spicy-proton 1.0.0 → 1.1.0

data/lib/disk-corpus.rb

@@ -0,0 +1,103 @@
+require 'yaml'
+require 'bindata'
+require 'seek'
+
+module Spicy
+end
+
+module Spicy::Disk
+  module Files
+    def self.dir
+      @@dir ||= File.join(File.dirname(__FILE__), 'corpus')
+    end
+
+    def self.adjectives
+      @@adjectives ||= File.join(dir, 'adjectives.bin')
+    end
+
+    def self.nouns
+      @@nouns ||= File.join(dir, 'nouns.bin')
+    end
+
+    def self.colors
+      @@colors ||= File.join(dir, 'colors.bin')
+    end
+  end
+
+  class Header < BinData::Record
+    endian :little
+    uint8 :width
+    uint8 :min_length
+    uint8 :group_count, :value => lambda { cumulative.length }
+    array :cumulative, :type => :uint32, :initial_length => :group_count
+  end
+
+  class WordList
+    include Spicy::Seek
+
+    def initialize(file_name)
+      @file = File.open(file_name, 'r')
+      header = Header.read(@file)
+      @origin = @file.pos
+
+      @width = header.width.to_i
+      @min = header.min_length.to_i
+      @max = @min + header.cumulative.count - 1
+
+      @cumulative = Hash[(@min..@max).zip(header.cumulative.to_a.map(&:to_i))]
+    end
+
+    def close
+      @file.close
+    end
+
+    def word(*args)
+      self.seek(*args) do |index, _|
+        @file.seek(@origin + index * @width, IO::SEEK_SET)
+        @file.read(@width).strip
+      end
+    end
+  end
+  
+  class Corpus
+    private_class_method :new
+
+    def self.use
+      corpus = new
+      begin
+        yield corpus
+      ensure
+        corpus.close
+      end
+    end
+
+    def initialize
+      @lists = {}
+    end
+
+    def close
+      @lists.values.each(&:close)
+    end
+
+    def adjective(*args)
+      generate(:adjectives, *args)
+    end
+
+    def noun(*args)
+      generate(:nouns, *args)
+    end
+
+    def color(*args)
+      generate(:colors, *args)
+    end
+
+    private
+
+    def generate(type, *args)
+      @lists[type] ||= begin
+        WordList.new(Files.send(type))
+      end
+      @lists[type].word(*args)
+    end
+  end
+end

data/lib/memory-corpus.rb

@@ -0,0 +1,71 @@
+require 'yaml'
+require 'seek'
+
+module Spicy
+end
+
+module Spicy::Memory
+  module Files
+    def self.dir
+      @@dir ||= File.join(File.dirname(__FILE__), 'corpus')
+    end
+
+    def self.adjectives
+      @@adjectives ||= File.join(dir, 'adjectives.yaml')
+    end
+
+    def self.nouns
+      @@nouns ||= File.join(dir, 'nouns.yaml')
+    end
+
+    def self.colors
+      @@colors ||= File.join(dir, 'colors.yaml')
+    end
+  end
+
+  class WordList
+    include Spicy::Seek
+
+    def initialize(file_name)
+      dataset = YAML.load_file(file_name)
+
+      @words = dataset['words']
+      @cumulative = dataset['cumulative']
+
+      @min = @cumulative.keys.min
+      @max = @cumulative.keys.max
+    end
+
+    def word(*args)
+      self.seek(*args) do |index, length|
+        index -= (@cumulative[length - 1] || 0)
+        @words[length][index]
+      end
+    end
+  end
+
+  class Corpus
+    def initialize
+      dir = File.dirname(__FILE__)
+      @adjectives = WordList.new(Files.adjectives)
+      @nouns = WordList.new(Files.nouns)
+      @colors = WordList.new(Files.colors)
+    end
+
+    def adjective(*args)
+      @adjectives.word(*args)
+    end
+
+    def noun(*args)
+      @nouns.word(*args)
+    end
+
+    def color(*args)
+      @colors.word(*args)
+    end
+
+    def pair(separator = '-')
+      "#{self.adjective}#{separator}#{self.noun}"
+    end
+  end
+end

data/lib/seek.rb

@@ -0,0 +1,30 @@
+require 'securerandom'
+
+module Spicy
+  module Seek
+    def seek(min: nil, max: nil, length: nil, &found)
+      if !length.nil? && (!min.nil? || !max.nil?)
+        raise ArgumentError.new('length cannot be specified if min or max are specified.')
+      end
+
+      if !min.nil? && !max.nil? && min > max
+        raise ArgumentError.new('min must be no more than max.')
+      end
+
+      min = [min || length || @min, @min].max
+      max = [max || length || @max, @max].min
+
+      rand_min = @cumulative[min - 1] || 0
+      rand_max = @cumulative[max] || @cumulative[@max]
+      index = SecureRandom.random_number(rand_min...rand_max)
+
+      min.upto(max) do |len|
+        if @cumulative[len] > index
+          return found.call(index, len)
+        end
+      end
+
+      nil
+    end
+  end
+end

data/lib/spicy-proton.rb

@@ -1,66 +1,55 @@
-require 'yaml'
-require 'binary'
-require 'files'
+require 'forwardable'
+require 'disk-corpus'
+require 'memory-corpus'
 
 module Spicy
   class Proton
+    extend Forwardable
+
     def initialize
-      dir = File.dirname(__FILE__)
-      @adjectives = YAML.load_file(Files.adjectives)
-      @nouns = YAML.load_file(Files.nouns)
-      @colors = YAML.load_file(Files.colors)
+      @corpus = Memory::Corpus.new
     end
 
