do |mapping|
     country = Country.new(mapping)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: normalize_country
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.3
 platform: ruby
 authors:
 - Skye Shaw
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-01-01 00:00:00.000000000 Z
+date: 2025-02-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -54,13 +54,12 @@ files:
 - bin/normalize_country
 - lib/normalize_country.rb
 - lib/normalize_country/countries/en.yml
-- lib/normalize_country/scanner.rb
 - spec/normalize_country_spec.rb
 homepage: http://github.com/sshaw/normalize_country
 licenses:
 - MIT
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -75,9 +74,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.6
-signing_key: 
+rubygems_version: 3.5.16
+signing_key:
 specification_version: 4
 summary: Convert country names and codes to a standard
 test_files:

data/lib/normalize_country/scanner.rb

@@ -1,121 +0,0 @@
-module NormalizeCountry
-  class Tokenizer
-    # Need numbers to match ISO codes
-    # Need "-" to match emoji names
-    # Need regex to match emojis
-    PATTERN = /[[:word:]]+/
-
-    def initialize(s)
-      @scanner = StringScanner.new(s)
-    end
-
-    def scan
-      @scanner.scan_until(PATTERN)
-      @scanner.matched
-    end
-
-    def peek
-      match = scan
-      @scanner.unscan if match
-      match
-    end
-
-    def end?
-      peek.nil?
-    end
-  end
-
-  class Scanner
-    def initialize(options = nil)
-      options ||= {}
-
-      @to = options[:to] || NormalizeCountry.to
-      @table = lookup_table(options[:from] || NormalizeCountry.formats)   # need aliases!!!
-    end
-
-    def convert(text)
-      s = Tokenizer.new(text)
-
-      matches = []
-      stack = []
-      match_position = @table
-
-      while !s.end?
-        word = s.peek.downcase
-        if !match_position[word]
-          s.scan
-          next
-        end
-
-        stack << s.scan
-        alternatives = match_position[stack[-1].downcase]
-
-        peek = s.peek
-        if alternatives && peek && alternatives[peek.downcase]
-          match_position = alternatives
-          next
-        end
-
-        if match_position[stack[-1].downcase][:match]
-          text = stack.join(" ")
-          matches << { :matched => text, :converted => NormalizeCountry(text, :to => @to) }
-
-        end
-
-        stack.clear
-        match_position = @table
-      end
-
-      matches
-    end
-
-    private
-
-    def lookup_table(from_formats)
-      table = {}
-
-      NormalizeCountry::Countries.values.uniq.each_with_object(table) do |country, o|
-        # country.name # eq all
-        # no way to get aliases
-        from_formats.each do |format|
-          name = country[format]
-          next unless name
-
-          head = o
-          parts = name.split(/[[:space:]]+/)
-          parts.each_with_index do |word, i|
-            # options[:case_sensitive_codes] = [x,y] # or true
-            # options[:case_sensitive_formats] = [x,y]
-            # options[:case_sensitive_formats] = true   # alpha2, alpha3, fifa, ioc
-            word.downcase! #unless parts.size == 1 && ABBRV_FORMATS.include?(format)
-
-            # if head[word]
-            #   if i == parts.size - 1
-            #     head[word][:match] = true
-            #   else
-            #     head = head[word]
-            #   end
-            # else
-            #   head[word] = {}
-            #   if i == parts.size - 1
-            #     head[word][:match] = true
-            #   else
-            #     head = head[word]
-            #   end
-            # end
-
-            head[word] = {} unless head[word]
-
-            if i == parts.size - 1
-              head[word][:match] = true
-            else
-              head = head[word]
-            end
-          end
-        end
-
-        o
-      end
-    end
-  end
-end