rambling-trie 0.8.1 → 0.9.0

data/lib/rambling/trie/plain_text_reader.rb

@@ -6,7 +6,7 @@ module Rambling
       # @param [String] filepath the full path of the file to load the words from.
       # @yield [String] Each line read from the file.
       def each_word filepath
-        each_line(filepath) { |line| yield line.chomp }
+        each_line(filepath) { |line| yield line.chomp! }
       end
 
       private

data/lib/rambling/trie/raw_node.rb

@@ -0,0 +1,90 @@
+module Rambling
+  module Trie
+    # A representation of a node in an uncompressed Trie data structure.
+    class RawNode < Rambling::Trie::Node
+      # Adds a branch to the current trie node based on the word
+      # @param [String] word the word to add the branch from.
+      # @return [Rambling::Trie::RawNode] the just added branch's root node.
+      # @note This method clears the contents of the word variable.
+      def add word
+        if word.empty?
+          terminal!
+        else
+          add_to_children_tree word
+        end
+      end
+
+      # Checks if a path for set of characters exists in the trie.
+      # @param [Array] chars the characters to look for in the trie.
+      # @return [Boolean] `true` if the characters are found, `false` otherwise.
+      def partial_word? chars = []
+        if chars.empty?
+          true
+        else
+          letter = chars.slice!(0).to_sym
+          child = children_tree[letter]
+          !!child && child.partial_word?(chars)
+        end
+      end
+
+      # Checks if a path for set of characters represents a word in the trie.
+      # @param [Array] chars the characters to look for in the trie.
+      # @return [Boolean] `true` if the characters are found and form a word,
+      # `false` otherwise.
+      def word? chars = []
+        if chars.empty?
+          terminal?
+        else
+          letter = chars.slice!(0).to_sym
+          child = children_tree[letter]
+          !!child && child.word?(chars)
+        end
+      end
+
+      # Returns all words that start with the specified characters.
+      # @param [Array] chars the characters to look for in the trie.
+      # @return [Array] all the words contained in the trie that start with the specified characters.
+      def scan chars
+        closest_node chars
+      end
+
+      # Always return `false` for a raw (uncompressed) node.
+      # @return [Boolean] always false for a raw (uncompressed) node.
+      def compressed?
+        false
+      end
+
+      protected
+
+      def closest_node chars
+        if chars.empty?
+          self
+        else
+          letter = chars.slice!(0).to_sym
+          child = children_tree[letter]
+
+          if child
+            child.closest_node chars
+          else
+            Rambling::Trie::MissingNode.new
+          end
+        end
+      end
+
+      private
+
+      def add_to_children_tree word
+        letter = word.slice!(0).to_sym
+        child = children_tree[letter] || new_node(letter)
+        child.add word
+        child
+      end
+
+      def new_node letter
+        node = Rambling::Trie::RawNode.new self
+        node.letter = letter
+        children_tree[letter] = node
+      end
+    end
+  end
+end

data/lib/rambling/trie/tasks/helpers/path.rb

@@ -0,0 +1,13 @@
+module Helpers
+  module Path
+    def path *filename
+      Pathname.new(full_path *filename).cleanpath
+    end
+
+    private
+
+    def full_path *filename
+      full_path = File.join File.dirname(__FILE__), '..', '..', '..', '..', '..', *filename
+    end
+  end
+end

data/lib/rambling/trie/tasks/helpers/time.rb

@@ -0,0 +1,7 @@
+module Helpers
+  module Time
+    def time
+      @time ||= ::Time.now.strftime '%Y-%m-%d %H.%M.%S.%L'
+    end
+  end
+end

