sapor 0.1b1

data/lib/sapor/regional_data/utopia.rb

@@ -0,0 +1,66 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+module Sapor
+  #
+  # The regional data for Utopia, an area for testing purposes.
+  #
+  class Utopia
+    include Singleton
+
+    LAST_ELECTION_RESULT = { 'Red' => 91_811, 'Green' => 190_934,
+                             'Blue' => 290_647, 'Yellow' => 356_473 }
+
+    LAST_DETAILED_ELECTION_RESULT = { 'North' => { 'Red' => 50, 'Green' => 70
+                                                 },
+                                      'South' => { 'Red' => 70, 'Green' => 50,
+                                                   'Blue' => 100 },
+                                      'East' => { 'Red' => 90, 'Green' => 70,
+                                                  'Blue' => 90 },
+                                      'West' => { 'Red' => 110, 'Green' => 50,
+                                                  'Yellow' => 120 } }
+
+    ELECTORAL_SYSTEM = FirstPastThePost.new(LAST_ELECTION_RESULT,
+                                            LAST_DETAILED_ELECTION_RESULT)
+
+    def area_code
+      'UT'
+    end
+
+    def coalitions
+      [['Red', 'Green'], ['Red', 'Blue']]
+    end
+
+    def population_size
+      1_000_000
+    end
+
+    def threshold
+      nil
+    end
+
+    def no_of_seats
+      LAST_DETAILED_ELECTION_RESULT.size
+    end
+
+    def seats(simulation)
+      ELECTORAL_SYSTEM.project(simulation)
+    end
+  end
+end

data/sapor.gemspec

@@ -0,0 +1,35 @@
+# -*- encoding: utf-8 -*-
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+Gem::Specification.new do |gem|
+  gem.name = 'sapor'
+  gem.version = '0.1b1'
+  gem.authors = ['Filip van Laenen']
+  gem.email = ['f.a.vanlaenen@ieee.org']
+
+  gem.description = 'SAPoR'
+  gem.summary = 'Statistical Analysis of Polling Results'
+  gem.homepage = 'https://github.com/filipvanlaenen/sapor'
+  gem.license = 'GPL'
+
+  gem.require_paths = ['lib']
+  gem.files = `git ls-files`.split("\n")
+  gem.test_files = `git ls-files -- spec`.split("\n")
+  gem.extra_rdoc_files = %w(LICENSE README.md)
+end

data/spec/integration/area_spec.rb

@@ -0,0 +1,28 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'spec_helper'
+
+describe Sapor::Area, '#load_election_results' do
+  it 'loads four constituencies for the 2012 election results for Catalonia' do
+    area = Sapor::Area.instance
+    catalonia_2012 = area.load_election_results('catalonia-2012.psv')
+    expect(catalonia_2012.size).to eq(4)
+  end
+end

data/spec/integration/poll_spec.rb

@@ -0,0 +1,107 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'spec_helper'
+
+SAMPLE_FILE = File.join('spec', 'integration', 'sample.poll')
+MAX_ERROR = 0.05
+
+#
+# Logger logging messages to an array.
+#
+class ArrayLogger
+  attr_reader :messages
+
+  def initialize
+    @messages = []
+  end
+
+  def info(message)
+    @messages << message
+  end
+end
+
+#
+# LoggerBuilder that creates ArrayLogger instances.
+#
+class ArrayLoggerBuilder
+  def create_logger
+    ArrayLogger.new
+  end
+end
+
+Sapor::LogFacade.logger_builder = ArrayLoggerBuilder.new
+
+describe Sapor::Poll, '#from_file' do
+  it 'reads a poll from a file and extracts area' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    expect(poll.area.area_code).to eq('UT')
+  end
+
+  it 'reads a poll from a file and extracts first result' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    expect(poll.result('Red')).to eq(1)
+  end
+
+  it 'reads a poll from a file and extracts last result' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    expect(poll.result('Blue')).to eq(3)
+  end
+end
+
+describe Sapor::Poll, '#analyze' do
+  it 'logs “Analyzing as a set of dichotomies...”' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    poll.analyze(MAX_ERROR)
+    expect(poll.logger.messages).to include('Analyzing as a set of' \
+                                            ' dichotomies...')
+  end
+
+  it 'logs “Analyzing as a polychotomy...”' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    poll.analyze(MAX_ERROR)
+    expect(poll.logger.messages).to include('Analyzing as a polychotomy...')
+  end
+
+  it 'logs “Done.” at the end' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    poll.analyze(MAX_ERROR)
+    expect(poll.logger.messages.last).to eq('Done.')
+  end
+
+  it 'logs error estimates' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    poll.analyze(MAX_ERROR)
+    expect(poll.logger.messages).to include('Error estimate: ε ≤ 33.3%.')
+  end
+
+  it 'logs MPV and CI(95%) reports from Dichotomies' do
+    poll = Sapor::Poll.from_file(SAMPLE_FILE)
+    poll.analyze(MAX_ERROR)
+    expected_report = 'Most probable fractions and 95% confidence' \
+                      " intervals:\n" \
+                      "Choice    MPF      CI(95%)\n" \
+                      "Yellow   46.3%   22.2%– 74.1%\n" \
+                      "Blue     24.1%    7.4%– 51.9%\n" \
+                      "Green    16.7%    3.7%– 44.4%\n" \
+                      "Red       9.3%    0.0%– 37.0%\n" \
+                      'Other     9.3%    0.0%– 37.0%'
+    expect(poll.logger.messages).to include(expected_report)
+  end
+end

