red-datasets 0.1.7 → 0.1.8

data/test/test-house-of-representative.rb

@@ -0,0 +1,54 @@
+class HouseOfRepresentativeTest < Test::Unit::TestCase
+  def setup
+    @dataset = Datasets::HouseOfRepresentative.new
+  end
+
+  def record(*args)
+    Datasets::HouseOfRepresentative::Record.new(*args)
+  end
+
+  test("#each") do
+    assert_equal(record(142,
+                        "衆法の一覧",
+                        nil,
+                        139,
+                        18,
+                        "市民活動促進法案",
+                        "成立",
+                        "経過",
+                        "https://www.shugiin.go.jp/internet/itdb_gian.nsf/html/gian/keika/5516.htm",
+                        nil,
+                        nil,
+                        "衆法",
+                        "熊代　昭彦君外四名",
+                        %w(自由民主党 社会民主党・市民連合 新党さきがけ),
+                        nil,
+                        nil,
+                        nil,
+                        Date.jisx0301("H10.03.04"),
+                        Date.jisx0301("H10.03.11"),
+                        "内閣",
+                        Date.jisx0301("H10.03.17"),
+                        "可決",
+                        Date.jisx0301("H10.03.19"),
+                        "可決",
+                        nil,
+                        nil,
+                        nil,
+                        nil,
+                        nil,
+                        nil,
+                        nil,
+                        Date.jisx0301("H10.01.12"),
+                        "労働・社会政策",
+                        Date.jisx0301("H10.03.03"),
+                        "修正",
+                        Date.jisx0301("H10.03.04"),
+                        "修正",
+                        Date.jisx0301("H10.03.25"),
+                        7,
+                        nil,
+                        nil),
+                 @dataset.each.next)
+  end
+end

data/test/test-nagoya-university-conversation-corpus.rb

@@ -5,116 +5,64 @@ class NagoyaUniversityConversationCorpusTest < Test::Unit::TestCase
 
   sub_test_case("each") do
     test("#sentences") do
-      records = @dataset.each.to_a
-      first_sentences = records[0].sentences
-      last_sentences = records[-1].sentences
+      first_sentences = @dataset.each.next.sentences
       assert_equal([
                      856,
                      {
                        participant_id: 'F107',
-                       content: '＊＊＊の町というのはちいちゃくって、城壁がこう町全体をぐるっと回ってて、それが城壁の上を歩いても１時間ぐらいですよね。'
+                       content: '＊＊＊の町というのはちいちゃくって、城壁がこう町全体をぐるっと回ってて、それが城壁の上を歩いても１時間ぐらいですよね。',
                      },
                      {
                        participant_id: nil,
-                       content: nil
+                       content: nil,
                      },
-                     603,
-                     {
-                       participant_id: 'F007',
-                       content: 'それでは話を始めまーす。'
-                     },
-                     {
-                       participant_id: nil,
-                       content: nil
-                     }
                    ],
                    [
                      first_sentences.size,
                      first_sentences[0].to_h,
                      first_sentences[-1].to_h,
-                     last_sentences.size,
-                     last_sentences[0].to_h,
-                     last_sentences[-1].to_h,
                    ])
     end
 
     test("#participants") do
-      records = @dataset.each.to_a
-      first_participants = records[0].participants
-      last_participants = records[-1].participants
+      first_participants = @dataset.each.next.participants
       assert_equal([
                      4,
                      {
                        id: 'F107',
                        attribute: '女性３０代後半',
                        birthplace: '愛知県幡豆郡出身',
-                       residence: '愛知県幡豆郡在住'
+                       residence: '愛知県幡豆郡在住',
                      },
                      {
                        id: 'F128',
                        attribute: '女性２０代前半',
                        birthplace: '愛知県西尾市出身',
-                       residence: '西尾市在住'
+                       residence: '西尾市在住',
                      },
-                     2,
-                     {
-                       id: 'F007',
-                       attribute: '女性５０代後半',
-                       birthplace: '東京都出身',
-                       residence: '東京都国分寺市在住'
-                     },
-                     {
-                       id: 'F003',
-                       attribute: '女性８０代後半',
-                       birthplace: '栃木県宇都宮市出身',
-                       residence: '国分寺市在住'
-                     }
                    ],
                    [
                      first_participants.size,
                      first_participants[0].to_h,
                      first_participants[-1].to_h,
-                     last_participants.size,
-                     last_participants[0].to_h,
-                     last_participants[-1].to_h
                    ])
     end
 
