Class: EBNF::Rule

Inherits:
Object
  • Object
show all
Defined in:
lib/ebnf/rule.rb

Overview

Represent individual parsed rules

Constant Summary collapse

BNF_OPS =

Operations which are flattened to seprate rules in to_bnf

%w{
  alt opt plus seq star
}.map(&:to_sym).freeze
TERM_OPS = 
%w{
  diff hex range
}.map(&:to_sym).freeze

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(sym, id, expr, kind: nil, ebnf: nil, first: nil, follow: nil, start: nil, top_rule: nil, cleanup: nil) ⇒ Rule

Returns a new instance of Rule.

Parameters:

  • id (Integer)
  • sym (Symbol)
  • expr (Array)
  • :kind (Symbol)
  • :ebnf (String)
  • :first (Array)
  • :follow (Array)
  • :start (Boolean)
  • :top_rule (Rule)
  • :cleanup (Boolean) 


70
71
72
73
74
75
76
77
78
79
80
 
# File 'lib/ebnf/rule.rb', line 70

def initialize(sym, id, expr, kind: nil, ebnf: nil, first: nil, follow: nil, start: nil, top_rule: nil, cleanup: nil)
  @sym, @id = sym, id
  @expr = expr.is_a?(Array) ? expr : [:seq, expr]
  @ebnf, @kind, @first, @follow, @start, @cleanup, @top_rule = ebnf, kind, first, follow, start, cleanup, top_rule
  @top_rule ||= self
  @kind ||= case
  when sym.to_s == sym.to_s.upcase then :terminal
  when !BNF_OPS.include?(@expr.first) then :terminal
  else :rule
  end
end

Instance Attribute Details

#cleanup ⇒ Object

Determines preparation and cleanup rules for reconstituting EBNF ? * + from BNF



58
59
60
 
# File 'lib/ebnf/rule.rb', line 58

def cleanup
  @cleanup
end

#comp ⇒ Rule

A comprehension is a sequence which contains all elements but the first of the original rule.

Returns:



25
26
27
 
# File 'lib/ebnf/rule.rb', line 25

def comp
  @comp
end

#expr ⇒ Array

Rule expression

Returns:

  • (Array) 


35
36
37
 
# File 'lib/ebnf/rule.rb', line 35

def expr
  @expr
end

#first ⇒ Array<Rule> (readonly)

Terminals that immediately procede this rule

Returns:



45
46
47
 
# File 'lib/ebnf/rule.rb', line 45

def first
  @first
end

#follow ⇒ Array<Rule> (readonly)

Terminals that immediately follow this rule

Returns:



50
51
52
 
# File 'lib/ebnf/rule.rb', line 50

def follow
  @follow
end

#id ⇒ String

ID of rule

Returns:

  • (String) 


20
21
22
 
# File 'lib/ebnf/rule.rb', line 20

def id
  @id
end

#kind ⇒ :rule, ...

Kind of rule

Returns:

  • (:rule, :terminal, or :pass) 


30
31
32
 
# File 'lib/ebnf/rule.rb', line 30

def kind
  @kind
end

#orig ⇒ String

Original EBNF

Returns:

  • (String) 


40
41
42
 
# File 'lib/ebnf/rule.rb', line 40

def orig
  @orig
end

#start ⇒ Boolean

Indicates that this is a starting rule

Returns:

  • (Boolean) 


55
56
57
 
# File 'lib/ebnf/rule.rb', line 55

def start
  @start
end

#sym ⇒ Symbol

Symbol of rule

Returns:

  • (Symbol) 


16
17
18
 
# File 'lib/ebnf/rule.rb', line 16

def sym
  @sym
end

Class Method Details

.from_sxp(sxp) ⇒ Rule

Return a rule from its SXP representation:

Also may have (first …), (follow …), or (start #t)

Examples:

inputs

(pass (plus (range "#x20\\t\\r\\n")))
(rule ebnf "1" (star (alt declaration rule)))
(terminal O_ENUM "17" (seq "[^" (plus CHAR) "]"))

Parameters:

  • sxp (Array)

Returns:



