Class: SuperExpressiveRuby
- Inherits:
-
Object
- Object
- SuperExpressiveRuby
- Defined in:
- lib/super-expressive-ruby/super-expressive-ruby.rb
Constant Summary collapse
- NamedGroupRegex =
/^[a-z]+\w*$/i.freeze
- QuantifierTable =
{ oneOrMore: '+', oneOrMoreLazy: '+?', zeroOrMore: '*', zeroOrMoreLazy: '*?', optional: '?', exactly: proc { |times| "{#{times}}" }, atLeast: proc { |times| "{#{times},}" }, between: proc { |times| "{#{times[0]},#{times[1]}}" }, betweenLazy: proc { |times| "{#{times[0]},#{times[1]}}?" } }.freeze
- @@t =
{ root: as_type('root').call, noop: as_type('noop').call, startOfInput: as_type('startOfInput').call, endOfInput: as_type('endOfInput').call, anyChar: as_type('anyChar').call, whitespaceChar: as_type('whitespaceChar').call, nonWhitespaceChar: as_type('nonWhitespaceChar').call, digit: as_type('digit').call, nonDigit: as_type('nonDigit').call, word: as_type('word').call, nonWord: as_type('nonWord').call, wordBoundary: as_type('wordBoundary').call, nonWordBoundary: as_type('nonWordBoundary').call, newline: as_type('newline').call, carriageReturn: as_type('carriageReturn').call, tab: as_type('tab').call, nullByte: as_type('nullByte').call, anyOfChars: as_type('anyOfChars'), anythingButString: as_type('anythingButString'), anythingButChars: as_type('anythingButChars'), anythingButRange: as_type('anythingButRange'), char: as_type('char'), range: as_type('range'), string: as_type('string', { quantifierRequiresGroup: true }), namedBackreference: proc { |name| deferred_type('namedBackreference', { name: name }) }, backreference: proc { |index| deferred_type('backreference', { index: index }) }, capture: deferred_type('capture', { containsChildren: true }), subexpression: as_type('subexpression', { containsChildren: true, quantifierRequiresGroup: true }), namedCapture: proc { |name| deferred_type('namedCapture', { name: name, containsChildren: true }) }, group: deferred_type('group', { containsChildren: true }), anyOf: deferred_type('anyOf', { containsChildren: true }), assertAhead: deferred_type('assertAhead', { containsChildren: true }), assertNotAhead: deferred_type('assertNotAhead', { containsChildren: true }), exactly: proc { |times| deferred_type('exactly', { times: times, containsChild: true }) }, atLeast: proc { |times| deferred_type('atLeast', { times: times, containsChild: true }) }, between: proc { |x, y| deferred_type('between', { times: [x, y], containsChild: true }) }, betweenLazy: proc { |x, y| deferred_type('betweenLazy', { times: [x, y], containsChild: true }) }, zeroOrMore: deferred_type('zeroOrMore', { containsChild: true }), zeroOrMoreLazy: deferred_type('zeroOrMoreLazy', { containsChild: true }), oneOrMore: deferred_type('oneOrMore', { containsChild: true }), oneOrMoreLazy: deferred_type('oneOrMoreLazy', { containsChild: true }), optional: deferred_type('optional', { containsChild: true }) }.freeze
Instance Attribute Summary collapse
-
#state ⇒ Object
Returns the value of attribute state.