     test("others") do
-      records = @dataset.each.to_a
+      first_record = @dataset.each.next
       assert_equal([
-                     129,
-                     [
-                       '１（約３５分）',
-                       '２００１年１０月１６日',
-                       'ファミリーレストラン',
-                       '英会話教室の友人',
-                       nil
-                     ],
-                     [
-                       '１２９（３６分）',
-                       '２００３年２月１６日',
-                       '二人の自宅',
-                       '母と娘',
-                       'F007は東京に３８年、F003は東京に６０年居住。'
-                    ]
+                     '１（約３５分）',
+                     '２００１年１０月１６日',
+                     'ファミリーレストラン',
+                     '英会話教室の友人',
+                     nil,
                    ],
                    [
-                     records.size,
-                     [
-                       records[0].name,
-                       records[0].date,
-                       records[0].place,
-                       records[0].relationships,
-                       records[0].note
-                     ],
-                     [
-                       records[-1].name,
-                       records[-1].date,
-                       records[-1].place,
-                       records[-1].relationships,
-                       records[-1].note
-                     ]
+                     first_record.name,
+                     first_record.date,
+                     first_record.place,
+                     first_record.relationships,
+                     first_record.note,
                    ])
     end
   end

data/test/test-postal-code-japan.rb

@@ -1,4 +1,11 @@
 class PostalCodeJapanTest < Test::Unit::TestCase
+  test("invalid") do
+    message = ":reading must be one of [:lowercase, :uppercase, :romaji]: :invalid"
+    assert_raise(ArgumentError.new(message)) do
+      Datasets::PostalCodeJapan.new(reading: :invalid)
+    end
+  end
+
   sub_test_case(":reading") do
     test(":lowercase") do
       dataset = Datasets::PostalCodeJapan.new(reading: :lowercase)

data/test/test-quora-duplicate-question-pair.rb

@@ -8,26 +8,12 @@ class QuoraDuplicateQuestionPairTest < Test::Unit::TestCase
   end
 
   test("#each") do
-    records = @dataset.each.to_a
-    assert_equal([
-                   404290,
-                   record(0,
-                          1,
-                          2,
-                          "What is the step by step guide to invest in share market in india?",
-                          "What is the step by step guide to invest in share market?",
-                          false),
-                   record(404289,
-                          537932,
-                          537933,
-                          "What is like to have sex with cousin?",
-                          "What is it like to have sex with your cousin?",
-                          false),
-                 ],
-                 [
-                   records.size,
-                   records.first,
-                   records.last,
-                 ])
+    assert_equal(record(0,
+                        1,
+                        2,
+                        "What is the step by step guide to invest in share market in india?",
+                        "What is the step by step guide to invest in share market?",
+                        false),
+                 @dataset.each.next)
   end
 end