94
95
96
97
98
99
100
101
102
103
104
105
106
 
# File 'lib/ebnf/rule.rb', line 94

def self.from_sxp(sxp)
  expr = sxp.detect {|e| e.is_a?(Array) && ![:first, :follow, :start].include?(e.first.to_sym)}
  first = sxp.detect {|e| e.is_a?(Array) && e.first.to_sym == :first}
  first = first[1..-1] if first
  follow = sxp.detect {|e| e.is_a?(Array) && e.first.to_sym == :follow}
  follow = follow[1..-1] if follow
  cleanup = sxp.detect {|e| e.is_a?(Array) && e.first.to_sym == :cleanup}
  cleanup = cleanup[1..-1] if cleanup
  start = sxp.any? {|e| e.is_a?(Array) && e.first.to_sym == :start}
  sym = sxp[1] if sxp[1].is_a?(Symbol)
  id = sxp[2] if sxp[2].is_a?(String)
  Rule.new(sym, id, expr, kind: sxp.first, first: first, follow: follow, cleanup: cleanup, start: start)
end

Instance Method Details

#<=>(other) ⇒ Object

Rules compare using their ids



386
387
388
389
390
391
392
 
# File 'lib/ebnf/rule.rb', line 386

def <=>(other)
  if id.to_i == other.id.to_i
    id.to_s <=> other.id.to_s
  else
    id.to_i <=> other.id.to_i
  end
end

#==(other) ⇒ Boolean

Two rules are equal if they have the same #sym, #kind and #expr

Parameters:

Returns:

  • (Boolean) 


357
358
359
360
361
 
# File 'lib/ebnf/rule.rb', line 357

def ==(other)
  sym   == other.sym &&
  kind  == other.kind &&
  expr  == other.expr
end

#add_first(terminals) ⇒ Integer

Add terminal as proceding this rule

Parameters:

  • terminals (Array<Rule, Symbol, String>)

Returns:

  • (Integer) —

    if number of terminals added



294
295
296
297
298
299
 
# File 'lib/ebnf/rule.rb', line 294

def add_first(terminals)
  @first ||= []
  terminals = terminals.map {|t| t.is_a?(Rule) ? t.sym : t} - @first
  @first += terminals
  terminals.length
end

#add_follow(terminals) ⇒ Integer

Add terminal as following this rule. Don’t add _eps as a follow

Parameters:

  • terminals (Array<Rule, Symbol, String>)

Returns:

  • (Integer) —

    if number of terminals added



305
306
307
308
309
310
311
312
313
 
# File 'lib/ebnf/rule.rb', line 305

def add_follow(terminals)
  # Remove terminals already in follows, and empty string
  terminals = terminals.map {|t| t.is_a?(Rule) ? t.sym : t} - (@follow || []) - [:_eps]
  unless terminals.empty?
    @follow ||= []
    @follow += terminals
  end
  terminals.length
end

#alt? ⇒ Boolean

Is this rule of the form (alt …)?

Returns:

  • (Boolean) 


334
335
336
 
# File 'lib/ebnf/rule.rb', line 334

def alt?
  expr.is_a?(Array) && expr.first == :alt
end

#build(expr, kind: nil, cleanup: nil, **options) ⇒ Object

Build a new rule creating a symbol and numbering from the current rule Symbol and number creation is handled by the top-most rule in such a chain

Parameters:

  • expr (Array)
  • options (Hash{Symbol => Object})
  • :kind (Symbol) 


114
115
116
117
118
119
120
121
122
 
# File 'lib/ebnf/rule.rb', line 114

def build(expr, kind: nil, cleanup: nil, **options)
  new_sym, new_id = (@top_rule ||self).send(:make_sym_id)
  Rule.new(new_sym, new_id, expr,
           kind: kind,
           ebnf: @ebnf,
           top_rule: (@top_rule || self),
           cleanup: cleanup,
           **options)
end

#equivalent?(other) ⇒ Boolean

Two rules are equivalent if they have the same #expr

Parameters:

Returns:

  • (Boolean) 


366
367
368
 
# File 'lib/ebnf/rule.rb', line 366

def equivalent?(other)
  expr  == other.expr
end

#first_includes_eps? ⇒ Boolean

Do the firsts of this rule include the empty string?