data/lib/rambling/trie/tasks/performance.rb

@@ -1,92 +1,11 @@
+require 'fileutils'
 require 'benchmark'
-
-namespace :performance do
-  def report name, trie, output
-    words = %w(hi help beautiful impressionism anthropological)
-    methods = [:word?, :partial_word?]
-
-    output.puts "==> #{name}"
-    methods.each do |method|
-      output.puts "`#{method}`:"
-      words.each do |word|
-        output.print "#{word} - #{trie.send method, word}".ljust 30
-        output.puts Benchmark.measure { 200_000.times { trie.send method, word }}
-      end
-    end
-  end
-
-  def generate_report filename = nil
-    output = filename.nil? ? $stdout : File.open(filename, 'a+')
-
-    output.puts "\nReport for rambling-trie version #{Rambling::Trie::VERSION}"
-
-    trie = nil
-    measure = Benchmark.measure { trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt') }
-
-    if ENV['profile_creation']
-      output.puts '==> Creation'
-      output.print 'Rambling::Trie.create'.ljust 30
-      output.puts measure
-    end
-
-    report 'Uncompressed', trie, output
-
-    return unless trie.respond_to? :compress!
-
-    trie.compress!
-    report 'Compressed', trie, output
-
-    output.close
-  end
-
-  def path *filename
-    File.join File.dirname(__FILE__), '..', '..', '..', '..', *filename
-  end
-
-  desc 'Generate performance report'
-  task :report do
-    puts 'Generating performance report...'
-    generate_report
-  end
-
-  namespace :report do
-    desc 'Generate performance report and append result to reports/performance'
-    task :save do
-      puts 'Generating performance report...'
-      generate_report path('reports', 'performance')
-      puts 'Report has been saved to reports/performance'
-    end
-  end
-
-  desc 'Generate application profiling reports'
-  task :profile do
-    require 'ruby-prof'
-
-    puts 'Generating profiling reports...'
-
-    rambling_trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
-    words = %w(hi help beautiful impressionism anthropological)
-    methods = [:word?, :partial_word?]
-    tries = [lambda {rambling_trie.clone}, lambda {rambling_trie.clone.compress!}]
-
-    methods.each do |method|
-      tries.each do |trie_generator|
-        trie = trie_generator.call
-        result = RubyProf.profile do
-          words.each do |word|
-            200_000.times { trie.send method, word }
-          end
-        end
-
-        File.open path('reports', "profile-#{trie.compressed? ? 'compressed' : 'uncompressed'}-#{method.to_s.sub(/\?/, '')}-#{Time.now.to_i}"), 'w' do |file|
-          RubyProf::CallTreePrinter.new(result).print file
-        end
-      end
-    end
-
-    puts 'Done'
-  end
-
-  desc 'Generate profiling and performance reports'
-  task all: [:profile, :report]
-end
+require 'ruby-prof'
+require 'memory_profiler'
+require 'benchmark/ips'
+require_relative 'helpers/path'
+require_relative 'performance/directory'
+require_relative 'performance/benchmark'
+require_relative 'performance/profile/call_tree'
+require_relative 'performance/profile/memory'
+require_relative 'performance/all'

data/lib/rambling/trie/tasks/performance/all.rb

@@ -0,0 +1,4 @@
+namespace :performance do
+  desc 'Generate profiling and performance reports'
+  task all: ['profile:call_tree', :report]
+end