-    def self.adjective
-      BinaryCorpus.use(&:adjective)
+    def self.adjective(*args)
+      Disk::Corpus.use do |c|
+        c.adjective(*args)
+      end
     end
 
-    def self.noun
-      BinaryCorpus.use(&:noun)
+    def self.noun(*args)
+      Disk::Corpus.use do |c|
+        c.noun(*args)
+      end
     end
 
-    def self.color
-      BinaryCorpus.use(&:color)
+    def self.color(*args)
+      Disk::Corpus.use do |c|
+        c.color(*args)
+      end
     end
 
     def self.pair(separator = '-')
-      BinaryCorpus.use do |b|
-        "#{b.adjective}#{separator}#{b.noun}"
+      Disk::Corpus.use do |c|
+        "#{c.adjective}#{separator}#{c.noun}"
       end
     end
 
     def self.format(format)
-      BinaryCorpus.use do |b|
-        self.format_with(b, format)
+      Disk::Corpus.use do |c|
+        self.format_with(c, format)
       end
     end
 
-    def adjective
-      random(adjectives)
-    end
-
-    def noun
-      random(nouns)
-    end
-
-    def color
-      random(colors)
-    end
-
-    def pair(separator = '-')
-      "#{self.adjective}#{separator}#{self.noun}"
-    end
-
     def format(format)
       self.class.format_with(self, format)
     end
 
-    attr_accessor :adjectives, :nouns, :colors
+    def_delegators :@corpus, :adjective, :noun, :color, :pair, :adjectives, :nouns, :colors
 
     private
 
     def self.format_with(source, format)
-      format.gsub(/%([anc])/) do
+      format.gsub(/%([anc%])/) do
         case $1
         when 'a'
           source.adjective
@@ -68,13 +57,10 @@ module Spicy
           source.noun
         when 'c'
           source.color
+        when '%'
+          '%'
         end
       end
     end
-
-    def random(set)
-      index = SecureRandom.random_number(set.count)
-      set[index]
-    end
   end
 end

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: spicy-proton
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 1.1.0
 platform: ruby
 authors:
 - Chris Schmich
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-01-29 00:00:00.000000000 Z
+date: 2017-01-30 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: bindata
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.3'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.3'
 - !ruby/object:Gem::Dependency
   name: minitest
   requirement: !ruby/object:Gem::Requirement
@@ -32,12 +46,15 @@ extra_rdoc_files: []
 files:
 - LICENSE
 - README.md
-- lib/binary.rb
+- lib/corpus/adjectives.bin
 - lib/corpus/adjectives.yaml
+- lib/corpus/colors.bin
 - lib/corpus/colors.yaml
-- lib/corpus/combined.bin
+- lib/corpus/nouns.bin
 - lib/corpus/nouns.yaml
-- lib/files.rb
+- lib/disk-corpus.rb
+- lib/memory-corpus.rb
+- lib/seek.rb
 - lib/spicy-proton.rb
 homepage: https://github.com/schmich/spicy-proton
 licenses:
@@ -51,7 +68,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 1.9.3
+      version: 2.0.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="

data/lib/binary.rb

@@ -1,102 +0,0 @@
-require 'json'
-require 'yaml'
-require 'securerandom'
-require 'files'
-
-module Spicy
-  class BinaryCorpus
-    private_class_method :new
-
-    def self.use
-      file = File.open(Files.combined, 'r')
-      begin
-        corpus = new(file)
-        yield corpus
-      ensure
-        file.close
-      end
-    end
-
-    def initialize(file)
-      @file = file
-      @meta = {}
-      self.class.corpus.keys.each do |set|
-        bytes = @file.read(3 * 4)
-        start, count, size = bytes.unpack('L*')
-        @meta[set] = {
-          start: start,
-          count: count,
-          size: size
-        }
-      end
-    end
-
-    def self.generate
-      dir = File.dirname(__FILE__)
-      sources = corpus.values
-
-      sets = sources.map do |source|
-        terms = YAML.load_file(source)
-        {
-          terms: terms,
-          start: 0,
-          count: terms.count,
-          size: terms.map(&:length).max
-        }
-      end
-
-      start = (3 * 4) * sets.length
-      sets.each do |set|
-        set[:start] = start
-        start += set[:count] * set[:size]
-      end
-
-      File.open(Files.combined, 'w+') do |f|
-        meta = []
-        sets.each do |set|
-          meta << set[:start]
-          meta << set[:count]
-          meta << set[:size]
-        end
-
-        f.write(meta.pack('L*'))
-
-        sets.each do |set|
-          set[:terms].each do |term|
-            padding = "\0" * (set[:size] - term.length)
-            f.write(term + padding)
-          end
-        end
-      end
-    end
-
-    def adjective
-      read_random(:adjectives)
-    end
-
-    def noun
-      read_random(:nouns)
-    end
-
-    def color
-      read_random(:colors)
-    end
-
-    private
-
-    def read_random(set)
-      set = @meta[set]
-      index = SecureRandom.random_number(set[:count])
-      @file.seek(set[:start] + index * set[:size], IO::SEEK_SET)
-      @file.read(set[:size]).strip
-    end
-
-    def self.corpus
-      {
-        adjectives: Files.adjectives,
-        nouns: Files.nouns,
-        colors: Files.colors
-      }
-    end
-  end
-end