Class Method Summary collapse
- .as_type(type, opts = {}) ⇒ Object
- .assert(condition, message) ⇒ Object
- .camelize(snake_case_str) ⇒ Object
- .deferred_type(type, opts = {}) ⇒ Object
- .evaluate(el) ⇒ Object
- .fuse_elements(elements) ⇒ Object
- .is_fusable(element) ⇒ Object
- .partition(a) ⇒ Object
Instance Method Summary collapse
- #allow_multiple_matches ⇒ Object
- #any_char ⇒ Object
- #any_of ⇒ Object
- #any_of_chars(s) ⇒ Object
- #anything_but_chars(chars) ⇒ Object
- #anything_but_range(a, b) ⇒ Object
- #anything_but_string(str) ⇒ Object
- #apply_quantifier(element) ⇒ Object
- #apply_subexpression_defaults(expr) ⇒ Object
- #assert(condition, message) ⇒ Object
- #assert_ahead ⇒ Object
- #assert_not_ahead ⇒ Object
- #at_least(n) ⇒ Object
- #backreference(index) ⇒ Object
- #between(x, y) ⇒ Object
- #between_lazy(x, y) ⇒ Object
- #capture ⇒ Object
- #carriage_return ⇒ Object
- #case_insensitive ⇒ Object
- #char(c) ⇒ Object
- #create_stack_frame(type) ⇒ Object
- #digit ⇒ Object
- #end ⇒ Object
- #end_of_input ⇒ Object
- #escape_special(s) ⇒ Object
- #exactly(n) ⇒ Object
- #frame_creating_element(type_fn) ⇒ Object
- #get_current_element_array ⇒ Object
- #get_current_frame ⇒ Object
- #get_regex_pattern_and_flags ⇒ Object
- #group ⇒ Object
-
#initialize ⇒ SuperExpressiveRuby
constructor
A new instance of SuperExpressiveRuby.
- #line_by_line ⇒ Object
- #match_element(type_fn) ⇒ Object
- #merge_subexpression(el, options, parent, increment_capture_groups) ⇒ Object
- #named_backreference(name) ⇒ Object
- #named_capture(name) ⇒ Object
- #newline ⇒ Object
- #non_digit ⇒ Object
- #non_whitespace_char ⇒ Object
- #non_word ⇒ Object
- #non_word_boundary ⇒ Object
- #null_byte ⇒ Object
- #one_or_more ⇒ Object
- #one_or_more_lazy ⇒ Object
- #optional ⇒ Object
- #quantifier_element(type_fn_name) ⇒ Object
- #range(a, b) ⇒ Object
- #single_line ⇒ Object
- #start_of_input ⇒ Object
- #sticky ⇒ Object
- #string(str) ⇒ Object
- #subexpression(expr, opts = {}) ⇒ Object
- #t ⇒ Object
- #tab ⇒ Object
- #to_regex ⇒ Object
- #to_regex_string ⇒ Object
- #track_named_group(name) ⇒ Object
- #unicode ⇒ Object
- #whitespace_char ⇒ Object
- #word ⇒ Object
- #word_boundary ⇒ Object
- #zero_or_more ⇒ Object
- #zero_or_more_lazy ⇒ Object
Constructor Details
#initialize ⇒ SuperExpressiveRuby
Returns a new instance of SuperExpressiveRuby.
230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 230 def initialize self.state = { hasDefinedStart: false, hasDefinedEnd: false, flags: { g: false, y: false, m: false, i: false, u: false, s: false }, stack: [create_stack_frame(t[:root])], namedGroups: [], totalCaptureGroups: 0 } end |
Instance Attribute Details
#state ⇒ Object
Returns the value of attribute state.
7 8 9 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 7 def state @state end |
Class Method Details
.as_type(type, opts = {}) ⇒ Object
131 132 133 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 131 def as_type(type, opts={}) proc { |value| { type: type, value: value }.merge(opts) } end |
.assert(condition, message) ⇒ Object
140 141 142 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 140 def assert(condition, ) raise StandardError, unless condition end |
.camelize(snake_case_str) ⇒ Object
174 175 176 177 178 179 180 181 182 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 174 def camelize(snake_case_str) snake_case_str.split('_').each_with_object([]).with_index do |(s, acc), idx| acc << if idx.zero? s else s.capitalize end end.join end |
.deferred_type(type, opts = {}) ⇒ Object
135 136 137 138 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 135 def deferred_type(type, opts={}) type_fn = as_type(type, opts) type_fn.call(type_fn) end |
.evaluate(el) ⇒ Object