Returns:

  • (Boolean) 


287
288
289
 
# File 'lib/ebnf/rule.rb', line 287

def first_includes_eps?
  @first && @first.include?(:_eps)
end

#for_sxp ⇒ Array

Return representation for building S-Expressions

Returns:

  • (Array) 


126
127
128
129
130
131
132
133
134
135
 
# File 'lib/ebnf/rule.rb', line 126

def for_sxp
  elements = [kind, sym]
  elements << id if id
  elements << [:start, true] if start
  elements << first.sort_by(&:to_s).unshift(:first) if first
  elements << follow.sort_by(&:to_s).unshift(:follow) if follow
  elements << [:cleanup, cleanup] if cleanup
  elements << expr
  elements
end

#inspect ⇒ Object 



348
349
350
351
352
 
# File 'lib/ebnf/rule.rb', line 348

def inspect
  "#<EBNF::Rule:#{object_id} " +
  {sym: sym, id: id, kind: kind, expr: expr}.inspect +
  ">"
end

#non_terminals(ast) ⇒ Array<Rule>

Return the non-terminals for this rule. For seq, this is the first non-terminals in the seq. For alt, this is every non-terminal ni the alt

Parameters:

  • ast (Array<Rule>) —

    The set of rules, used to turn symbols into rules

Returns:



238
239
240
241
242
243
244
245
246
247
248
 
# File 'lib/ebnf/rule.rb', line 238

def non_terminals(ast)
  @non_terms ||= (alt? ? expr[1..-1] : expr[1,1]).map do |sym|
    case sym
    when Symbol
      r = ast.detect {|r| r.sym == sym}
      r if r && r.rule?
    else
      nil
    end
  end.compact
end

#pass? ⇒ Boolean

Is this a pass?

Returns:

  • (Boolean) 


323
324
325
 
# File 'lib/ebnf/rule.rb', line 323

def pass?
  kind == :pass
end

#rewrite(src_rule, dst_rule) ⇒ Rule

Rewrite the rule substituting src_rule for dst_rule wherever it is used in the production (first level only).

Parameters:

Returns:



375
376
377
378
379
380
381
382
383
 
# File 'lib/ebnf/rule.rb', line 375

def rewrite(src_rule, dst_rule)
  case @expr
  when Array
    @expr = @expr.map {|e| e == src_rule.sym ? dst_rule.sym : e}
  else
    @expr = dst_rule.sym if @expr == src_rule.sym
  end
  self
end

#rule? ⇒ Boolean

Is this a rule?

Returns:

  • (Boolean) 


329
330
331
 
# File 'lib/ebnf/rule.rb', line 329

def rule?
  kind == :rule
end

#seq? ⇒ Boolean

Is this rule of the form (seq …)?

Returns:

  • (Boolean) 


339
340
341
 
# File 'lib/ebnf/rule.rb', line 339

def seq?
  expr.is_a?(Array) && expr.first == :seq
end

#starts_with?(sym) ⇒ Array<Symbol, String>

Does this rule start with a sym? It does if expr is that sym, expr starts with alt and contains that sym, or expr starts with seq and the next element is that sym

Parameters:

  • sym (Symbol, class) —

    Symbol matching any start element, or if it is String, any start element which is a String

Returns:

  • (Array<Symbol, String>) —

    list of symbol (singular), or strings which are start symbol, or nil if there are none



275
276
277
278
279
280
281
282
283
 
# File 'lib/ebnf/rule.rb', line 275

def starts_with?(sym)
  if seq? && sym === (v = expr.fetch(1, nil))
    [v]
  elsif alt? && expr.any? {|e| sym === e}
    expr.select {|e| sym === e}
  else
    nil
  end
end

#terminal? ⇒ Boolean

Is this a terminal?

