rley 0.3.06 → 0.3.07

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 1bbd7b9c764c5f24a73b8f3f158e6c25f6f7ccd6
-  data.tar.gz: 63676212ceb56eb1b09f511a6fade48abcd7e3ab
+  metadata.gz: b9b70296885dc161ee4fdedf19ff1c2e04256086
+  data.tar.gz: 993b5b46f0edf140140d4713161c88b3cc846ca6
 SHA512:
-  metadata.gz: 2a17d1c3447e72f59f4f8268b7e90ccfda098fb0769b73219866d42fa77095e2e06f6bc980e6b0e5928ca1217c9c1eb8b5b973ff12b511fa3327745353caf5bd
-  data.tar.gz: e03b4dc4cebf8fca9049e367641248601850398ed77a725ab92e3cb22f2552224dc143a8121d142f0e38522a5cd1dc5ef570341542d0b39845a3ab7b02596a97
+  metadata.gz: c390b234b9412d2b48d92ad725bfad8860dea3c1b9d21ef75ffb6191f05cc23a3bccfa8975c6a79e9651b29fba345298e1c053454523390d118eeab9f8b6ad24
+  data.tar.gz: 5358fb0d0201b821be251d47f804dd9362ea970e4765462346658106ecf8e452e6088d239b0ce63073bc7bc9db9299018f523ac9024de6cac3b85e664ba36fe8

data/CHANGELOG.md

@@ -1,3 +1,8 @@
+### 0.3.07 / 2016-11-08
+* [FIX] The sharing a of forest node could be repeated in a production in a revisit event.
+* [CHANGE] Method `ParseWalkerFactory#process_end_entry`. Added a guard condition to avoid repeated node sharing
+* [NEW] RSpec file `ambiguous_parse_spec.rb` added in order to test the fix.
+
 ### 0.3.06 / 2016-11-06
 * [FIX] There were missing links to shared parse forest nodes for ambiguous parses.
 * [NEW] RSpec file `ambiguous_parse_spec.rb` added in order to test the parse forest building for an ambiguous parse.

data/lib/rley/constants.rb

@@ -3,7 +3,7 @@
 
 module Rley # Module used as a namespace
   # The version number of the gem.
-  Version = '0.3.06'.freeze
+  Version = '0.3.07'.freeze
 
   # Brief description of the gem.
   Description = "Ruby implementation of the Earley's parsing algorithm".freeze

data/lib/rley/parser/parse_forest_builder.rb

@@ -92,9 +92,10 @@ module Rley # This module is used as a namespace
             # Retrieve the already existing node corresponding 
             # to re-visited entry
             popular = @entry2node[anEntry]
-
-            # Share with parent
-            curr_parent.add_subnode(popular)
+            
+            # Share with parent (if needed)...
+            children = curr_parent.subnodes            
+            curr_parent.add_subnode(popular) unless children.include? popular
 
