Class: EBNF::Writer

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

Constant Summary collapse

LINE_LENGTH = 
80
LINE_LENGTH_HTML = 
200
UNICODE_ESCAPE_NAMES =

UNICODE escape names From en.wikipedia.org/wiki/List_of_Unicode_characters

{
  0x00 => 'null',
  0x01 => 'start of heading',
  0x02 => 'start of text',
  0x03 => 'end of text',
  0x04 => 'end of transmission',
  0x05 => 'enquiry',
  0x06 => 'acknowledge',
  0x07 => 'bell',
  0x08 => 'backspace',
  0x09 => 'horizontal tab',
  0x0A => 'new line',
  0x0B => 'vertical tab',
  0x0C => 'form feed',
  0x0D => 'carriage return',
  0x0E => 'shift out',
  0x0F => 'shift in',
  0x10 => 'data link escape',
  0x11 => 'device control 1',
  0x12 => 'device control 2',
  0x13 => 'device control 3',
  0x14 => 'device control 4',
  0x15 => 'negative acknowledge',
  0x16 => 'synchronous idle',
  0x17 => 'end of trans. block',
  0x18 => 'cancel',
  0x19 => 'end of medium',
  0x1A => 'substitute',
  0x1B => 'escape',
  0x1C => 'file separator',
  0x1D => 'group separator',
  0x1E => 'record separator',
  0x1F => 'unit separator',
  0x20 => 'space',
  0x22 => 'dquote',
  0x27 => 'apos',
  0x2F => 'slash',
  0x5C => 'backslash',
  0x60 => 'grave',
  0x7F => 'delete',
  0x80 => 'padding character',
  0x81 => 'high octet preset',
  0x82 => 'break permitted here',
  0x83 => 'no break here',
  0x84 => 'index',
  0x85 => 'next line',
  0x86 => 'start of selected area',
  0x87 => 'end of selected area',
  0x88 => 'character tabulation set',
  0x89 => 'character tabulation with justification',
  0x8A => 'line tabulation set',
  0x8B => 'partial line forward',
  0x8C => 'partial line backward',
  0x8D => 'reverse line feed',
  0x8E => 'single-shift two',
  0x8F => 'single-shift three',
  0x90 => 'device control string',
  0x91 => 'private use 1',
  0x92 => 'private use 2',
  0x93 => 'set transmit state',
  0x94 => 'cancel character',
  0x95 => 'message waiting',
  0x96 => 'start of protected area',
  0x97 => 'end of protected area',
  0x98 => 'start of string',
  0x99 => 'single graphic character introducer',
  0x9A => 'single character intro introducer',
  0x9B => 'control sequence introducer',
  0x9C => 'string terminator',
  0x9D => 'operating system command',
  0x9E => 'private message',
  0x9F => 'application program command',
}

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(rules, out: $stdout, html: false, format: :ebnf, validate: false, **options) ⇒ Writer

Returns a new instance of Writer.

Parameters:

  • rules (Array<Rule>)
  • format (:abnf, :ebnf, :isoebnf) (defaults to: :ebnf) —

    (:ebnf)

  • html (Boolean) (defaults to: false) —

    (false) generate HTML output

  • validate (Boolean) (defaults to: false) —

    (false) validate generated HTML.

  • options (Hash{Symbol => Object})
  • out (#write) (defaults to: $stdout) —

    ($stdout)



145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
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
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
 
# File 'lib/ebnf/writer.rb', line 145

def initialize(rules, out: $stdout, html: false, format: :ebnf, validate: false, **options)
  @options = options.merge(html: html)
  return if rules.empty?

  # Determine max LHS length
  format_meth = "format_#{format}".to_sym
  max_id = rules.max_by {|r| r.id.to_s.length}.id.to_s.length
  max_sym = rules.max_by {|r| r.sym.to_s.length}.sym.to_s.length
  lhs_length = max_sym + 1
  lhs_fmt = case format
  when :abnf    then "%<sym>-#{max_sym}s = "
  when :ebnf    then "%<sym>-#{max_sym}s ::= "
  when :isoebnf then "%<sym>-#{max_sym}s = "
  end
  if format == :ebnf && max_id > 0
    lhs_fmt = "%<id>-#{max_id+2}s " + lhs_fmt
    lhs_length += max_id + 3
  end
  rhs_length = (html ? LINE_LENGTH_HTML : LINE_LENGTH) - lhs_length

