rley 0.3.04 → 0.3.05
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.rubocop.yml +7 -3
- data/CHANGELOG.md +3 -0
- data/Rakefile +30 -30
- data/examples/parsers/parsing_L0.rb +1 -1
- data/examples/parsers/parsing_L1.rb +1 -1
- data/examples/parsers/parsing_abc.rb +1 -1
- data/examples/parsers/parsing_ambig.rb +1 -1
- data/examples/parsers/parsing_another.rb +1 -1
- data/examples/parsers/parsing_b_expr.rb +1 -1
- data/examples/parsers/parsing_err_expr.rb +1 -1
- data/examples/parsers/parsing_groucho.rb +1 -1
- data/examples/parsers/parsing_right_recursive.rb +1 -1
- data/examples/parsers/parsing_tricky.rb +1 -1
- data/lib/rley/constants.rb +2 -2
- data/lib/rley/formatter/base_formatter.rb +0 -2
- data/lib/rley/formatter/debug.rb +0 -2
- data/lib/rley/formatter/json.rb +1 -3
- data/lib/rley/gfg/call_edge.rb +31 -30
- data/lib/rley/gfg/edge.rb +22 -23
- data/lib/rley/gfg/end_vertex.rb +22 -24
- data/lib/rley/gfg/epsilon_edge.rb +20 -21
- data/lib/rley/gfg/grm_flow_graph.rb +39 -39
- data/lib/rley/gfg/item_vertex.rb +16 -17
- data/lib/rley/gfg/non_terminal_vertex.rb +3 -4
- data/lib/rley/gfg/return_edge.rb +32 -31
- data/lib/rley/gfg/scan_edge.rb +25 -26
- data/lib/rley/gfg/shortcut_edge.rb +25 -26
- data/lib/rley/gfg/start_vertex.rb +0 -2
- data/lib/rley/gfg/vertex.rb +8 -8
- data/lib/rley/parse_forest_visitor.rb +113 -115
- data/lib/rley/parse_tree_visitor.rb +0 -2
- data/lib/rley/parser/base_parser.rb +27 -27
- data/lib/rley/parser/chart.rb +14 -14
- data/lib/rley/parser/dotted_item.rb +33 -33
- data/lib/rley/parser/earley_parser.rb +6 -6
- data/lib/rley/parser/gfg_chart.rb +8 -15
- data/lib/rley/parser/gfg_earley_parser.rb +15 -13
- data/lib/rley/parser/gfg_parsing.rb +26 -22
- data/lib/rley/parser/grm_items_builder.rb +3 -2
- data/lib/rley/parser/parse_entry.rb +3 -9
- data/lib/rley/parser/parse_entry_set.rb +14 -19
- data/lib/rley/parser/parse_entry_tracker.rb +56 -56
- data/lib/rley/parser/parse_forest_builder.rb +215 -214
- data/lib/rley/parser/parse_forest_factory.rb +57 -56
- data/lib/rley/parser/parse_state.rb +8 -11
- data/lib/rley/parser/parse_state_tracker.rb +56 -56
- data/lib/rley/parser/parse_tracer.rb +3 -3
- data/lib/rley/parser/parse_tree_builder.rb +10 -10
- data/lib/rley/parser/parse_walker_factory.rb +30 -33
- data/lib/rley/parser/parsing.rb +8 -8
- data/lib/rley/parser/state_set.rb +23 -26
- data/lib/rley/ptree/non_terminal_node.rb +1 -1
- data/lib/rley/ptree/token_range.rb +2 -2
- data/lib/rley/sppf/alternative_node.rb +32 -34
- data/lib/rley/sppf/composite_node.rb +27 -27
- data/lib/rley/sppf/epsilon_node.rb +26 -27
- data/lib/rley/sppf/leaf_node.rb +11 -12
- data/lib/rley/sppf/non_terminal_node.rb +37 -38
- data/lib/rley/sppf/sppf_node.rb +1 -1
- data/lib/rley/sppf/token_node.rb +29 -29
- data/lib/rley/syntax/grammar.rb +1 -3
- data/lib/rley/syntax/grammar_builder.rb +8 -8
- data/lib/rley/syntax/non_terminal.rb +2 -4
- data/lib/rley/syntax/production.rb +3 -3
- data/lib/rley/syntax/symbol_seq.rb +1 -1
- data/spec/rley/gfg/call_edge_spec.rb +50 -51
- data/spec/rley/gfg/edge_spec.rb +33 -33
- data/spec/rley/gfg/end_vertex_spec.rb +26 -27
- data/spec/rley/gfg/epsilon_edge_spec.rb +25 -25
- data/spec/rley/gfg/grm_flow_graph_spec.rb +1 -1
- data/spec/rley/gfg/item_vertex_spec.rb +3 -4
- data/spec/rley/gfg/return_edge_spec.rb +51 -51
- data/spec/rley/gfg/scan_edge_spec.rb +32 -30
- data/spec/rley/gfg/shortcut_edge_spec.rb +1 -1
- data/spec/rley/gfg/vertex_spec.rb +3 -3
- data/spec/rley/parse_forest_visitor_spec.rb +239 -238
- data/spec/rley/parser/dotted_item_spec.rb +1 -1