data/lib/rambling/trie/tasks/performance/benchmark.rb

@@ -0,0 +1,172 @@
+require_relative '../helpers/path'
+
+namespace :performance do
+  include Helpers::Path
+
+  class BenchmarkMeasurement
+    def initialize output
+      @output = output
+    end
+
+    def perform times, params = nil
+      params = Array params
+      params << nil unless params.any?
+
+      params.each do |param|
+        output.print "#{param}".ljust 20
+
+        measure times, param do |param|
+          yield param
+        end
+      end
+    end
+
+    def banner
+      output.puts "\nBenchmark for rambling-trie version #{Rambling::Trie::VERSION}"
+    end
+
+    private
+
+    attr_reader :output
+
+    def measure times, param = nil
+      result = nil
+
+      measure = Benchmark.measure do
+        times.times do
+          result = yield param
+        end
+      end
+
+      output.print "#{result}".ljust 10
+      output.puts measure
+    end
+  end
+
+  def with_file filename = nil
+    output = filename.nil? ? IO.new(1) : File.open(filename, 'a+')
+
+    yield output
+
+    output.close
+  end
+
+  def generate_lookups_benchmark filename = nil
+    with_file filename do |output|
+      measure = BenchmarkMeasurement.new output
+      measure.banner
+
+      trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+      [ trie, trie.clone.compress! ].each do |trie|
+        output.puts "==> #{trie.compressed? ? 'Compressed' : 'Uncompressed'}"
+        words = %w(hi help beautiful impressionism anthropological)
+
+        output.puts '`word?`'
+        measure.perform 200_000, words do |word|
+          trie.word? word
+        end
+
+        output.puts '`partial_word?`'
+        measure.perform 200_000, words do |word|
+          trie.partial_word? word
+        end
+      end
+    end
+  end
+
+  def generate_scans_benchmark filename = nil
+    with_file filename do |output|
+      measure = BenchmarkMeasurement.new output
+      measure.banner
+
+      words = {
+        hi: 1_000,
+        help: 100_000,
+        beautiful: 100_000,
+        impressionism: 200_000,
+        anthropological: 200_000,
+      }
+      trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+
+      [ trie, trie.clone.compress! ].each do |trie|
+        output.puts "==> #{trie.compressed? ? 'Compressed' : 'Uncompressed'}"
+        output.puts "`scan`"
+        words.each do |word, times|
+          measure.perform times, word.to_s do |word|
+            trie.scan(word).size
+          end
+        end
+      end
+    end
+  end
+
+  namespace :benchmark do
+    desc 'Generate lookups performance benchmark report'
+    task :lookups do
+      generate_lookups_benchmark
+    end
+
+    desc 'Generate scans performance benchmark report'
+    task :scans do
+      generate_scans_benchmark
+    end
+
+    namespace :lookups do
+      desc 'Generate performance benchmark report store results in reports/'
+      task save: ['performance:directory'] do
+        puts 'Generating performance benchmark report for lookups...'
+        generate_lookups_benchmark path('reports', Rambling::Trie::VERSION, 'benchmark')
+        puts "Benchmarks have been saved to reports/#{Rambling::Trie::VERSION}/benchmark"
+      end
+    end
+
+    task :creation do
+      with_file do |output|
+        measure = BenchmarkMeasurement.new output
+        measure.banner
+
+        output.puts '==> Creation'
+        output.puts '`Rambling::Trie.create`'
+        measure.perform 5 do
+          trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+        end
+      end
+    end
+
+    task :compression do
+      with_file do |output|
+        measure = BenchmarkMeasurement.new output
+        measure.banner
+
+        output.puts '==> Compression'
+        output.puts '`compress!`'
+
+        trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+        measure.perform 5 do
+          trie.clone.compress!
+        end
+      end
+    end
+
+    task all: [
+      'performance:benchmark:creation',
+      'performance:benchmark:compression',
+      'performance:benchmark:lookups',
+      'performance:benchmark:scans',
+    ]
+
+    task :compare do
+      Benchmark.ips do |b|
+        hash = { 'thing' => 'gniht' }
+
+        b.report 'has_key?' do
+          hash.has_key? 'thing'
+        end
+
+        b.report '[]' do
+          !!hash['thing']
+        end
+      end
+    end
+  end
+end

data/lib/rambling/trie/tasks/performance/directory.rb

@@ -0,0 +1,11 @@
+require 'fileutils'
+require_relative '../helpers/path'
+
+namespace :performance do
+  include Helpers::Path
+
+  desc 'Create report dir'
+  task :directory do
+    FileUtils.mkdir_p path('reports', Rambling::Trie::VERSION)
+  end
+end