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 23 def evaluate(el) case el[:type] when 'noop' '' when 'anyChar' '.' when 'whitespaceChar' '\\s' when 'nonWhitespaceChar' '\\S' when 'digit' '\\d' when 'nonDigit' '\\D' when 'word' '\\w' when 'nonWord' '\\W' when 'wordBoundary' '\\b' when 'nonWordBoundary' '\\B' when 'startOfInput' '^' when 'endOfInput' '$' when 'newline' '\\n' when 'carriageReturn' '\\r' when 'tab' '\\t' when 'nullByte' '\\0' when 'string' el[:value] when 'char' el[:value] when 'range' "[#{el[:value][0]}-#{el[:value][1]}]" when 'anythingButRange' "[^#{el[:value][0]}-#{el[:value][1]}]" when 'anyOfChars' "[#{el[:value]}]" when 'anythingButChars' "[^#{el[:value]}]" when 'namedBackreference' "\\k<#{el[:name]}>" when 'backreference' "\\#{el[:index]}" when 'subexpression' el[:value].map { |value| evaluate(value) }.join('') when 'optional', 'zeroOrMore', 'zeroOrMoreLazy', 'oneOrMore', 'oneOrMoreLazy' inner = evaluate(el[:value]) with_group = if el[:value][:quantifierRequiresGroup] "(?:#{inner})" else inner end symbol = QuantifierTable[el[:type].to_sym] "#{with_group}#{symbol}" when 'betweenLazy', 'between', 'atLeast', 'exactly' inner = evaluate(el[:value]) withGroup = if el[:value][:quantifierRequiresGroup] "(?:#{inner})" else inner end "#{withGroup}#{QuantifierTable[el[:type].to_sym].call(el[:times])}" when 'anythingButString' chars = el[:value].split('').map { |c| "[^#{c}]" }.join('') "(?:#{chars})" when 'assertAhead' evaluated = el[:value].map { |v| evaluate(v) }.join('') "(?=#{evaluated})" when 'assertNotAhead' evaluated = el[:value].map { |v| evaluate(v) }.join('') "(?!#{evaluated})" when 'anyOf' fused, rest = fuse_elements(el[:value]) return "[#{fused}]" unless rest.length evaluatedRest = rest.map { |v| evaluate(v) } separator = evaluatedRest.length > 0 && fused.length > 0 ? '|' : '' "(?:#{evaluatedRest.join('|')}#{separator}#{fused ? "[#{fused}]" : ''})" when 'capture' evaluated = el[:value].map { |v| evaluate(v) } "(#{evaluated.join('')})" when 'namedCapture' evaluated = el[:value].map { |v| evaluate(v) } "(?<#{el[:name]}>#{evaluated.join('')})" when 'group' evaluated = el[:value].map { |v| evaluate(v) } "(?:#{evaluated.join('')})" else raise "Can't process unsupported element type: #{el[:type]}" end end |
.fuse_elements(elements) ⇒ Object
162 163 164 165 166 167 168 169 170 171 172 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 162 def fuse_elements(elements) fusables, rest = partition(elements) fused = fusables.map do |el| if %w[char anyOfChars].include?(el[:type]) el[:value] else "#{el[:value][0]}-#{el[:value][1]}" end end.join('') [fused, rest] end |
.is_fusable(element) ⇒ Object
156 157 158 159 160 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 156 def is_fusable(element) element[:type] == 'range' || element[:type] == 'char' || element[:type] == 'anyOfChars' end |
.partition(a) ⇒ Object
144 145 146 147 148 149 150 151 152 153 154 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 144 def partition(a) r = a.each_with_object([[], []]) do |cur, acc| if is_fusable(cur) acc[0].push(cur) else acc[1].push(cur) end acc end [r[0], r[1]] end |
Instance Method Details
#allow_multiple_matches ⇒ Object
260 261 262 263 264 265 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 260 def allow_multiple_matches # warn("Warning: Ruby does not have a allow multiple matches option. use String#gsub or String#scan") n = clone n.state[:flags][:g] = true n end |
#any_char ⇒ Object
305 306 307 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 305 def any_char match_element(t[:anyChar]) end |
#any_of ⇒ Object
376 377 378 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 376 def any_of frame_creating_element(t[:anyOf]) end |
#any_of_chars(s) ⇒ Object
524 525 526 527 528 529 530 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 524 def any_of_chars(s) n = clone element_value = t[:anyOfChars].call(escape_special(s)) current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#anything_but_chars(chars) ⇒ Object
553 554 555 556 557 558 559 560 561 562 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 553 def anything_but_chars(chars) assert(chars.is_a?(String), "chars must be a string (got #{chars})") assert(chars.length > 0, 'chars must have at least one character') n = clone element_value = t[:anythingButChars].call(escape_special(chars)) current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#anything_but_range(a, b) ⇒ Object
564 565 566 567 568 569 570 571 572 573 574 575 576 577 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 564 def anything_but_range(a, b) str_a = a.to_s str_b = b.to_s assert(str_a.length === 1, "a must be a single character or number (got #{str_a})") assert(str_b.length === 1, "b must be a single character or number (got #{str_b})") assert(str_a[0].ord < str_b[0].ord, "a must have a smaller character value than b (a = #{str_a[0].ord}, b = #{str_b[0].ord})") n = clone element_value = t[:anythingButRange].call([a, b]) current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#anything_but_string(str) ⇒ Object
