Class: SDL4R::Reader

Inherits:
Object
  • Object
show all
Defined in:
lib/sdl4r/reader.rb

Overview

Implementation of a pull parser for SDL designed after the model of Nokogiri::XML::Reader.

Constant Summary collapse

TYPE_ELEMENT = 
:ELEMENT
TYPE_END_ELEMENT = 
:END_ELEMENT
@@SKIP_PROC = 
lambda { |reader| false }
@@ON_SELF_CLOSING_TAG_PROG = 
lambda { |reader| reader.on_self_closing_tag }
@@comment_handler_set = 
{
  :INLINE_COMMENT => lambda { |reader| reader.on_simple_comment },
  :ONE_LINE_COMMENT => lambda { |reader| reader.on_simple_comment },
  :MULTILINE_COMMENT_START => lambda { |reader| reader.on_multiline_comment },
}
@@common_tag_set =

Handlers that work the same at the top level or in any normal tag body.

{
  :WHITESPACE => @@SKIP_PROC,
  :EOL => @@SKIP_PROC,
  :SEMICOLON => @@SKIP_PROC,
  :IDENTIFIER => lambda { |reader| reader.on_tag_start },
}
@@handler_sets =

The handlers are object with #call() (like Proc, etc) that should return false if the corresponding token is ignored, true otherwise.

{}
@@value_handlers = 
{
  :NULL => lambda { |s, reader| nil },
  :INTEGER => lambda { |s, reader| reader.parse_integer(s) },
  :FLOAT => lambda { |s, reader| reader.parse_float(s) },
  :BOOLEAN => lambda { |s, reader| (s =~ /\A(?:true|on)\Z/) ? true : false },
  :CHARACTER => lambda { |s, reader| reader.parse_character(s) },
  :INLINE_BACKQUOTE_STRING => lambda { |s, reader| s },
  :INLINE_DOUBLE_QUOTE_STRING => lambda { |s, reader| reader.parse_double_quote_string(s) },
  :MULTILINE_BACKQUOTE_STRING_START =>
    lambda { |s, reader| reader.parse_multiline_backquote_string(s) },
  :MULTILINE_DOUBLE_QUOTE_STRING_START =>
    lambda { |s, reader| reader.parse_multiline_double_quote_string(s) },
  :INLINE_BINARY => lambda { |s, reader| SdlBinary.decode64(s) },
  :MULTILINE_BINARY_START => lambda { |s, reader| reader.parse_multiline_binary(s) },
  :DATE => lambda { |s, reader| reader.parse_date(s) },
  :TIME_OR_TIMESPAN => lambda { |s, reader| reader.parse_time_span(s) },
}

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(io) ⇒ Reader

Returns a new instance of Reader.

Raises:

  • (ArgumentError) 


154
155
156
157
158
159
160
161
162
163
164
165
166
 
# File 'lib/sdl4r/reader.rb', line 154

def initialize(io)
  raise ArgumentError, "io == nil" if io.nil?
  raise ArgumentError, "io is not an IO" unless io.respond_to?(:gets)

  @io = io
  @tokenizer = Tokenizer.new(@io)
  @element = nil
  @element_pool = []
  @depth = 1

  clear_node()
  set_mode(:top)
end

Instance Attribute Details

#depth ⇒ Object (readonly)

Depth of the current SDL node. Depth of top nodes is 1 (0 would be the root that the Reader doesn’t traverse).



152
153
154
 
# File 'lib/sdl4r/reader.rb', line 152

def depth
  @depth
end

#name ⇒ Object (readonly)

Name of the traversed SDL node.



148
149
150
 
# File 'lib/sdl4r/reader.rb', line 148

def name
  @name
end

#node_type ⇒ Object (readonly)

Type of the traversed SDL node (e.g. TYPE_ELEMENT).



142
143
144
 
# File 'lib/sdl4r/reader.rb', line 142

def node_type
  @node_type
end

#prefix ⇒ Object (readonly)

Prefix (namespace) of the traversed SDL node.