data/spec/integration/sample.poll

@@ -0,0 +1,7 @@
+Area=UT
+==
+Red=1
+Green=2
+Blue=3
+Yellow=6
+Other=1

data/spec/spec_helper.rb

@@ -0,0 +1,31 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+$LOAD_PATH << 'lib'
+require 'sapor'
+
+require 'timeout'
+
+RSpec.configure do |config|
+  config.around do |example|
+    Timeout.timeout(1) do
+      example.run
+    end
+  end
+end

data/spec/unit/area_spec.rb

@@ -0,0 +1,115 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'spec_helper'
+
+NORTH = 'North'
+SOUTH = 'South'
+
+RED = 'Red'
+BLUE = 'Blue'
+
+NORTH_RED_LINE = "#{NORTH} | #{RED} | 5"
+NORTH_BLUE_LINE = "#{NORTH} | #{BLUE} | 6"
+SOUTH_RED_LINE = "#{SOUTH} | #{RED} | 7"
+SOUTH_BLUE_LINE_WITH_THOUSANDS_SEPARATOR = "#{SOUTH} | #{BLUE} | 8,888,888"
+EMPTY_LINE = ' '
+COMMENT = '# A comment'
+BROKEN_LINE = 'Broken line'
+
+describe Sapor::Area, '#lines_to_election_results' do
+  it 'converts a single line to 1 constituency' do
+    area = Sapor::Area.instance
+    election_results = area.lines_to_election_results(NORTH_RED_LINE)
+    expect(election_results.size).to eq(1)
+  end
+
+  it 'converts two lines for the same constituency to 1 constituency' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, NORTH_BLUE_LINE].join("\n")
+    election_results = area.lines_to_election_results(lines)
+    expect(election_results.size).to eq(1)
+  end
+
+  it 'converts two lines for two constituencies to 2 constituencies' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, SOUTH_RED_LINE].join("\n")
+    election_results = area.lines_to_election_results(lines)
+    expect(election_results.size).to eq(2)
+  end
+
+  it 'ignores empty lines' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, EMPTY_LINE].join("\n")
+    election_results = area.lines_to_election_results(lines)
+    expect(election_results.size).to eq(1)
+  end
+
+  it 'ignores comments' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, COMMENT].join("\n")
+    election_results = area.lines_to_election_results(lines)
+    expect(election_results.size).to eq(1)
+  end
+
+  it 'sets the number of votes for a party in a constituency correctly' do
+    area = Sapor::Area.instance
+    election_results = area.lines_to_election_results(NORTH_RED_LINE)
+    expect(election_results[NORTH][RED]).to eq(5)
+  end
+
+  it 'ignores thousands separators' do
+    area = Sapor::Area.instance
+    election_results = area.lines_to_election_results(\
+      SOUTH_BLUE_LINE_WITH_THOUSANDS_SEPARATOR)
+    expect(election_results[SOUTH][BLUE]).to eq(8_888_888)
+  end
+
+  it 'raises an ArgumentError when a line is broken' do
+    area = Sapor::Area.instance
+    expect { area.lines_to_election_results(BROKEN_LINE) }.to \
+      raise_error(ArgumentError, "Broken line: #{BROKEN_LINE}.")
+  end
+
+  it 'raises an ArgumentError if the same choice is added twice to the same ' \
+     'constituency' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, NORTH_RED_LINE].join("\n")
+    expect { area.lines_to_election_results(lines) }.to \
+      raise_error(ArgumentError, "Choice #{RED} appeared twice in " \
+                                 "constituency #{NORTH}.")
+  end
+end
+
+describe Sapor::Area, '#summarize_election_results' do
+  it 'summarizes one result to one result' do
+    area = Sapor::Area.instance
+    election_results = area.lines_to_election_results(NORTH_RED_LINE)
+    summary = area.summarize_election_results(election_results)
+    expect(summary[RED]).to eq(5)
+  end
+
+  it 'summarizes results of same choice in two constituencies' do
+    area = Sapor::Area.instance
+    lines = [NORTH_RED_LINE, SOUTH_RED_LINE].join("\n")
+    election_results = area.lines_to_election_results(lines)
+    summary = area.summarize_election_results(election_results)
+    expect(summary[RED]).to eq(12)
+  end
+end