542 543 544 545 546 547 548 549 550 551 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 542 def anything_but_string(str) assert(str.is_a?(String), "str must be a string (got #{str})") assert(str.length > 0, 'str must have least one character') n = clone element_value - t[:anythingButString].call(escape_special(str)) current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#apply_quantifier(element) ⇒ Object
783 784 785 786 787 788 789 790 791 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 783 def apply_quantifier(element) current_frame = get_current_frame if current_frame[:quantifier] wrapped = current_frame[:quantifier][:value].call(element) current_frame[:quantifier] = nil return wrapped end element end |
#apply_subexpression_defaults(expr) ⇒ Object
700 701 702 703 704 705 706 707 708 709 710 711 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 700 def apply_subexpression_defaults(expr) out = {}.merge(expr) out[:namespace] = out.has_key?(:namespace) ? out[:namespace] : '' out[:ignoreFlags] = out.has_key?(:ignoreFlags) ? out[:ignoreFlags] : true out[:ignoreStartAndEnd] = out.has_key?(:ignoreStartAndEnd) ? out[:ignoreStartAndEnd] : true assert(out[:namespace].is_a?(String), 'namespace must be a string') assert(out[:ignoreFlags].is_a?(TrueClass) || out[:ignoreFlags].is_a?(FalseClass), 'ignoreFlags must be a boolean') assert(out[:ignoreStartAndEnd].is_a?(TrueClass) || out[:ignoreStartAndEnd].is_a?(FalseClass), 'ignoreStartAndEnd must be a boolean') out end |
#assert(condition, message) ⇒ Object
801 802 803 804 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 801 def assert(condition, ) self.class.assert(condition, ) raise StandardError, unless condition end |
#assert_ahead ⇒ Object
384 385 386 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 384 def assert_ahead frame_creating_element(t[:assertAhead]) end |
#assert_not_ahead ⇒ Object
388 389 390 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 388 def assert_not_ahead frame_creating_element(t[:assertNotAhead]) end |
#at_least(n) ⇒ Object
463 464 465 466 467 468 469 470 471 472 473 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 463 def at_least(n) assert(n.is_a?(Integer) && n > 0, "n must be a positive integer (got #{n})") nxt = clone current_frame = nxt.get_current_frame if current_frame[:quantifier] raise StandardError, "cannot quantify regular expression with 'atLeast' because it's already being quantified with '#{currentFrame.quantifier.type}'" end current_frame[:quantifier] = t[:atLeast].call(n) nxt end |
#backreference(index) ⇒ Object
362 363 364 365 366 367 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 362 def backreference(index) assert(index.is_a?(Integer), 'index must be a number') assert(index > 0 && index <= state[:totalCaptureGroups], "invalid index #{index}. There are #{state[:totalCaptureGroups]} capture groups on this SuperExpression") match_element(t[:backreference].call(index)) end |
#between(x, y) ⇒ Object
475 476 477 478 479 480 481 482 483 484 485 486 487 488 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 475 def between(x, y) assert(x.is_a?(Integer) && x >= 0, "x must be an integer (got #{x})") assert(y.is_a?(Integer) && y > 0, "y must be an integer greater than 0 (got #{y})") assert(x < y, "x must be less than y (x = #{x}, y = #{y})") nxt = clone current_frame = nxt.get_current_frame if current_frame[:quantifier] raise StandardError, "cannot quantify regular expression with 'between' because it's already being quantified with '#{currentFrame.quantifier.type}'" end current_frame[:quantifier] = t[:between].call(x, y) nxt end |
#between_lazy(x, y) ⇒ Object