145
146
147
 
# File 'lib/sdl4r/reader.rb', line 145

def prefix
  @prefix
end

Class Method Details

.add_values_handler(map, handler) ⇒ Object 



40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
 
# File 'lib/sdl4r/reader.rb', line 40

def self.add_values_handler(map, handler)
  map[:NULL] = handler
  map[:INTEGER] = handler
  map[:FLOAT] = handler
  map[:BOOLEAN] = handler
  map[:CHARACTER] = handler
  map[:INLINE_BACKQUOTE_STRING] = handler
  map[:INLINE_DOUBLE_QUOTE_STRING] = handler
  map[:MULTILINE_BACKQUOTE_STRING_START] = handler
  map[:MULTILINE_DOUBLE_QUOTE_STRING_START] = handler
  map[:INLINE_BINARY] = handler
  map[:MULTILINE_BINARY_START] = handler
  map[:DATE] = handler
  map[:TIME_OR_TIMESPAN] = handler
end

.from_io(io) ⇒ Object 



280
281
282
 
# File 'lib/sdl4r/reader.rb', line 280

def self.from_io(io)
  self.new(io)
end

.from_memory(s) ⇒ Object 



284
285
286
 
# File 'lib/sdl4r/reader.rb', line 284

def self.from_memory(s)
  self.new(StringIO.new(s))
end

Instance Method Details

#attribute(name) ⇒ Object #attribute(prefix, name) ⇒ Object

Returns the value of the specified attribute.

Returns:

  • the value of the specified attribute.



178
179
180
181
182
183
184
185
186
187
188
189
190
191
 
# File 'lib/sdl4r/reader.rb', line 178

def attribute(prefix, name = nil)
  return nil unless @element

  if name
    prefix, name = prefix.to_s, name.to_s
  else
    prefix, name = '', prefix.to_s
  end

  @element.attributes.each do |attr|
    return attr[1] if attr[0][0] == prefix && attr[0][1] == name
  end
  return nil
end

#attribute_at(index) ⇒ Object

Returns the value of the attribute at the specified index.

Returns:

  • the value of the attribute at the specified index.



194
195
196
197
198
199
200
 
# File 'lib/sdl4r/reader.rb', line 194

def attribute_at(index)
  if @element
    @element.attributes[index]
  else
    nil
  end
end

#attribute_count ⇒ Integer

Returns number of attributes in the current element.

Returns:

  • (Integer) —

    number of attributes in the current element



203
204
205
 
# File 'lib/sdl4r/reader.rb', line 203

def attribute_count
  @element ? @element.attributes.size : 0
end

#attributes ⇒ Array

Returns an array of the attributes structured as follows: [ [["ns1", "attr1"], 123], [["", "attr2"], true] ].

Returns:

  • (Array) —

    an array of the attributes structured as follows: [ [["ns1", "attr1"], 123], [["", "attr2"], true] ]



170
171
172
 
# File 'lib/sdl4r/reader.rb', line 170

def attributes
  @element ? @element.attributes.clone : nil
end

#attributes? ⇒ boolean

Returns whether the current element has attributes.

Returns:

  • (boolean) —

    whether the current element has attributes.



208
209
210
 
# File 'lib/sdl4r/reader.rb', line 208

def attributes?
  @element && @element.attributes.size > 0
end

#each {|Reader| ... } ⇒ Object

Enumerates all the parsed nodes and calls the given block.

Examples:

open("sample.sdl") do |io|
  SDL4R::Reader.from_io(io).each do |node]
    puts node.node_type
  end
end

Yields:

  • (Reader) —

    the current node



299
300
301
302
303
 
# File 'lib/sdl4r/reader.rb', line 299

def each(&block)
  while node = self.read
    block.call(node)
  end
end

#each_tag(only_top_tags = false) {|Tag| ... } ⇒ Object

Calls the given block for each encountered Tag. The block is called when the Tag definition is complete.