data/lib/rambling/trie/tasks/performance/profile/call_tree.rb

@@ -0,0 +1,132 @@
+require_relative '../../helpers/path'
+require_relative '../../helpers/time'
+
+namespace :performance do
+  namespace :profile do
+    include Helpers::Path
+    include Helpers::Time
+
+    def profile times, params, path
+      params = Array params
+      params << nil unless params.any?
+
+      result = RubyProf.profile merge_fibers: true do
+        params.each do |param|
+          times.times do
+            yield param
+          end
+        end
+      end
+
+      printer = RubyProf::CallTreePrinter.new result
+      printer.print path: path
+    end
+
+    def generate_lookups_call_tree
+      puts 'Generating call tree profiling reports for lookups...'
+
+      puts "\nCall Tree profile for rambling-trie version #{Rambling::Trie::VERSION}"
+      trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+      tries = [ trie, trie.clone.compress! ]
+
+      words = %w(hi help beautiful impressionism anthropological)
+
+      tries.each do |trie|
+        filename = "profile-#{trie.compressed? ? 'compressed' : 'uncompressed'}-word"
+        path = path 'reports', Rambling::Trie::VERSION, 'call-tree', time, filename
+        FileUtils.mkdir_p path
+
+        profile 200_000, words, path do
+          trie.word? word
+        end
+
+        filename = "profile-#{trie.compressed? ? 'compressed' : 'uncompressed'}-partial-word"
+        path = path 'reports', Rambling::Trie::VERSION, 'call-tree', time, filename
+        FileUtils.mkdir_p path
+
+        profile 200_000, words, path do
+          trie.partial_word? word
+        end
+      end
+
+      puts 'Done'
+    end
+
+    def generate_scans_call_tree
+      puts 'Generating call tree profiling reports for scans...'
+
+      puts "\nCall Tree profile for rambling-trie version #{Rambling::Trie::VERSION}"
+      trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+      tries = [ trie, trie.clone.compress! ]
+
+      words = {
+        hi: 1_000,
+        help: 100_000,
+        beautiful: 100_000,
+        impressionism: 200_000,
+        anthropological: 200_000,
+      }
+
+      tries.each do |trie|
+        filename = "profile-#{trie.compressed? ? 'compressed' : 'uncompressed'}-scan"
+        path = path 'reports', Rambling::Trie::VERSION, 'call-tree', time, filename
+        FileUtils.mkdir_p path
+
+        words.each do |word, times|
+          profile times, word.to_s, path do |word|
+            trie.scan(word).size
+          end
+        end
+      end
+
+      puts 'Done'
+    end
+
+    namespace :call_tree do
+      desc 'Generate call tree profiling reports for creation'
+      task creation: ['performance:directory'] do
+        puts 'Generating call tree profiling reports for creation...'
+        puts "\nCall Tree profile for rambling-trie version #{Rambling::Trie::VERSION}"
+        filename = "profile-new-trie"
+        path = path 'reports', Rambling::Trie::VERSION, 'call-tree', time, filename
+        FileUtils.mkdir_p path
+
+        profile 5, nil, path do
+          trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+        end
+      end
+
+      desc 'Generate call tree profiling reports for compression'
+      task compression: ['performance:directory'] do
+        puts 'Generating call tree profiling reports for compression...'
+        puts "\nCall Tree profile for rambling-trie version #{Rambling::Trie::VERSION}"
+
+        filename = "profile-compressed-trie"
+        path = path 'reports', Rambling::Trie::VERSION, 'call-tree', time, filename
+        FileUtils.mkdir_p path
+
+        trie = Rambling::Trie.create path('assets', 'dictionaries', 'words_with_friends.txt')
+        profile 5, nil, path do
+          trie.clone.compress!
+        end
+      end
+
+      desc 'Generate call tree profiling reports for lookups'
+      task lookups: ['performance:directory'] do
+        generate_lookups_call_tree
+      end
+
+      desc 'Generate call tree profiling reports for scans'
+      task scans: ['performance:directory'] do
+        generate_scans_call_tree
+      end
+
+      task all: [
+        'performance:profile:call_tree:creation',
+        'performance:profile:call_tree:compression',
+        'performance:profile:call_tree:lookups',
+        'performance:profile:call_tree:scans',
+      ]
+    end
+  end
+end