490 491 492 493 494 495 496 497 498 499 500 501 502 503 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 490 def between_lazy(x, y) assert(x.is_a?(Integer) && x >= 0, "x must be an integer (got #{x})") assert(y.is_a?(Integer) && y > 0, "y must be an integer greater than 0 (got #{y})") assert(x < y, "x must be less than y (x = #{x}, y = #{y})") n = clone current_frame = n.get_current_frame if current_frame[:quantifier] raise StandardError, "cannot quantify regular expression with 'betweenLazy' because it's already being quantified with '#{current_frame[:quantifier][:type]}'" end current_frame[:quantifier] = t[:betweenLazy].call(x, y) n end |
#capture ⇒ Object
392 393 394 395 396 397 398 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 392 def capture n = clone new_frame = create_stack_frame(t[:capture]) n.state[:stack].push(new_frame) n.state[:totalCaptureGroups] += 1 n end |
#carriage_return ⇒ Object
345 346 347 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 345 def carriage_return match_element(t[:carriageReturn]) end |
#case_insensitive ⇒ Object
274 275 276 277 278 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 274 def case_insensitive n = clone n.state[:flags][:i] = true n end |
#char(c) ⇒ Object
596 597 598 599 600 601 602 603 604 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 596 def char(c) assert(c.is_a?(String), "c must be a string (got #{c})") assert(c.length == 1, "char() can only be called with a single character (got #{c})") n = clone current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(t[:char].call(escape_special(c)))) n end |
#create_stack_frame(type) ⇒ Object
256 257 258 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 256 def create_stack_frame(type) { type: type, quantifier: nil, elements: [] } end |
#digit ⇒ Object
317 318 319 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 317 def digit match_element(t[:digit]) end |
#end ⇒ Object
532 533 534 535 536 537 538 539 540 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 532 def end assert(state[:stack].length > 1, 'Cannot call end while building the root expression.') n = clone old_frame = n.state[:stack].pop current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(old_frame[:type][:value].call(old_frame[:elements]))) n end |
#end_of_input ⇒ Object
515 516 517 518 519 520 521 522 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 515 def end_of_input assert(!state[:hasDefinedEnd], 'This regex already has a defined end of input') n = clone n.state[:hasDefinedEnd] = true n.get_current_element_array.push(t[:endOfInput]) n end |
#escape_special(s) ⇒ Object
252 253 254 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 252 def escape_special(s) Regexp.escape(s) end |
#exactly(n) ⇒ Object
450 451 452 453 454 455 456 457 458 459 460 461 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 450 def exactly(n) assert(n.is_a?(Integer) && n > 0, "n must be a positive integer (got #{n})") nxt = clone current_frame = nxt.get_current_frame if current_frame[:quantifier] raise StandardError, "cannot quantify regular expression with 'exactly' because it's already being quantified with '#{current_frame[:quantifier][:type]}'" end current_frame[:quantifier] = t[:exactly].call(n) nxt end |
#frame_creating_element(type_fn) ⇒ Object
369 370 371 372 373 374 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 369 def frame_creating_element(type_fn) n = clone new_frame = create_stack_frame(type_fn) n.state[:stack].push(new_frame) n end |
#get_current_element_array ⇒ Object
797 798 799 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 797 def get_current_element_array get_current_frame[:elements] end |
#get_current_frame ⇒ Object
793 794 795 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 793 def get_current_frame state[:stack][state[:stack].length - 1] end |
#get_regex_pattern_and_flags ⇒ Object
760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 760 def get_regex_pattern_and_flags assert state[:stack].length === 1, "Cannot compute the value of a not yet fully specified regex object. \n(Try adding a .end() call to match the '#{get_current_frame[:type][:type]}')\n" pattern = get_current_element_array.map { |el| self.class.evaluate(el) }.join('') flag = nil state[:flags].map do |name, is_on| if is_on flag = 0 if !flag case name when :s flag = flag | Regexp::MULTILINE when :i flag = flag | Regexp::IGNORECASE when :x flag = flag | Regexp::EXTENDED end end end pat = (pattern == '' ? '(?:)' : pattern) [pat, flag] end |
#group ⇒ Object
