Dictionary and dictionary reader

.gitignore

@@ -0,0 +1,9 @@
+#Ignore editor backup and temporary files
+*~
+#*
+
+#Ignore org files
+*.org
+
+#Ignore gemset file
+.ruby-gemset

.rubocop.yml

@@ -0,0 +1,13 @@
+require: rubocop-rspec
+
+Metrics/LineLength:
+  Max: 100
+
+Metrics/MethodLength:
+  Max: 20
+
+Metrics/ClassLength:
+  Max: 100
+
+Style/Documentation:
+  Enabled: false

.ruby-version

@@ -0,0 +1 @@
+2.5.1

Gemfile

@@ -0,0 +1,8 @@
+# frozen_string_literal: true
+
+source 'https://rubygems.org'
+
+gem 'pry'
+gem 'rspec'
+gem 'rubocop', '~> 0.56.0', require: false
+gem 'rubocop-rspec'

Gemfile.lock

@@ -0,0 +1,51 @@
+GEM
+  remote: https://rubygems.org/
+  specs:
+    ast (2.4.0)
+    coderay (1.1.2)
+    diff-lcs (1.3)
+    method_source (0.9.0)
+    parallel (1.12.1)
+    parser (2.5.1.0)
+      ast (~> 2.4.0)
+    powerpack (0.1.2)
+    pry (0.11.3)
+      coderay (~> 1.1.0)
+      method_source (~> 0.9.0)
+    rainbow (3.0.0)
+    rspec (3.7.0)
+      rspec-core (~> 3.7.0)
+      rspec-expectations (~> 3.7.0)
+      rspec-mocks (~> 3.7.0)
+    rspec-core (3.7.1)
+      rspec-support (~> 3.7.0)
+    rspec-expectations (3.7.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.7.0)
+    rspec-mocks (3.7.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.7.0)
+    rspec-support (3.7.1)
+    rubocop (0.56.0)
+      parallel (~> 1.10)
+      parser (>= 2.5)
+      powerpack (~> 0.1)
+      rainbow (>= 2.2.2, < 4.0)
+      ruby-progressbar (~> 1.7)
+      unicode-display_width (~> 1.0, >= 1.0.1)
+    rubocop-rspec (1.27.0)
+      rubocop (>= 0.56.0)
+    ruby-progressbar (1.9.0)
+    unicode-display_width (1.4.0)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  pry
+  rspec
+  rubocop (~> 0.56.0)
+  rubocop-rspec
+
+BUNDLED WITH
+   1.16.1

dictionaries/personal.dic

@@ -0,0 +1,3 @@
+foo
+Foo/Simpson
+*bar

dictionaries/test.aff

@@ -0,0 +1,13 @@
+SET UTF-8
+TRY esianrtolcdugmphbyfvkwzESIANRTOLCDUGMPHBYFVKWZ'
+
+REP 2
+REP f ph
+REP ph f
+
+PFX A Y 1
+PFX A 0 re .
+
+SFX B Y 2
+SFX B 0 ed [^y]
+SFX B y ied y

dictionaries/test.dic

@@ -0,0 +1,4 @@
+3
+hello
+try/B
+work/AB

lib/dictionary.rb

@@ -0,0 +1,6 @@
+# frozen_string_literal: true
+
+require 'ostruct'
+
+class Dictionary < OpenStruct
+end

lib/dictionary_reader.rb

@@ -0,0 +1,30 @@
+# frozen_string_literal: true
+
+require_relative 'parsers/dictionary_parser'
+require_relative 'dictionary'
+
+class DictionaryReader
+  class << self
+    def read(*args)
+      dic_files = read_files_from_arguments(args)
+      dic_files_data = get_data_from_dic_files(dic_files)
+      create_dictionary(dic_files_data)
+    end
+
+    private
+
+    def read_files_from_arguments(args)
+      args.map(&File.method(:new))
+    end
+
+    def get_data_from_dic_files(dic_files)
+      dic_files.inject({}) do |acc, file|
+        acc.merge(Parsers::DictionaryParser.parse_file(file))
+      end
+    end
+
+    def create_dictionary(data)
+      Dictionary.new(data)
+    end
+  end
+end

lib/parsers/aff_parser.rb

@@ -0,0 +1,44 @@
+# frozen_string_literal: true
+
+module Parsers
+  class AffParser
+    AFFIX_REGEX = /^(PFX|SFX)(\s|\w|.)*/
+    AFFIX_GROUP_REGEX = /^(PFX|SFX)\s\w/
+
+    def initialize(file)
+      @file = file
+    end
+
+    def parse
+      {
+        affixes: @file.to_a
+                      .map { |ln| ln.tr("\n\t", '') }
+                      .grep(AFFIX_REGEX)
+                      .group_by { |el| el[AFFIX_GROUP_REGEX] }.values
+                      .map(&method(:parse_affix_line))
+      }
+    end
+
+    private
+
+    def parse_affix_line(aff_group)
+      header, *rule_lines = aff_group
+      name, flag, cross_product, line_count = header.split(/[\s*]/)
+      rules = fetch_rules(rule_lines)
+      {
+        name: name,
+        flag: flag,
+        cross_product: cross_product,
+        line_count: line_count,
+        rules: rules
+      }
+    end
+
+    def fetch_rules(lines)
+      lines.map do |el|
+        _, _, stripping_rule, affixes, condition = el.split(/[\s*]/)
+        { stripping_rule: stripping_rule, affixes: affixes, condition: condition }
+      end
+    end
+  end
+end