data/test/test-rdataset.rb

@@ -8,7 +8,7 @@ class RdatasetTest < Test::Unit::TestCase
       test("with package_name") do
         records = @dataset.filter(package: "datasets").to_a
         assert_equal([
-                       102,
+                       104,
                        {
                          package: "datasets",
                          dataset: "ability.cov",
@@ -48,7 +48,7 @@ class RdatasetTest < Test::Unit::TestCase
       test("without package_name") do
         records = @dataset.each.to_a
         assert_equal([
-                       2142,
+                       2337,
                        {
                          package: "AER",
                          dataset: "Affairs",
@@ -110,8 +110,8 @@ class RdatasetTest < Test::Unit::TestCase
           records = @dataset.each.to_a
           assert_equal([
                          144,
-                         { time: 1949,             value: 112 },
-                         { time: 1960.91666666667, value: 432 },
+                         { time: 1949,             value: 112, rownames: 1 },
+                         { time: 1960.91666666667, value: 432, rownames: 144 },
                        ],
                        [
                          records.size,
@@ -141,8 +141,8 @@ class RdatasetTest < Test::Unit::TestCase
           records = @dataset.each.to_a
           assert_equal([
                          153,
-                         { Ozone: nil, "Solar.R": nil, Wind: 14.3, Temp: 56, Month: 5, Day: 5 },
-                         { Ozone: 20, "Solar.R": 223, Wind: 11.5, Temp: 68, Month: 9, Day: 30 },
+                         { Ozone: nil, "Solar.R": nil, Wind: 14.3, Temp: 56, Month: 5, Day: 5, rownames: 5 },
+                         { Ozone: 20, "Solar.R": 223, Wind: 11.5, Temp: 68, Month: 9, Day: 30, rownames: 153 },
                        ],
                        [
                          records.size,
@@ -161,10 +161,10 @@ class RdatasetTest < Test::Unit::TestCase
           records = @dataset.each.to_a
           assert_equal([
                          182,
-                         { event: 1, mag: 7, station: "117", dist: 12, accel: 0.359 },
-                         { event: 16, mag: 5.1, station: nil, dist: 7.6, accel: 0.28 },
-                         { event: 23, mag: 5.3, station: "c168", dist: 25.3, accel: 0.23 },
-                         { event: 23, mag: 5.3, station: "5072", dist: 53.1, accel: 0.022 }
+                         { event: 1, mag: 7, station: 117, dist: 12, accel: 0.359, rownames: 1 },
+                         { event: 16, mag: 5.1, station: nil, dist: 7.6, accel: 0.28, rownames: 79 },
+                         { event: 23, mag: 5.3, station: "c168", dist: 25.3, accel: 0.23, rownames: 170 },
+                         { event: 23, mag: 5.3, station: 5072, dist: 53.1, accel: 0.022, rownames: 182 }
                        ],
                        [
                          records.size,
@@ -187,8 +187,8 @@ class RdatasetTest < Test::Unit::TestCase
           records = @dataset.each.to_a
           assert_equal([
                          192,
-                         { temp: 10, species: 'wheat', start: 0, end: 1.0, germinated: 0 },
-                         { temp: 40, species: 'rice', start: 18, end: Float::INFINITY, germinated: 12 }
+                         { temp: 10, species: 'wheat', start: 0, end: 1, germinated: 0, rownames: 1 },
+                         { temp: 40, species: 'rice', start: 18, end: Float::INFINITY, germinated: 12, rownames: 192 }
                        ],
                        [
                          records.size,
@@ -213,25 +213,27 @@ class RdatasetTest < Test::Unit::TestCase
                           Order: 398_481,
                           Level: 1,
                           Code: 'A',
-                          Parent: '',
+                          Parent: nil,
                           Description: 'AGRICULTURE, FORESTRY AND FISHING',
                           This_item_includes: 'This section includes the exploitation of vegetal and animal natural resources, comprising the activities of growing of crops, raising and breeding of animals, harvesting of timber and other plants, animals or animal products from a farm or their natural habitats.',
-                          This_item_also_includes: '',
-                          Rulings: '',
-                          This_item_excludes: '',
-                          "Reference_to_ISIC_Rev._4": 'A'
+                          This_item_also_includes: nil,
+                          Rulings: nil,
+                          This_item_excludes: nil,
+                          "Reference_to_ISIC_Rev._4": 'A',
+                          rownames: 1
                         },
                         {
                           Order: 399_476,
                           Level: 4,
-                          Code: '99.00',
-                          Parent: '99.0',
+                          Code: 99.0,
+                          Parent: 99.0,
                           Description: 'Activities of extraterritorial organisations and bodies',
                           This_item_includes: "This class includes:\n- activities of international organisations such as the United Nations and the specialised agencies of the United Nations system, regional bodies etc., the International Monetary Fund, the World Bank, the World Customs Organisation, the Organisation for Economic Co-operation and Development, the organisation of Petroleum Exporting Countries, the European Communities, the European Free Trade Association etc.",
                           This_item_also_includes: "This class also includes:\n- activities of diplomatic and consular missions when being determined by the country of their location rather than by the country they represent",
-                          Rulings: '',
-                          This_item_excludes: '',
-                          "Reference_to_ISIC_Rev._4": '9900'
+                          Rulings: nil,
+                          This_item_excludes: nil,
+                          "Reference_to_ISIC_Rev._4": 9900,
+                          rownames: 996
                         }
                       ],
                       [

data/test/test-sudachi-synonym-dictionary.rb

@@ -4,37 +4,18 @@ class SudachiSynonymDictionaryTest < Test::Unit::TestCase
   end
 