380 381 382 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 380 def group frame_creating_element(t[:group]) end |
#line_by_line ⇒ Object
267 268 269 270 271 272 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 267 def line_by_line # warn("Warning: Ruby does not have a line by line option. use \A or \z as an alternative") n = clone n.state[:flags][:m] = true n end |
#match_element(type_fn) ⇒ Object
299 300 301 302 303 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 299 def match_element(type_fn) n = clone n.get_current_element_array.push(n.apply_quantifier(type_fn)) n end |
#merge_subexpression(el, options, parent, increment_capture_groups) ⇒ Object
622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 622 def merge_subexpression(el, , parent, increment_capture_groups) next_el = el.clone next_el[:index] += parent.state[:totalCaptureGroups] if next_el[:type] == 'backreference' increment_capture_groups.call if next_el[:type] == 'capture' if next_el[:type] === 'namedCapture' group_name = if [:namespace] "#{options[:namespace]}#{next_el[:name]}" else next_el[:name] end parent.track_named_group(group_name) next_el[:name] = group_name end if next_el[:type] == 'namedBackreference' next_el[:name] = if [:namespace] "#{options[:namespace]}#{next_el[:name]}" else next_el[:name] end end if next_el[:containsChild] next_el[:value] = merge_subexpression( next_el[:value], , parent, increment_capture_groups ) elsif next_el[:containsChildren] next_el[:value] = next_el[:value].map do |e| merge_subexpression( e, , parent, increment_capture_groups ) end end if next_el[:type] == 'startOfInput' return @@t[:noop] if [:ignoreStartAndEnd] assert( !parent.state[:hasDefinedStart], 'The parent regex already has a defined start of input. ' + 'You can ignore a subexpressions startOfInput/endOfInput markers with the ignoreStartAndEnd option' ) assert( !parent.state[:hasDefinedEnd], 'The parent regex already has a defined end of input. ' + 'You can ignore a subexpressions startOfInput/endOfInput markers with the ignoreStartAndEnd option' ) parent.state[:hasDefinedStart] = true end if next_el[:type] == 'endOfInput' return @@t[:noop] if [:ignoreStartAndEnd] assert( !parent.state[:hasDefinedEnd], 'The parent regex already has a defined start of input. ' + 'You can ignore a subexpressions startOfInput/endOfInput markers with the ignoreStartAndEnd option' ) parent.state[:hasDefinedEnd] = true end next_el end |
#named_backreference(name) ⇒ Object
357 358 359 360 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 357 def named_backreference(name) assert(state[:namedGroups].include?(name), "no capture group called '#{name}' exists (create one with .namedCapture())") match_element(t[:namedBackreference].call(name)) end |
#named_capture(name) ⇒ Object
409 410 411 412 413 414 415 416 417 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 409 def named_capture(name) n = clone new_frame = create_stack_frame(t[:namedCapture].call(name)) n.track_named_group(name) n.state[:stack].push(new_frame) n.state[:totalCaptureGroups] += 1 n end |
#newline ⇒ Object
341 342 343 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 341 def newline match_element(t[:newline]) end |
#non_digit ⇒ Object
321 322 323 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 321 def non_digit match_element(t[:nonDigit]) end |
#non_whitespace_char ⇒ Object
313 314 315 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 313 def non_whitespace_char match_element(t[:nonWhitespaceChar]) end |
#non_word ⇒ Object
329 330 331 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 329 def non_word match_element(t[:nonWord]) end |
#non_word_boundary ⇒ Object
337 338 339 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 337 def non_word_boundary match_element(t[:nonWordBoundary]) end |
#null_byte ⇒ Object
353 354 355 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 353 def null_byte match_element(t[:nullByte]) end |
#one_or_more ⇒ Object
442 443 444 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 442 def one_or_more quantifier_element('oneOrMore') end |
#one_or_more_lazy ⇒ Object
446 447 448 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 446 def one_or_more_lazy quantifier_element('oneOrMoreLazy') end |
#optional ⇒ Object