data/spec/unit/binomials_cache_spec.rb

@@ -0,0 +1,34 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'spec_helper'
+
+describe Sapor::BinomialsCache, '#binomial' do
+  it 'returns the correct binomial' do
+    actual = Sapor::BinomialsCache.binomial(10, 20)
+    expected = 10.large_float_binomial_by_product_of_divisions(20)
+    expect(actual).to eq(expected)
+  end
+
+  it 'returns the same binomial when asked a second time' do
+    first = Sapor::BinomialsCache.binomial(10, 21)
+    second = Sapor::BinomialsCache.binomial(10, 21)
+    expect(second).to eq(first)
+  end
+end

data/spec/unit/catalonia_spec.rb

@@ -0,0 +1,82 @@
+# encoding: utf-8
+#
+# Statistical Analysis of Polling Results (SAPoR)
+# Copyright (C) 2014 Filip van Laenen <f.a.vanlaenen@ieee.org>
+#
+# This file is part of SAPoR.
+#
+# SAPoR is free software: you can redistribute it and/or modify it under the
+# terms of the GNU General Public License as published by the Free Software
+# Foundation, either version 3 of the License, or (at your option) any later
+# version.
+#
+# SAPoR is distributed in the hope that it will be useful, but WITHOUT ANY
+# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
+# A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+#
+# You can find a copy of the GNU General Public License in /doc/gpl.txt
+#
+
+require 'spec_helper'
+
+describe Sapor::Catalonia, '#area_code' do
+  it 'returns ES-CAT as the area code' do
+    expect(Sapor::Catalonia.instance.area_code).to eq('ES-CT')
+  end
+end
+
+describe Sapor::Catalonia, '#coalitions' do
+  it 'returns JS and JS+CUP as the coalitions to track' do
+    expect(Sapor::Catalonia.instance.coalitions).to include(['Junts pel Sí'])
+    expect(Sapor::Catalonia.instance.coalitions).to \
+      include(['Junts pel Sí',
+               "Candidatura d'Unitat Popular – Alternativa d'Esquerres"])
+  end
+end
+
+describe Sapor::Catalonia, '#no_of_seats' do
+  it 'returns 135 as the number of seats' do
+    expect(Sapor::Catalonia.instance.no_of_seats).to eq(135)
+  end
+end
+
+describe Sapor::Catalonia, '#population_size' do
+  it 'returns a population size of 3,668,310' do
+    expect(Sapor::Catalonia.instance.population_size).to eq(3_668_310)
+  end
+end
+
+describe Sapor::Catalonia, '#seats' do
+  it 'calculates the number of seats for the election of 2012 (adapted to' \
+     ' 2015) correctly' do
+    catalonia = Sapor::Catalonia.instance
+    results = catalonia.overall_election_results_of_2012_adapted_to_2015
+    seats = catalonia.seats(results)
+    expect(seats['Junts pel Sí']).to eq(64)
+    expect(seats['Partit dels Socialistes de Catalunya']).to eq(20)
+    expect(seats['Partit Popular de Catalunya']).to eq(19)
+    expect(seats['Catalunya Sí que es Pot']).to eq(13)
+    expect(seats['Ciutadans-Partido de la Ciudadanía']).to eq(9)
+    expect(seats['Unió Democràtica de Catalunya']).to eq(7)
+    expect(seats["Candidatura d'Unitat Popular – Alternativa d'Esquerres"]).to \
+      eq(3)
+    expect(seats['Plataforma per Catalunya']).to eq(0)
+    expect(seats['Solidaritat Catalana per la Independència']).to eq(0)
+    expect(seats['Escons en Blanc']).to eq(0)
+    expect(seats['Partido Animalista Contra el Maltrato Animal']).to eq(0)
+    expect(seats['Pirates de Catalunya']).to eq(0)
+    expect(seats['Unió, Progrés i Democràcia']).to eq(0)
+    expect(seats['Farts.cat']).to eq(0)
+    expect(seats['Via Democràtica']).to eq(0)
+    expect(seats["Unificació Comunista d'Espanya"]).to eq(0)
+    expect(seats["Partit Republicà d'Esquerra – Izquierda Republicana"]).to \
+      eq(0)
+    expect(seats['Socialistes i Republicans']).to eq(0)
+  end
+end
+
+describe Sapor::Catalonia, '#threshold' do
+  it 'returns a threshold of 3%' do
+    expect(Sapor::Catalonia.instance.threshold).to eq(0.03)
+  end
+end