- data/spec/rley/parser/earley_parser_spec.rb +16 -16
- data/spec/rley/parser/gfg_earley_parser_spec.rb +30 -31
- data/spec/rley/parser/gfg_parsing_spec.rb +11 -10
- data/spec/rley/parser/grm_items_builder_spec.rb +2 -2
- data/spec/rley/parser/parse_entry_set_spec.rb +4 -4
- data/spec/rley/parser/parse_entry_spec.rb +0 -2
- data/spec/rley/parser/parse_forest_builder_spec.rb +82 -57
- data/spec/rley/parser/parse_forest_factory_spec.rb +84 -82
- data/spec/rley/parser/parse_walker_factory_spec.rb +10 -9
- data/spec/rley/parser/parsing_spec.rb +0 -1
- data/spec/rley/sppf/alternative_node_spec.rb +2 -2
- data/spec/rley/sppf/non_terminal_node_spec.rb +0 -1
- data/spec/rley/support/ambiguous_grammar_helper.rb +1 -1
- data/spec/rley/support/expectation_helper.rb +37 -36
- data/spec/rley/support/grammar_abc_helper.rb +17 -17
- data/spec/rley/support/grammar_b_expr_helper.rb +40 -39
- data/spec/rley/support/grammar_helper.rb +2 -1
- data/spec/rley/support/{grammar_L0_helper.rb → grammar_l0_helper.rb} +82 -81
- data/spec/rley/support/grammar_sppf_helper.rb +24 -25
- data/spec/rley/syntax/grammar_spec.rb +1 -1
- metadata +2 -2
@@ -1,81 +1,82 @@
|
|
1
|
-
# Load the builder class
|
2
|
-
require_relative '../../../lib/rley/syntax/grammar_builder'
|
3
|
-
require_relative '../../../lib/rley/parser/token'
|
4
|
-
|
5
|
-
|
6
|
-
module GrammarL0Helper
|
7
|
-
########################################
|
8
|
-
# Factory method. Define a grammar for a micro English-like language
|
9
|
-
# based on Jurafky & Martin L0 language (chapter 12 of the book).
|
10
|
-
# It defines the syntax of a sentence in a language with a
|
11
|
-
# very limited syntax and lexicon in the context of airline reservation.
|
12
|
-
def
|
13
|
-
builder = Rley::Syntax::GrammarBuilder.new
|
14
|
-
builder.add_terminals('Noun', 'Verb', 'Pronoun', 'Proper-Noun')
|
15
|
-
builder.add_terminals('Determiner', 'Preposition'
|
16
|
-
builder.add_production('S' => %w
|
17
|
-
builder.add_production('NP' => 'Pronoun')
|
18
|
-
builder.add_production('NP' => 'Proper-Noun')
|
19
|
-
builder.add_production('NP' => %w
|
20
|
-
builder.add_production('Nominal' => %w
|
21
|
-
builder.add_production('Nominal' => 'Noun')
|
22
|
-
builder.add_production('VP' => 'Verb')
|
23
|
-
builder.add_production('VP' => %w
|
24
|
-
builder.add_production('VP' => %w
|
25
|
-
builder.add_production('VP' => %w
|
26
|
-
builder.add_production('PP' => %w
|
27
|
-
builder
|
28
|
-
end
|
29
|
-
|
30
|
-
# Return the language lexicon.
|
31
|
-
# A lexicon is just a Hash with pairs of the form:
|
32
|
-
# word => terminal symbol name
|
33
|
-
def
|
34
|
-
|
35
|
-
'flight' => 'Noun',
|
36
|
-
'breeze' => 'Noun',
|
37
|
-
'trip' => 'Noun',
|
38
|
-
'morning' => 'Noun',
|
39
|
-
'is' => 'Verb',
|
40
|
-
'prefer' => 'Verb',
|
41
|
-
'like' => 'Verb',
|
42
|
-
'need' => 'Verb',
|
43
|
-
'want' => 'Verb',
|
44
|
-
'fly' => 'Verb',
|
45
|
-
'me' => 'Pronoun',
|
46
|
-
'I' => 'Pronoun',
|
47
|
-
'you' => 'Pronoun',
|
48
|
-
'it' => 'Pronoun',
|
49
|
-
'Alaska' => 'Proper-Noun',
|
50
|
-
'Baltimore' => 'Proper-Noun',
|
51
|
-
'Chicago' => 'Proper-Noun',
|
52
|
-
'United' => 'Proper-Noun',
|
53
|
-
'American' => 'Proper-Noun',
|
54
|
-
'the' => 'Determiner',
|
55
|
-
'a' => 'Determiner',
|
56
|
-
'an' => 'Determiner',
|
57
|
-
'this' => 'Determiner',
|
58
|
-
'these' => 'Determiner',
|
59
|
-
'that' => 'Determiner',
|
60
|
-
'from' => 'Preposition',
|
61
|
-
'to' => 'Preposition',
|
62
|
-
'on' => 'Preposition',
|
63
|
-
'near' => 'Preposition'