           else
             raise NotImplementedError

data/spec/rley/parser/groucho_spec.rb

@@ -0,0 +1,156 @@
+# Purpose: to test the parse forest generation for an emblematic
+# ambiguous sentence
+# Based on example found at: http://www.nltk.org/book_1ed/ch08.html
+require_relative '../../spec_helper'
+
+require_relative '../../../lib/rley/parser/gfg_earley_parser'
+require_relative '../../../lib/rley/parser/parse_walker_factory'
+
+require_relative '../support/grammar_helper'
+require_relative '../support/expectation_helper'
+require_relative '../support/grammar_ambig01_helper'
+
+# Load the class under test
+require_relative '../../../lib/rley/parser/parse_forest_builder'
+
+module Rley # Open this namespace to avoid module qualifier prefixes
+  module Parser
+    describe 'Coping with a NLP ambiguous toy grammar' do
+      include GrammarHelper     # Mix-in with token factory method
+      include ExpectationHelper # Mix-in with expectation on parse entry sets
+
+      let(:sample_grammar) do
+        builder = Rley::Syntax::GrammarBuilder.new
+        builder.add_terminals('N', 'V', 'Pro')  # N(oun), V(erb), Pro(noun)
+        builder.add_terminals('Det', 'P')       # Det(erminer), P(reposition)
+        builder.add_production('S' => %w(NP VP))
+        builder.add_production('NP' => %w(Det N))
+        builder.add_production('NP' => %w(Det N PP))
+        builder.add_production('NP' => 'Pro')
+        builder.add_production('VP' => %w(V NP))
+        builder.add_production('VP' => %w(VP PP))
+        builder.add_production('PP' => %w(P NP))
+        builder.grammar
+      end
+
+      # The lexicon is just a Hash with pairs of the form:
+      # word => terminal symbol name
+      Groucho_lexicon = {
+        'elephant' => 'N',
+        'pajamas' => 'N',
+        'shot' => 'V',
+        'I' => 'Pro',
+        'an' => 'Det',
+        'my' => 'Det',
+        'in' => 'P'
+      }.freeze
+
+      # Highly simplified tokenizer implementation.
+      def tokenizer(aText, aGrammar)
+        tokens = aText.scan(/\S+/).map do |word|
+          term_name = Groucho_lexicon[word]
+          if term_name.nil?
+            raise StandardError, "Word '#{word}' not found in lexicon"
+          end
+          terminal = aGrammar.name2symbol[term_name]
+          Rley::Parser::Token.new(word, terminal)
+        end
+
+        return tokens
+      end
+
+      let(:sentence_tokens) do
+        sentence = 'I shot an elephant in my pajamas'
+        tokenizer(sentence, sample_grammar)
+      end
+
+      let(:sentence_result) do
+        parser = Parser::GFGEarleyParser.new(sample_grammar)
+        parser.parse(sentence_tokens)
+      end
+
+      # Emit a text representation of the current path.
+      def path_to_s()
+        text_parts = subject.curr_path.map do |path_element|
+          path_element.to_string(0)
+        end
+        return text_parts.join('/')
+      end
+
+      def next_event(eventType, anEntryText)
+        event = @walker.next
+        subject.receive_event(*event)
+        expect(event[0]).to eq(eventType)
+        expect(event[1].to_s).to eq(anEntryText)
+      end
+
+      def expected_curr_parent(anExpectation)
+        expect(subject.curr_parent.to_string(0)).to eq(anExpectation)
+      end
+
+      def expected_curr_path(anExpectation)
+        expect(path_to_s).to eq(anExpectation)
+      end
+
+      def expected_first_child(anExpectation)
+          child = subject.curr_parent.subnodes.first
+          expect(child.to_string(0)).to eq(anExpectation)
+      end
+      
+      def root_children
+          subject.forest.root.subnodes    
+      end      
+
+
+      before(:each) do
+        factory = ParseWalkerFactory.new
+        accept_entry = sentence_result.accepting_entry
+        accept_index = sentence_result.chart.last_index
+        @walker = factory.build_walker(accept_entry, accept_index)
+      end
+
+      context 'Parse ambiguous sentence' do
+
+        subject { ParseForestBuilder.new(sentence_tokens) }
+
+        it 'should build a parse forest with a correct root node' do
+          next_event(:visit, 'S. | 0') # Event 1
+          expected_curr_path('S[0, 7]')
+          # Root node should have no child
+          expect(root_children.size).to be_zero
+
+          next_event(:visit, 'S => NP VP . | 0') # Event 2
+          expected_curr_path('S[0, 7]')
+
+          next_event(:visit, 'VP. | 1') # Event 3
+          expected_curr_path('S[0, 7]/VP[1, 7]')
+          # Root node should have one child          
+          expect(root_children.size).to eq(1)
+          expect(root_children.first.to_string(0)).to eq('VP[1, 7]')
+
+          25.times do
+            event = @walker.next
+            subject.receive_event(*event)
+          end
+
+          next_event(:visit, 'NP. | 0') # Event 29
+          expected_curr_path('S[0, 7]/NP[0, 1]')
+          # Root node should have two children
+          expect(root_children.size).to eq(2)
+          expect(root_children.first.to_string(0)).to eq('NP[0, 1]')
+          
+          18.times do
+            event = @walker.next
+            subject.receive_event(*event)
+          end
+         
+          next_event(:revisit, 'NP. | 0') # Event 48
+          expected_curr_path('S[0, 7]')
+          # Root node should still have two children          
+          expect(root_children.size).to eq(2)         
+        end
+      end # context
+    end # describe
+  end # module
+end # module
+# End of file

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rley
 version: !ruby/object:Gem::Version
-  version: 0.3.06
+  version: 0.3.07
 platform: ruby
 authors:
 - Dimitri Geshef
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-11-06 00:00:00.000000000 Z
+date: 2016-11-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -233,6 +233,7 @@ files:
 - spec/rley/parser/gfg_earley_parser_spec.rb
 - spec/rley/parser/gfg_parsing_spec.rb
 - spec/rley/parser/grm_items_builder_spec.rb
+- spec/rley/parser/groucho_spec.rb
 - spec/rley/parser/parse_entry_set_spec.rb
 - spec/rley/parser/parse_entry_spec.rb
 - spec/rley/parser/parse_forest_builder_spec.rb
@@ -320,6 +321,7 @@ test_files:
 - spec/rley/parser/gfg_earley_parser_spec.rb
 - spec/rley/parser/gfg_parsing_spec.rb
 - spec/rley/parser/grm_items_builder_spec.rb
+- spec/rley/parser/groucho_spec.rb
 - spec/rley/parser/parse_entry_set_spec.rb
 - spec/rley/parser/parse_entry_spec.rb
 - spec/rley/parser/parse_forest_builder_spec.rb