Returns:

  • (Boolean) 


317
318
319
 
# File 'lib/ebnf/rule.rb', line 317

def terminal?
  kind == :terminal
end

#terminals(ast) ⇒ Array<Rule>

Return the terminals for this rule. For seq, this is the first terminals or strings in the seq. For alt, this is every non-terminal ni the alt

Parameters:

  • ast (Array<Rule>) —

    The set of rules, used to turn symbols into rules

Returns:



255
256
257
258
259
260
261
262
263
264
265
266
267
 
# File 'lib/ebnf/rule.rb', line 255

def terminals(ast)
  @terms ||= (alt? ? expr[1..-1] : expr[1,1]).map do |sym|
    case sym
    when Symbol
      r = ast.detect {|r| r.sym == sym}
      r if r && r.terminal?
    when String
      sym
    else
      nil
    end
  end.compact
end

#to_bnf ⇒ Array<Rule>

Transform EBNF rule to BNF rules:

* Transform (a [n] rule (op1 (op2))) into two rules:
  (a [n] rule (op1 _a_1))
  (_a_1 [n.1] rule (op2))
* Transform (a rule (opt b)) into (a rule (alt _empty b))
* Transform (a rule (star b)) into (a rule (alt _empty (seq b a)))
* Transform (a rule (plus b)) into (a rule (seq b (star b)

Transformation includes information used to re-construct non-transformed AST representation

Returns:



177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
 
# File 'lib/ebnf/rule.rb', line 177

def to_bnf
  return [self] unless rule?
  new_rules = []

  # Look for rules containing recursive definition and rewrite to multiple rules. If `expr` contains elements which are in array form, where the first element of that array is a symbol, create a new rule for it.
  if expr.any? {|e| e.is_a?(Array) && (BNF_OPS + TERM_OPS).include?(e.first)}
    #   * Transform (a [n] rule (op1 (op2))) into two rules:
    #     (a.1 [n.1] rule (op1 a.2))
    #     (a.2 [n.2] rule (op2))
    # duplicate ourselves for rewriting
    this = dup
    new_rules << this

    expr.each_with_index do |e, index|
      next unless e.is_a?(Array) && e.first.is_a?(Symbol)
      new_rule = build(e)
      this.expr[index] = new_rule.sym
      new_rules << new_rule
    end

    # Return new rules after recursively applying #to_bnf
    new_rules = new_rules.map {|r| r.to_bnf}.flatten
  elsif expr.first == :opt
    this = dup
    #   * Transform (a rule (opt b)) into (a rule (alt _empty b))
    this.expr = [:alt, :_empty, expr.last]
    this.cleanup = :opt
    new_rules = this.to_bnf
  elsif expr.first == :star
    #   * Transform (a rule (star b)) into (a rule (alt _empty (seq b a)))
    this = dup
    this.cleanup = :star
    new_rule = this.build([:seq, expr.last, this.sym], cleanup: :merge)
    this.expr = [:alt, :_empty, new_rule.sym]
    new_rules = [this] + new_rule.to_bnf
  elsif expr.first == :plus
    #   * Transform (a rule (plus b)) into (a rule (seq b (star b)
    this = dup
    this.cleanup = :plus
    this.expr = [:seq, expr.last, [:star, expr.last]]
    new_rules = this.to_bnf
  elsif [:alt, :seq].include?(expr.first)
    # Otherwise, no further transformation necessary
    new_rules << self
  elsif [:diff, :hex, :range].include?(expr.first)
    # This rules are fine, the just need to be terminals
    raise "Encountered #{expr.first.inspect}, which is a #{self.kind}, not :terminal" unless self.terminal?
    new_rules << self
  else
    # Some case we didn't think of
    raise "Error trying to transform #{expr.inspect} to BNF"
  end
  
  return new_rules
end

#to_sxp ⇒ String Also known as: to_s

Return SXP representation of this rule

Returns:

  • (String) 


139
140
141
142
 
# File 'lib/ebnf/rule.rb', line 139

def to_sxp
  require 'sxp' unless defined?(SXP)
  for_sxp.to_sxp
end

#to_ttl ⇒ String

Serializes this rule to an Turtle

Returns:

  • (String) 


148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
 
# File 'lib/ebnf/rule.rb', line 148

def to_ttl
  @ebnf.debug("to_ttl") {inspect} if @ebnf
  comment = orig.to_s.strip.
    gsub(/"""/, '\"\"\"').
    gsub("\\", "\\\\").
    sub(/^\"/, '\"').
    sub(/\"$/m, '\"')
  statements = [
    %{:#{id} rdfs:label "#{id}"; rdf:value "#{sym}";},
    %{  rdfs:comment #{comment.inspect};},
  ]
  
  statements += ttl_expr(expr, terminal? ? "re" : "g", 1, false)
  "\n" + statements.join("\n")
end