|
64
|
-
}
|
65
|
-
end
|
66
|
-
|
67
|
-
|
68
|
-
# Highly simplified tokenizer implementation.
|
69
|
-
def
|
70
|
-
tokens = aText.scan(/\S+/).map do |word|
|
71
|
-
term_name =
|
72
|
-
if term_name.nil?
|
73
|
-
|
74
|
-
end
|
75
|
-
terminal = aGrammar.name2symbol[term_name]
|
76
|
-
Rley::Parser::Token.new(word, terminal)
|
77
|
-
end
|
78
|
-
|
79
|
-
return tokens
|
80
|
-
end
|
81
|
-
end # module
|
1
|
+
# Load the builder class
|
2
|
+
require_relative '../../../lib/rley/syntax/grammar_builder'
|
3
|
+
require_relative '../../../lib/rley/parser/token'
|
4
|
+
|
5
|
+
|
6
|
+
module GrammarL0Helper
|
7
|
+
########################################
|
8
|
+
# Factory method. Define a grammar for a micro English-like language
|
9
|
+
# based on Jurafky & Martin L0 language (chapter 12 of the book).
|
10
|
+
# It defines the syntax of a sentence in a language with a
|
11
|
+
# very limited syntax and lexicon in the context of airline reservation.
|
12
|
+
def grammar_l0_builder()
|
13
|
+
builder = Rley::Syntax::GrammarBuilder.new
|
14
|
+
builder.add_terminals('Noun', 'Verb', 'Pronoun', 'Proper-Noun')
|
15
|
+
builder.add_terminals('Determiner', 'Preposition')
|
16
|
+
builder.add_production('S' => %w(NP VP))
|
17
|
+
builder.add_production('NP' => 'Pronoun')
|
18
|
+
builder.add_production('NP' => 'Proper-Noun')
|
19
|
+
builder.add_production('NP' => %w(Determiner Nominal))
|
20
|
+
builder.add_production('Nominal' => %w(Nominal Noun))
|
21
|
+
builder.add_production('Nominal' => 'Noun')
|
22
|
+
builder.add_production('VP' => 'Verb')
|
23
|
+
builder.add_production('VP' => %w(Verb NP))
|
24
|
+
builder.add_production('VP' => %w(Verb NP PP))
|
25
|
+
builder.add_production('VP' => %w(Verb PP))
|
26
|
+
builder.add_production('PP' => %w(Preposition PP))
|
27
|
+
builder
|
28
|
+
end
|
29
|
+
|
30
|
+
# Return the language lexicon.
|
31
|
+
# A lexicon is just a Hash with pairs of the form:
|
32
|
+
# word => terminal symbol name
|
33
|
+
def lexicon_l0()
|
34
|
+
return {
|
35
|
+
'flight' => 'Noun',
|
36
|
+
'breeze' => 'Noun',
|
37
|
+
'trip' => 'Noun',
|
38
|
+
'morning' => 'Noun',
|
39
|
+
'is' => 'Verb',
|
40
|
+
'prefer' => 'Verb',
|
41
|
+
'like' => 'Verb',
|
42
|
+
'need' => 'Verb',
|
43
|
+
'want' => 'Verb',
|
44
|
+
'fly' => 'Verb',
|
45
|
+
'me' => 'Pronoun',
|
46
|
+
'I' => 'Pronoun',
|
47
|
+
'you' => 'Pronoun',
|
48
|
+
'it' => 'Pronoun',
|
49
|
+
'Alaska' => 'Proper-Noun',
|
50
|
+
'Baltimore' => 'Proper-Noun',
|
51
|
+
'Chicago' => 'Proper-Noun',
|
52
|
+
'United' => 'Proper-Noun',
|
53
|
+
'American' => 'Proper-Noun',
|
54
|
+
'the' => 'Determiner',
|
55
|
+
'a' => 'Determiner',
|
56
|
+
'an' => 'Determiner',
|
57
|
+
'this' => 'Determiner',
|
58
|
+
'these' => 'Determiner',
|
59
|
+
'that' => 'Determiner',
|
60
|
+
'from' => 'Preposition',
|
61
|
+
'to' => 'Preposition',
|
62
|
+
'on' => 'Preposition',
|
63
|
+
'near' => 'Preposition'
|
64
|
+
}
|
65
|
+
end
|
66
|
+
|
67
|
+
|
68
|
+
# Highly simplified tokenizer implementation.
|
69
|
+
def tokenizer_l0(aText, aGrammar)
|
70
|
+
tokens = aText.scan(/\S+/).map do |word|
|
71
|
+
term_name = lexicon_l0[word]
|
72
|
+
if term_name.nil?
