greeb 0.0.1

data/.gitignore

@@ -0,0 +1,35 @@
+## MAC OS
+.DS_Store
+
+## TEXTMATE
+*.tmproj
+tmtags
+
+## EMACS
+*~
+\#*
+.\#*
+
+## VIM
+*.swp
+
+## RUBINIUS
+*.rbc
+
+## NETBEANS
+nbproject
+
+## REDCAR
+.redcar
+
+## RVM
+.rvmrc
+
+## BUNDLER
+.bundle
+
+## PROJECT::GENERAL
+coverage
+pkg
+
+## PROJECT::SPECIFIC

data/Gemfile

@@ -0,0 +1,3 @@
+source 'http://rubygems.org'
+
+gemspec

data/Rakefile

@@ -0,0 +1,4 @@
+# encoding: utf-8
+
+require 'bundler'
+Bundler::GemHelper.install_tasks

data/greeb-test.rb

@@ -0,0 +1,141 @@
+#!/usr/bin/env ruby
+# encoding: utf-8
+
+require 'rubygems'
+require 'graphviz'
+
+$:.unshift('./lib')
+require 'greeb'
+
+origin = <<-END
+ - Сынок, чего это     от тебя   зигами пахнет,
+опять на Манежную площадь ходил?
+
+ - Нет мама, я в метро ехал, там  назиговано было!!
+
+  
+
+Четырнадцать, восемьдесять восемь: 14/88.
+Вот так блять
+END
+origin.chomp!
+
+def identify(token)
+  case token
+    when Greeb::RU_LEX then 'RU_LEX'
+    when Greeb::EN_LEX then 'EN_LEX'
+    when Greeb::EOL then 'EOL'
+    when Greeb::SEP then 'SEP'
+    when Greeb::PUN then 'PUN'
+    when Greeb::SPUN then 'SPUN'
+    when Greeb::DIG then 'DIG'
+    when Greeb::DIL then 'DIL'
+  else
+    '?!'
+  end
+end
+
+greeb = Greeb::Parser.new(origin)
+text = greeb.tree
+
+g = GraphViz.new('graphematics', 'type' => 'graph')
+
+g.node[:color]    = '#ddaa66'
+g.node[:style]    = 'filled'
+g.node[:shape]    = 'box'
+g.node[:penwidth] = '1'
+g.node[:fontname] = 'PT Sans'
+g.node[:fontsize] = '8'
+g.node[:fillcolor]= '#ffeecc'
+g.node[:fontcolor]= '#775500'
+g.node[:margin]   = '0.0'
+
+g.edge[:color]    = '#999999'
+g.edge[:weight]   = '1'
+g.edge[:fontname] = 'PT Sans'
+g.edge[:fontcolor]= '#444444'
+g.edge[:fontsize] = '6'
+g.edge[:dir]      = 'forward'
+g.edge[:arrowsize]= '0.5'
+
+bid = 'begin'
+g.add_node(bid).tap do |node|
+  node.label = "Начало\nтекста"
+  node.shape = 'ellipse'
+  node.style = ''
+end
+
+eid = 'end'
+g.add_node(eid).tap do |node|
+  node.label = "Конец\nтекста"
+  node.shape = 'ellipse'
+  node.style = ''
+end
+
+tree = text.map_with_index do |paragraph, i|
+  pid = "p#{i}"
+  sentences = paragraph.map_with_index do |sentence, j|
+    sid = "#{pid}s#{j}"
+    subsentences = sentence.map_with_index do |subsentence, k|
+      ssid = "#{sid}ss#{k}"
+      tokens = subsentence.map_with_index do |token, l|
+        next if ' ' == token
+        [ "#{ssid}t#{l}", token, l ]
+      end
+      tokens.delete(nil)
+      [ ssid, tokens, k ]
+    end
+    [ sid, subsentences, j ]
+  end
+  [ pid, sentences, i ]
+end
+
+tree.each do |pid, paragraph, i|
+  g.add_node(pid).tap do |node|
+    node.label = "Абзац\n№#{i + 1}"
+    node.shape = 'ellipse'
+  end
+  g.add_edge(bid, pid)
+
+  paragraph.each do |sid, sentence, j|
+    g.add_node(sid).tap do |node|
+      node.label = "Предложение\n№#{j + 1}"
+      node.shape = 'ellipse'
+    end
+    g.add_edge(pid, sid)
+
+    sentence.each do |ssid, subsentence, k|
+      g.add_node(ssid).tap do |node|
+        node.label = "Подпредложение\n№#{k + 1}"
+        node.shape = 'ellipse'
+      end
+      g.add_edge(sid, ssid)
+
+      subsentence.each do |tid, token, l|
+        g.add_node(tid).label = token
+        g.add_edge(ssid, tid).label = identify(token)
+        g.add_edge(tid, eid)
+      end
+
+      subsentence.each_cons(2) do |(tid1, token1, l1),
+                                   (tid2, token2, l2)|
+        g.add_edge(tid1, tid2).tap do |edge|
+          edge.weight = 0.25
+          edge.style = 'dashed'
+        end
+      end
+    end
+
+    sentence.each_cons(2) do |(ssid1, subsentence1, k1),
+                              (ssid2, subsentence2, k2)|
+      tid1, token1, l1 = subsentence1.last
+      tid2, token2, l2 = subsentence2.first
+      g.add_edge(tid1, tid2).tap do |edge|
+        edge.weight = 0.5
+        edge.style = 'dashed'
+      end
+    end
+  end
+end
+
+g.output(:output => 'png', :file => 'graph.png')