   test('#each') do
-    records = @dataset.each.to_a
-    assert_equal([
-                   65206,
-                   {
-                     group_id: "000001",
-                     is_noun: true,
-                     expansion_type: :always,
-                     lexeme_id: 1,
-                     form_type: :typical,
-                     acronym_type: :typical,
-                     variant_type: :typical,
-                     categories: [],
-                     notation: "曖昧",
-                   },
-                   {
-                     group_id: "024916",
-                     is_noun: true,
-                     expansion_type: :expanded,
-                     lexeme_id: 1,
-                     form_type: :typical,
-                     acronym_type: :alphabet,
-                     variant_type: :typical,
-                     categories: ["ビジネス"],
-                     notation: "SCM",
-                   },
-                 ],
-                 [
-                   records.size,
-                   records[0].to_h,
-                   records[-1].to_h,
-                 ])
+    assert_equal({
+                   group_id: "000001",
+                   is_noun: true,
+                   expansion_type: :always,
+                   lexeme_id: 1,
+                   form_type: :typical,
+                   acronym_type: :typical,
+                   variant_type: :typical,
+                   categories: [],
+                   notation: "曖昧",
+                 },
+                 @dataset.each.next.to_h)
   end
 
   sub_test_case('#metadata') do

data/test/test-wikipedia.rb

@@ -7,13 +7,13 @@ class WikipediaTest < Test::Unit::TestCase
       end
 
       test("#each") do
-        contributor = Datasets::Wikipedia::Contributor.new("Elli", 20842734)
+        contributor = Datasets::Wikipedia::Contributor.new("Asparagusus", 43603280)
         revision = Datasets::Wikipedia::Revision.new
-        revision.id = 1002250816
-        revision.parent_id = 854851586
-        revision.timestamp = Time.iso8601("2021-01-23T15:15:01Z")
+        revision.id = 1219062925
+        revision.parent_id = 1219062840
+        revision.timestamp = Time.iso8601("2024-04-15T14:38:04Z")
         revision.contributor = contributor
-        revision.comment = "shel"
+        revision.comment = "Restored revision 1002250816 by [[Special:Contributions/Elli|Elli]] ([[User talk:Elli|talk]]): Unexplained redirect breaking"
         revision.model = "wikitext"
         revision.format = "text/x-wiki"
         revision.text = <<-TEXT.chomp

metadata

@@ -1,15 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: red-datasets
 version: !ruby/object:Gem::Version
-  version: 0.1.7
+  version: 0.1.8
 platform: ruby
 authors:
 - tomisuker
 - Kouhei Sutou
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-05-29 00:00:00.000000000 Z
+date: 2025-02-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: csv
@@ -160,8 +159,11 @@ files:
 - lib/datasets/geolonia.rb
 - lib/datasets/ggplot2-dataset.rb
 - lib/datasets/hepatitis.rb
+- lib/datasets/house-of-councillor.rb
+- lib/datasets/house-of-representative.rb
 - lib/datasets/iris.rb
 - lib/datasets/ita-corpus.rb
+- lib/datasets/japanese-date-parser.rb
 - lib/datasets/kuzushiji-mnist.rb
 - lib/datasets/lazy.rb
 - lib/datasets/libsvm-dataset-list.rb
@@ -189,6 +191,7 @@ files:
 - lib/datasets/zip-extractor.rb
 - red-datasets.gemspec
 - test/helper.rb
+- test/japanese-date-parser-test.rb
 - test/run-test.rb
 - test/test-adult.rb
 - test/test-afinn.rb
@@ -206,6 +209,8 @@ files:
 - test/test-fuel-economy.rb
 - test/test-geolonia.rb
 - test/test-hepatitis.rb
+- test/test-house-of-councillor.rb
+- test/test-house-of-representative.rb
 - test/test-iris.rb
 - test/test-ita-corpus.rb
 - test/test-kuzushiji-mnist.rb
@@ -233,7 +238,6 @@ homepage: https://github.com/red-data-tools/red-datasets
 licenses:
 - MIT
 metadata: {}
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -248,12 +252,12 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.0.dev
-signing_key:
+rubygems_version: 3.6.2
 specification_version: 4
 summary: Red Datasets provides classes that provide common datasets such as iris dataset.
 test_files:
 - test/helper.rb
+- test/japanese-date-parser-test.rb
 - test/run-test.rb
 - test/test-adult.rb
 - test/test-afinn.rb
@@ -271,6 +275,8 @@ test_files:
 - test/test-fuel-economy.rb
 - test/test-geolonia.rb
 - test/test-hepatitis.rb
+- test/test-house-of-councillor.rb
+- test/test-house-of-representative.rb
 - test/test-iris.rb
 - test/test-ita-corpus.rb
 - test/test-kuzushiji-mnist.rb