Parameters:

  • only_top_tags (boolean) (defaults to: false) —

    if true only top Tags are enumerated

Yields:

  • (Tag) —

    called at each Tag



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
 
# File 'lib/sdl4r/reader.rb', line 218

def each_tag(only_top_tags = false)
  stack = []
  tag = nil # Only used during definition (values + attributes)

  while node = read
    case node.node_type

    when TYPE_ELEMENT
      tag = Tag.new @element.prefix, @element.name
      node.attributes.each do |attribute|
        tag.set_attribute(attribute[0][0], attribute[0][1], attribute[1])
      end
      values = node.values
      tag.values = values if values
      stack.last.add_child(tag) unless stack.empty?

      if node.self_closing?
        yield tag if !only_top_tags or @depth <= 1
      else
        stack << tag
      end

      tag = nil # definition ended here

    when TYPE_END_ELEMENT
      tag = stack.pop
      yield tag if !only_top_tags or depth <= 1
    end
  end
end

#new_time(year, month, day, hour, min, sec, msec, timezone_code) ⇒ Object

Creates and returns the object representing a datetime (calls SDL4R#new_time by default). Can be overriden.

def new_time(year, month, day, hour, min, sec, msec, timezone_code)
  Time.utc(year, month, day, hour, min, sec, msec, timezone_code)
end



359
360
361
 
# File 'lib/sdl4r/reader.rb', line 359

def new_time(year, month, day, hour, min, sec, msec, timezone_code)
  SDL4R::new_time(year, month, day, hour, min, sec, msec, timezone_code)
end

#on_anonymous_value ⇒ Object

Should only be called from :top or :tag_body modes.



437
438
439
440
441
442
 
# File 'lib/sdl4r/reader.rb', line 437

def on_anonymous_value # :nodoc:
  set_mode :tag_values
  @element = Element.new '', SDL4R::ANONYMOUS_TAG_NAME

  on_value
end

#on_attribute ⇒ Object 



416
417
418
419
420
421
422
423
424
 
# File 'lib/sdl4r/reader.rb', line 416

def on_attribute # :nodoc:
  read_name
  read_equal
  read_value
  set_mode :tag_attributes
  @element.add_attribute(@prefix, @name, @value)

  false
end

#on_eof ⇒ Object 



370
371
372
373
 
# File 'lib/sdl4r/reader.rb', line 370

def on_eof # :nodoc:
  @node_type = nil
  true
end

#on_multiline_comment ⇒ Object 



376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
 
# File 'lib/sdl4r/reader.rb', line 376

def on_multiline_comment # :nodoc:
#        @node_type = TYPE_COMMENT
  @value = @tokenizer.token

  while @tokenizer.read
    case @tokenizer.token_type
    when :EOL
      @value << ?\n
    when :MULTILINE_COMMENT_PART
      @value << @tokenizer.token
    when :MULTILINE_COMMENT_END
      @value << @tokenizer.token
      break
    else
      raise_unexpected_token
    end
  end

  false
end

#on_self_closing_tag ⇒ Object 



407
408
409
410
411
412
413
 
# File 'lib/sdl4r/reader.rb', line 407

def on_self_closing_tag # :nodoc:
  @node_type = TYPE_ELEMENT
  @prefix = @element.prefix
  @name = @element.name
  @element.self_closing = true
  set_mode(@depth <= 1 ? :top : :tag_body)
end