  if html
    # Output as formatted HTML
    begin
      require 'erubis'
      require 'htmlentities'
      @coder = HTMLEntities.new
      eruby = Erubis::Eruby.new(ERB_DESC)
      formatted_rules = rules.map do |rule|
        if rule.kind == :terminals || rule.kind == :pass
          OpenStruct.new(id: ("@#{rule.kind}"),
                         class: :declaration,
                         sym: rule.kind,
                         assign: nil,
                         formatted: (
                           rule.kind == :terminals ?
                             "<strong># Productions for terminals</strong>" :
                             self.send(format_meth, rule.expr)))
        else
          formatted_expr = self.send(format_meth, rule.expr)
          # Measure text without markup
          formatted_expr_text = formatted_expr.gsub(%r{</?\w+[^>]*>}, '')
          if formatted_expr_text.length > rhs_length && (format != :abnf || rule.alt?)
            lines = []
            # Can only reasonably split apart alts
            self.send(format_meth, rule.expr, sep: "--rule-extensions--").
            split(/\s*--rule-extensions--\s*/).each_with_index do |formatted, ndx|
              assign = case format
              when :ebnf
                formatted.sub!(%r{\s*<code[^>]*>\|</code>\s*}, '')
                (ndx > 0 ? (rule.alt? ? '<code class="grammar-alt">|</code>' : '') : '::=')
              when :abnf
                formatted.sub!(%r{\s*<code[^>]>/</code>\s*}, '')
                (ndx > 0 ? '<code class="grammar-alt">=/</code>' : '=')
              else
                formatted.sub!(%r{\s*<code[^>]>\|</code>\s*}, '')
                (ndx > 0 ? (rule.alt? ? '<code class="grammar-alt">|</code>' : '') : '=')
              end
              lines << OpenStruct.new(id: ((ndx == 0 ? "[#{rule.id}]" : "") if rule.id),
                                      sym: (rule.sym if ndx == 0 || format == :abnf),
                                      class: :production,
                                      assign: assign,
                                      formatted: formatted)
            end
            if format == :isoebnf
              lines << OpenStruct.new(assign: ';')
            end
            lines
          else
            OpenStruct.new(id: ("[#{rule.id}]" if rule.id),
                           class: :production,
                           sym: rule.sym,
                           assign: (format == :ebnf ? '::=' : '='),
                           formatted: (formatted_expr + (format == :isoebnf ? ' ;' : '')))
          end
        end
      end.flatten

      html_result = eruby.evaluate(format: format, rules: formatted_rules)

      if validate
        begin
          require 'nokogiri'
          # Validate the output HTML
          doc = ::Nokogiri::HTML5("<!DOCTYPE html>" + html_result, max_errors: 10)
          raise EncodingError, "Errors found in generated HTML:\n  " +
            doc.errors.map(&:to_s).join("\n  ") unless doc.errors.empty?
        rescue LoadError, NoMethodError
          # Skip
        end
      end

      out.write html_result
      return
    rescue LoadError
      $stderr.puts "Generating HTML requires erubis and htmlentities gems to be loaded"
    end
  end

  # Format each rule, considering the available rhs size
  rules.each do |rule|
    buffer = if rule.pass?
      "\n%-#{lhs_length-2}s      " % '@pass'
    elsif rule.kind == :terminals
      "\n%-#{lhs_length-2}s" % '@terminals'
    else
      lhs_fmt % {id: "[#{rule.id}]", sym: rule.sym}
    end
    formatted_expr = self.send(format_meth, rule.expr)
    if formatted_expr.length > rhs_length && (format != :abnf || rule.alt?)
      if format == :abnf
        # No whitespace, use =/
        self.send(format_meth, rule.expr, sep: "--rule-extensions--").
        split(/\s*--rule-extensions--\s*/).each_with_index do |formatted, ndx|
          if ndx > 0
            buffer << "\n" + lhs_fmt.sub('= ', '=/') % {id: "[#{rule.id}]", sym: rule.sym}
          end
          buffer << formatted.sub(/\s*\/\s*/, '')
        end
      else
        # Space out past "= "
        buffer << self.send(format_meth, rule.expr, sep: ("\n" + " " * (lhs_length + (rule.alt? ? 2 : 4) - (format == :ebnf ? 0 : 2))))
        buffer << ("\n" + " " * (lhs_length) + ';') if format == :isoebnf
      end
    else
      buffer << formatted_expr + (format == :isoebnf ? ' ;'  : '')
    end
    buffer << "\n\n" if [:terminals, :pass].include?(rule.kind)
    out.puts(buffer)
  end
end

Class Method Details

.html(*rules, format: :ebnf, validate: false) ⇒ Object

Write formatted rules to an IO like object as HTML

Parameters:

  • rules (Array<Rule>)
  • format (:abnf, :ebnf, :isoebnf) (defaults to: :ebnf) —

    (:ebnf)

  • validate (Boolean) (defaults to: false) —

    (false) validate generated HTML.

Returns:

  • (Object) 


131
132
133
134
135
136
 
# File 'lib/ebnf/writer.rb', line 131

def self.html(*rules, format: :ebnf, validate: false)
  require 'stringio' unless defined?(StringIO)
  buf = StringIO.new
  Writer.new(rules, out: buf, html: true, format: format, validate: validate)
  buf.string
end

.print(*rules, format: :ebnf) ⇒ Object

Format rules to $stdout

Parameters:

  • rules (Array<Rule>)
  • format (:abnf, :ebnf, :isoebnf) (defaults to: :ebnf) —

    (:ebnf)

Returns:

  • (Object) 


109
110
111
 
# File 'lib/ebnf/writer.rb', line 109

def self.print(*rules, format: :ebnf)
  write($stdout, *rules, format: format)
end

.string(*rules, format: :ebnf) ⇒ Object

Format rules to a String

Parameters:

  • rules (Array<Rule>)
  • format (:abnf, :ebnf, :isoebnf) (defaults to: :ebnf) —

    (:ebnf)

Returns:

  • (Object) 


96
97
98
99
100
101
 
# File 'lib/ebnf/writer.rb', line 96

def self.string(*rules, format: :ebnf)
  require 'stringio' unless defined?(StringIO)
  buf = StringIO.new
  write(buf, *rules, format: format)
  buf.string
end

.write(out, *rules, format: :ebnf) ⇒ Object

Write formatted rules to an IO like object

Parameters:

  • out (Object)
  • rules (Array<Rule>)
  • format (:abnf, :ebnf, :isoebnf) (defaults to: :ebnf) —

    (:ebnf)

Returns:

  • (Object) 


120
121
122
 
# File 'lib/ebnf/writer.rb', line 120

def self.write(out, *rules, format: :ebnf)
  Writer.new(rules, out: out, format: format)
end