data/greeb.gemspec

@@ -0,0 +1,23 @@
+# encoding: utf-8
+
+$:.push File.expand_path('../lib', __FILE__)
+require 'greeb/version'
+
+Gem::Specification.new do |s|
+  s.name        = 'greeb'
+  s.version     = Greeb::VERSION
+  s.platform    = Gem::Platform::RUBY
+  s.authors     = [ 'Dmitry A. Ustalov' ]
+  s.email       = [ 'dmitry@eveel.ru' ]
+  s.homepage    = 'https://github.com/eveel/greeb'
+  s.summary     = 'Greeb is a Graphematical Analyzer.'
+  s.description = 'Greeb is a Graphematical Analyzer, ' \
+                  'written in Ruby.'
+
+  s.rubyforge_project = 'greeb'
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = [ 'lib' ]
+end

data/lib/enumerable.rb

@@ -0,0 +1,8 @@
+# encoding: utf-8
+
+module Enumerable
+  def collect_with_index(i = -1)
+    collect { |e| yield(e, i += 1) }
+  end
+  alias map_with_index collect_with_index
+end

data/lib/greeb.rb

@@ -0,0 +1,144 @@
+# encoding: utf-8
+
+require 'meta_array'
+require 'enumerable'
+
+module Greeb
+  RU_LEX = /^[А-Яа-я]+$/u
+  EN_LEX = /^[A-Za-z]+$/u
+  EOL = /^\n+$/u
+  SEP = /^[*=_\/\\ ]$/u
+  PUN = /^(\.|\!|\?)$/u
+  SPUN = /^(\,|\[|\]|\(|\)|\-|:|;)$/u
+  DIG = /^[0-9]+$/u
+  DIL = /^[А-Яа-яA-Za-z0-9]+$/u
+  EMPTY = ''
+
+  class Parser
+    attr_accessor :origin
+    private :origin=
+
+    attr_writer :tree
+    private :tree=
+
+    def initialize(origin)
+      self.origin = origin
+    end
+
+    def tree
+      @tree ||= parse(origin)
+    end
+
+    private
+      def parse(origin) # :nodoc:
+        tree = MetaArray.new
+
+        # paragraph
+        p_id = 0
+
+        # sentence
+        s_id = 0
+
+        # subsentence
+        ss_id = 0
+
+        token = ''
+
+        origin.each_char do |c|
+          puts "[#{token.inspect}] ← #{c.inspect}"
+          case c
+            when EOL then begin
+              case token
+                when EMPTY then token << c
+                when EOL then begin
+                  token = ''
+                  p_id += 1
+                  s_id = 0
+                  ss_id = 0
+                end
+              else
+                tree[p_id][s_id][ss_id] << token
+                token = c
+              end
+            end
+            when SEP then begin
+              case token
+                when EMPTY
+              else
+                tree[p_id][s_id][ss_id] << token
+                while tree[p_id][s_id][ss_id].last == c
+                  tree[p_id][s_id][ss_id].pop
+                end
+                tree[p_id][s_id][ss_id] << c
+                token = ''
+              end
+            end
+            when PUN then begin
+              case token
+                when EMPTY
+              else
+                tree[p_id][s_id][ss_id] << token
+                tree[p_id][s_id][ss_id] << c
+                token = ''
+                s_id += 1
+                ss_id = 0