|
73
|
+
raise StandardError, "Word '#{word}' not found in lexicon"
|
74
|
+
end
|
75
|
+
terminal = aGrammar.name2symbol[term_name]
|
76
|
+
Rley::Parser::Token.new(word, terminal)
|
77
|
+
end
|
78
|
+
|
79
|
+
return tokens
|
80
|
+
end
|
81
|
+
end # module
|
82
|
+
# End of file
|
@@ -1,25 +1,24 @@
|
|
1
|
-
# Load the builder class
|
2
|
-
require_relative '../../../lib/rley/syntax/grammar_builder'
|
3
|
-
|
4
|
-
|
5
|
-
module GrammarSPPFHelper
|
6
|
-
# Factory method. Creates a grammar builder for a
|
7
|
-
# grammar described in paper from Elisabeth Scott
|
8
|
-
# "SPPF-Style Parsing From Earley Recognizers" in
|
9
|
-
# Notes in Theoretical Computer Science 203, (2008), pp. 53-67
|
10
|
-
# contains a hidden left recursion and a cycle
|
11
|
-
def grammar_sppf_builder()
|
12
|
-
builder = Rley::Syntax::GrammarBuilder.new
|
13
|
-
builder.add_terminals('a', 'b')
|
14
|
-
builder.add_production('Phi' =>
|
15
|
-
builder.add_production('S' => %w
|
16
|
-
builder.add_production('S' => %w
|
17
|
-
builder.add_production('A' => 'a')
|
18
|
-
builder.add_production('A' => %w
|
19
|
-
builder.add_production('B' => [])
|
20
|
-
builder.add_production('T' => %w(
|
21
|
-
|
22
|
-
return builder
|
23
|
-
end
|
24
|
-
|
25
|
-
end # module
|
1
|
+
# Load the builder class
|
2
|
+
require_relative '../../../lib/rley/syntax/grammar_builder'
|
3
|
+
|
4
|
+
|
5
|
+
module GrammarSPPFHelper
|
6
|
+
# Factory method. Creates a grammar builder for a
|
7
|
+
# grammar described in paper from Elisabeth Scott
|
8
|
+
# "SPPF-Style Parsing From Earley Recognizers" in
|
9
|
+
# Notes in Theoretical Computer Science 203, (2008), pp. 53-67
|
10
|
+
# contains a hidden left recursion and a cycle
|
11
|
+
def grammar_sppf_builder()
|
12
|
+
builder = Rley::Syntax::GrammarBuilder.new
|
13
|
+
builder.add_terminals('a', 'b')
|
14
|
+
builder.add_production('Phi' => 'S')
|
15
|
+
builder.add_production('S' => %w(A T))
|
16
|
+
builder.add_production('S' => %w(a T))
|
17
|
+
builder.add_production('A' => 'a')
|
18
|
+
builder.add_production('A' => %w(B A))
|
19
|
+
builder.add_production('B' => [])
|
20
|
+
builder.add_production('T' => %w(b b b))
|
21
|
+
|
22
|
+
return builder
|
23
|
+
end
|
24
|
+
end # module
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: rley
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.3.
|
4
|
+
version: 0.3.05
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Dimitri Geshef
|
@@ -252,10 +252,10 @@ files:
|
|
252
252
|
- spec/rley/sppf/non_terminal_node_spec.rb
|
253
253
|
- spec/rley/support/ambiguous_grammar_helper.rb
|
254
254
|
- spec/rley/support/expectation_helper.rb
|
255
|
-
- spec/rley/support/grammar_L0_helper.rb
|
256
255
|
- spec/rley/support/grammar_abc_helper.rb
|
257
256
|
- spec/rley/support/grammar_b_expr_helper.rb
|
258
257
|
- spec/rley/support/grammar_helper.rb
|
258
|
+
- spec/rley/support/grammar_l0_helper.rb
|
259
259
|
- spec/rley/support/grammar_sppf_helper.rb
|
260
260
|
- spec/rley/syntax/grammar_builder_spec.rb
|
261
261
|
- spec/rley/syntax/grammar_spec.rb
|