lib/parsers/dic_parser.rb

@@ -0,0 +1,37 @@
+# frozen_string_literal: true
+
+module Parsers
+  class DicParser
+    WORD_AFF_REGEX = %r{^(\*?)([^\/]+)(?:\/(.+))?$}
+
+    def initialize(file)
+      @file = file
+    end
+
+    def parse
+      @file.to_a
+           .map { |ln| ln.tr("\n\t", '') }
+           .yield_self do |cnt|
+             { approx_word_count: word_count_try(cnt) }.merge(fetch_words(cnt))
+           end
+    end
+
+    private
+
+    def word_count_try(content)
+      Integer(content.first)
+    rescue ArgumentError
+      nil
+    end
+
+    def fetch_words(content)
+      content.yield_self { |cnt| cnt - [word_count_try(cnt).to_s] }
+             .map { |ln| ln.match(WORD_AFF_REGEX).values_at(1, 2, 3) }
+             .map { |star, word, affixes| star == '*' ? word : { word: word, affixes: affixes } }
+             .partition { |word| word.is_a?(String) }
+             .yield_self do |forbidden, regular|
+               { words: regular, forbidden_words: forbidden }
+             end
+    end
+  end
+end

lib/parsers/dictionary_parser.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+require_relative 'dic_parser'
+require_relative 'aff_parser'
+
+module Parsers
+  class DictionaryParser
+    def self.parse_file(file)
+      file_type = File.extname(file).delete('.')
+      class_name = 'Parsers::' + file_type.to_s.capitalize + 'Parser'
+      self.class.const_get(class_name).new(File.new(file)).parse
+    end
+  end
+end

spec/dictionaries/test_affix.aff

@@ -0,0 +1,13 @@
+SET UTF-8
+TRY esianrtolcdugmphbyfvkwzESIANRTOLCDUGMPHBYFVKWZ'
+
+REP 2
+REP f ph
+REP ph f
+
+PFX A Y 1
+PFX A 0 re .
+
+SFX B Y 2
+SFX B 0 ed [^y]
+SFX B y ied y

spec/dictionaries/test_dict.dic

@@ -0,0 +1,4 @@
+3
+hello
+try/B
+work/AB

spec/dictionaries/test_personal.dic

@@ -0,0 +1,3 @@
+foo
+Foo/Simpson
+*bar

spec/dictionaries/unreadable_file

@@ -0,0 +1 @@
+this file will be unreadable

spec/dictionary_reader_spec.rb

@@ -0,0 +1,57 @@
+require_relative '../lib/dictionary_reader'
+require_relative './reference_parsers_results'
+
+RSpec.describe DictionaryReader do
+  include_context 'reference_results'
+
+  subject { described_class.read(*args) }
+
+  context 'unsuccessfull cases:' do
+    context 'file doesn\'t exist' do
+      let(:args) { 'this/file/doesnt/exist' }
+
+      it 'returns no_such_file error' do
+        expect { subject }.to raise_error(Errno::ENOENT, /No such file or directory/)
+      end
+    end
+
+    context 'input wrong type' do
+      let(:args) { :some_symbol }
+
+      it 'returns type error' do
+        expect { subject }.to raise_error(TypeError, /no implicit conversion of/)
+      end
+    end
+
+    context 'file not readable' do
+      let(:args) { './spec/dictionaries/unreadable_file' }
+
+      before do
+        FileUtils.chmod 'u=,go=', args
+      end
+
+      after do
+        FileUtils.chmod 'u=wr,go=rr', args
+      end
+
+      it 'returns permission error' do
+        expect { subject }.to raise_error(Errno::EACCES, /Permission denied/)
+      end
+    end
+  end
+
+  context 'successfull cases' do
+    let(:args) { ['./spec/dictionaries/test_affix.aff', './spec/dictionaries/test_dict.dic'] }
+    let(:result) do
+      Dictionary.new(reference_reader_result)
+    end
+
+    it 'returns Dictionary type' do
+      expect(subject).to be_a(Dictionary)
+    end
+
+    it 'returns reference result' do
+      expect(subject).to eq result
+    end
+  end
+end

spec/dictionary_spec.rb

@@ -0,0 +1,9 @@
+require_relative '../lib/dictionary'
+
+RSpec.describe Dictionary do
+  subject { described_class }
+
+  it 'should be OpenStruct' do
+    expect(subject).to be < OpenStruct
+  end
+end

spec/parsers/aff_parser_spec.rb

@@ -0,0 +1,18 @@
+require_relative '../../lib/parsers/aff_parser'
+require_relative '../reference_parsers_results'
+
+RSpec.describe Parsers::AffParser do
+  include_context 'reference_results'
+
+  let(:file) { File.new('./spec/dictionaries/test_affix.aff') }
+
+  subject { described_class.new(file).parse }
+
+  it 'returns hash' do
+    expect(subject).to be_a_kind_of(Hash)
+  end
+
+  it 'returns reference result' do
+    expect(subject).to eq reference_aff_result
+  end
+end

spec/parsers/dic_parser_spec.rb

@@ -0,0 +1,26 @@
+require_relative '../../lib/parsers/dic_parser'
+require_relative '../reference_parsers_results'
+
+RSpec.describe Parsers::DicParser do
+  include_context 'reference_results'
+
+  let(:file) { File.new('./spec/dictionaries/test_dict.dic') }
+
+  subject { described_class.new(file).parse }
+
+  it 'returns hash' do
+    expect(subject).to be_a_kind_of(Hash)
+  end
+
+  it 'returns reference result' do
+    expect(subject).to eq reference_dic_result
+  end
+
+  context 'personal dictionary' do
+    let(:file) { File.new('./spec/dictionaries/test_personal.dic') }
+
+    it 'returns reference result' do
+      expect(subject).to eq reference_personal_dic_result
+    end
+  end
+end