#on_simple_comment ⇒ Object 



364
365
366
367
 
# File 'lib/sdl4r/reader.rb', line 364

def on_simple_comment # :nodoc:
#        @node_type = TYPE_COMMENT
  false
end

#on_tag_body_end ⇒ Object 



456
457
458
459
460
461
462
463
464
465
466
467
 
# File 'lib/sdl4r/reader.rb', line 456

def on_tag_body_end # :nodoc:
  if @depth <= 1
    raise "unexpected end of tag"
  end

  clear_node
  @node_type = TYPE_END_ELEMENT
  @depth -= 1
  set_mode(@depth <= 1 ? :top : :tag_body)

  true
end

#on_tag_body_start ⇒ Object 



445
446
447
448
449
450
451
452
453
 
# File 'lib/sdl4r/reader.rb', line 445

def on_tag_body_start # :nodoc:
  @node_type = TYPE_ELEMENT
  @prefix = @element.prefix
  @name = @element.name
  @depth += 1
  set_mode :tag_body

  true
end

#on_tag_start ⇒ Object 



398
399
400
401
402
403
404
 
# File 'lib/sdl4r/reader.rb', line 398

def on_tag_start # :nodoc:
  read_name
  set_mode :tag_values
  @element = Element.new @prefix, @name

  false
end

#on_value ⇒ Object 



427
428
429
430
431
432
433
 
# File 'lib/sdl4r/reader.rb', line 427

def on_value # :nodoc:
  read_value
  set_mode :tag_values
  @element.add_value(@value)

  false
end

#parse_character(s) ⇒ Object 



548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
 
# File 'lib/sdl4r/reader.rb', line 548

def parse_character(s)
  case s
  when /\A.\Z/
    s
  when "\\\\"
    "\\"
  when "\\'"
    "'"
  when "\\n"
    "\n"
  when "\\r"
    "\r"
  when "\\t"
    "\t"
  else
    raise "illegal character literal #{s.inspect}"
  end
end

#parse_date(literal) ⇒ Object

Parses the literal into a returned Date object.

Raises an ArgumentError if literal has a bad format.



592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
 
# File 'lib/sdl4r/reader.rb', line 592

def parse_date(literal)
  # here, we're being stricter than strptime() alone as we forbid trailing chars (also faster)
  if literal =~ /\A(-?\d+)\/(\d+)\/(\d+)\Z/
    date_year = $1.to_i
    date_month = $2.to_i
    date_day = $3.to_i

    skip_whitespaces(false)

    # Check whether the next tag is the time part
    if @tokenizer.token_type == :TIME_OR_TIMESPAN
      # Is it a time or timespan?
      day, hour, min, sec, msec, zone =
        parse_time_span_and_time_zone(@tokenizer.token, true, true)

      if day
        @tokenizer.unread
        return Date.civil(date_year, date_month, date_day)
      else
        return new_time(date_year, date_month, date_day, hour, min, sec, msec, zone)
      end

    else
      @tokenizer.unread
      return Date.civil(date_year, date_month, date_day)
    end

  else
    raise ArgumentError, "Malformed Date <#{literal}>"
  end
end

#parse_double_quote_string(s) ⇒ Object 



470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
 
# File 'lib/sdl4r/reader.rb', line 470

def parse_double_quote_string(s)
  return s if s.empty?

  string = ""
  escaped = false

  s.each_char do |c|
    if escaped
      escaped = false

      case c
      when "\\", "\""
        string << c
      when "n"
        string << ?\n
      when "r"
        string << ?\r
      when "t"
        string << ?\t
      else
        @tokenizer.raise_parse_error("Illegal escape character in string literal: '#{c}'.")
      end

    elsif c == "\\"
      escaped = true

    else
      string << c
    end
  end

  @tokenizer.raise_parse_error("orphan backslash") if escaped

  string
end

#parse_float(s) ⇒ Object 



577
578
579
580
581
582
583
584
585
 
# File 'lib/sdl4r/reader.rb', line 577

def parse_float(s)
  if s =~ /\A([^BDF]+)BD\Z/i
    return BigDecimal($1)
  elsif s =~ /\A([^BDF]+)[FD]\Z/i
    return Float($1) rescue @tokenizer.raise_parse_error("not a float '#{$1}'")
  else
    return Float(s) rescue @tokenizer.raise_parse_error("not a float '#{s}'")
  end
end