+              end
+            end
+            when SPUN then begin
+              case token
+                when EMPTY
+              else
+                tree[p_id][s_id][ss_id] << token
+                tree[p_id][s_id][ss_id] << c
+                token = ''
+                ss_id += 1
+              end
+            end
+            when RU_LEX then begin
+              case token
+                when EOL then begin
+                  tree[p_id][s_id][ss_id] << ' '
+                  token = c
+                end
+              else
+                token << c
+              end
+            end
+            when EN_LEX then begin
+              case token
+                when EOL then begin
+                  tree[p_id][s_id][ss_id] << ' '
+                  token = c
+                end
+              else
+                token << c
+              end
+            end
+            when DIG then begin
+              case token
+                when EOL then begin
+                  tree[p_id][s_id][ss_id] << ' '
+                  token = c
+                end
+              else
+                token << c
+              end
+            end
+            when DIL then begin
+              case token
+                when EOL then begin
+                  tree[p_id][s_id][ss_id] << token
+                  token = c
+                end
+              else
+                token << c
+              end
+            end
+          end
+        end
+        tree[p_id][s_id][ss_id] << token
+        tree.delete(nil)
+        tree.to_a
+      end
+  end
+end

data/lib/greeb/version.rb

@@ -0,0 +1,5 @@
+# encoding: utf-8
+
+module Greeb
+  VERSION = "0.0.1"
+end

data/lib/meta_array.rb

@@ -0,0 +1,11 @@
+# encoding: utf-8
+
+class MetaArray < Array
+  def [] id
+    super(id) or begin
+      self.class.new.tap do |element|
+        self[id] = element
+      end
+    end
+  end
+end

metadata

@@ -0,0 +1,57 @@
+--- !ruby/object:Gem::Specification
+name: greeb
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+  prerelease: !!null 
+platform: ruby
+authors:
+- Dmitry A. Ustalov
+autorequire: !!null 
+bindir: bin
+cert_chain: []
+date: 2011-02-06 00:00:00.000000000 +05:00
+default_executable: !!null 
+dependencies: []
+description: Greeb is a Graphematical Analyzer, written in Ruby.
+email:
+- dmitry@eveel.ru
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- .gitignore
+- Gemfile
+- README
+- Rakefile
+- greeb-test.rb
+- greeb.gemspec
+- lib/enumerable.rb
+- lib/greeb.rb
+- lib/greeb/version.rb
+- lib/meta_array.rb
+has_rdoc: true
+homepage: https://github.com/eveel/greeb
+licenses: []
+post_install_message: !!null 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: greeb
+rubygems_version: 1.5.0
+signing_key: !!null 
+specification_version: 3
+summary: Greeb is a Graphematical Analyzer.
+test_files: []