#parse_integer(s) ⇒ Object 



568
569
570
571
572
573
574
 
# File 'lib/sdl4r/reader.rb', line 568

def parse_integer(s)
  if s =~ /\A([^L]+)L\Z/i
    return Integer($1)
  else
    return Integer(s)
  end
end

#parse_multiline_backquote_string(s) ⇒ Object 



530
531
532
 
# File 'lib/sdl4r/reader.rb', line 530

def parse_multiline_backquote_string(s)
  parse_multiline_string s, :MULTILINE_BACKQUOTE_STRING_PART, :MULTILINE_BACKQUOTE_STRING_END
end

#parse_multiline_binary(s) ⇒ Object 



542
543
544
545
 
# File 'lib/sdl4r/reader.rb', line 542

def parse_multiline_binary(s)
  literal = parse_multiline_string s, :MULTILINE_BINARY_PART, :MULTILINE_BINARY_END
  return SdlBinary.decode64(literal)
end

#parse_multiline_double_quote_string(s) ⇒ Object 



535
536
537
538
539
 
# File 'lib/sdl4r/reader.rb', line 535

def parse_multiline_double_quote_string(s)
  parse_double_quote_string(
    parse_multiline_string(
      s, :MULTILINE_DOUBLE_QUOTE_STRING_PART, :MULTILINE_DOUBLE_QUOTE_STRING_END))
end

#parse_time_span(literal) ⇒ Object

Parses literal (String) into the corresponding SDLTimeSpan, which is then returned.

Raises an ArgumentError if the literal is not a correct timespan literal.



630
631
632
633
634
635
636
637
638
639
 
# File 'lib/sdl4r/reader.rb', line 630

def parse_time_span(literal)
  days, hours, minutes, seconds, milliseconds, zone_code =
    parse_time_span_and_time_zone(literal, true, false)

  if zone_code
    @tokenizer.raise_parse_error("got a time when expecting a timespan: \"#{literal}\"")
  end

  return SDL4R::SdlTimeSpan.new(days || 0, hours, minutes, seconds, milliseconds)
end

#raise_unexpected_token ⇒ Object 



337
338
339
340
341
342
 
# File 'lib/sdl4r/reader.rb', line 337

def raise_unexpected_token
  @tokenizer.raise_parse_error(
    "unexpected token #{@tokenizer.token_type} #{@tokenizer.token.inspect}",
    @tokenizer.token_line_no,
    @tokenizer.token_pos)
end

#read ⇒ Reader

Reads the next node in the SDL structure.

Examples:

open("sample.sdl") do |io|
  reader = SDL4R::Reader.from_io(io)
  while node = reader.read
    puts node.node_type
  end
end

Returns:

  • (Reader) —

    returns a Reader if a new node has been reached or nil if the end of file has been reached.



317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
 
# File 'lib/sdl4r/reader.rb', line 317

def read
  clear_node

  node = nil

  while @tokenizer.read
    handler = @handler_set[@tokenizer.token_type]
    unless handler
      raise_unexpected_token
    end
    if handler.call(self)
      node = self if @node_type # otherwise, we reached the end of the file
      break
    end
  end

  node
end

#self_closing? ⇒ Boolean

Returns:

  • (Boolean) 


249
250
251
 
# File 'lib/sdl4r/reader.rb', line 249

def self_closing?
  @element ? @element.self_closing : false
end

#values ⇒ Object Also known as: value

Returns the values of the current node, nil if there are none.

Returns:

  • the values of the current node, nil if there are none.



261
262
263
264
265
266
267
268
 
# File 'lib/sdl4r/reader.rb', line 261

def values
  if @element
    values = @element.values
    values.empty? ? nil : values.clone
  else
    @value
  end
end

#values? ⇒ Boolean Also known as: value?

Returns:

  • (Boolean) 


271
272
273
274
275
276
277
 
# File 'lib/sdl4r/reader.rb', line 271

def values?
  if @element
    !@element.values.empty?
  else
    !@value.nil?
  end
end