430 431 432 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 430 def optional quantifier_element('optional') end |
#quantifier_element(type_fn_name) ⇒ Object
419 420 421 422 423 424 425 426 427 428 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 419 def quantifier_element(type_fn_name) n = clone current_frame = n.get_current_frame if current_frame[:quantifier] raise StandardError, "cannot quantify regular expression with '#{type_fn_name}' because it's already being quantified with '#{current_frame[:quantifier][:type]}'" end current_frame[:quantifier] = t[type_fn_name.to_sym] n end |
#range(a, b) ⇒ Object
606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 606 def range(a, b) str_a = a.to_s str_b = b.to_s assert(str_a.length == 1, "a must be a single character or number (got #{str_a})") assert(str_b.length == 1, "b must be a single character or number (got #{str_b})") assert(str_a[0].ord < str_b[0].ord, "a must have a smaller character value than b (a = #{str_a[0].ord}, b = #{str_b[0].ord})") n = clone element_value = t[:range].call([str_a, str_b]) current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#single_line ⇒ Object
293 294 295 296 297 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 293 def single_line n = clone n.state[:flags][:s] = true n end |
#start_of_input ⇒ Object
505 506 507 508 509 510 511 512 513 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 505 def start_of_input assert(!state[:hasDefinedStart], 'This regex already has a defined start of input') assert(!state[:hasDefinedEnd], 'Cannot define the start of input after the end of input') n = clone n.state[:hasDefinedStart] = true n.get_current_element_array.push(t[:startOfInput]) n end |
#sticky ⇒ Object
280 281 282 283 284 285 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 280 def sticky # warn("Warning: Ruby does not have a sticky option") n = clone n.state[:flags][:y] = true n end |
#string(str) ⇒ Object
579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 579 def string(str) assert('' != str, 'str cannot be an empty string') n = clone element_value = if str.length > 1 t[:string].call(escape_special(str)) else t[:char].call(str) end current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(element_value)) n end |
#subexpression(expr, opts = {}) ⇒ Object
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 713 def subexpression(expr, opts = {}) assert(expr.is_a?(SuperExpressiveRuby), 'expr must be a SuperExpressive instance') assert( expr.state[:stack].length === 1, 'Cannot call subexpression with a not yet fully specified regex object.' + "\n(Try adding a .end() call to match the '#{expr.get_current_frame[:type][:type]}' on the subexpression)\n" ) = apply_subexpression_defaults(opts) expr_n = expr.clone expr_n.state = expr.state.deep_dup n = clone additional_capture_groups = 0 expr_frame = expr_n.get_current_frame closure = proc { additional_capture_groups += 1 } expr_frame[:elements] = expr_frame[:elements].map do |e| merge_subexpression(e, , n, closure) end n.state[:totalCaptureGroups] += additional_capture_groups unless [:ignoreFlags] expr_n.state[:flags].to_a.each do |e| flag_name = e[0] enabled = e[1] n.state[:flags][flag_name] = enabled || n.state[:flags][flag_name] end end current_frame = n.get_current_frame current_frame[:elements].push(n.apply_quantifier(t[:subexpression].call(expr_frame[:elements]))) n end |
#t ⇒ Object
248 249 250 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 248 def t @@t end |
#tab ⇒ Object
349 350 351 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 349 def tab match_element(t[:tab]) end |
#to_regex ⇒ Object
755 756 757 758 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 755 def to_regex pattern, flags = get_regex_pattern_and_flags Regexp.new(pattern, flags) end |
#to_regex_string ⇒ Object
750 751 752 753 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 750 def to_regex_string pattern, flags = get_regex_pattern_and_flags Regexp.new(pattern, flags).to_s end |
#track_named_group(name) ⇒ Object
400 401 402 403 404 405 406 407 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 400 def track_named_group(name) assert(name.is_a?(String), "name must be a string (got #{name})") assert(name.length > 0, 'name must be at least one character') assert(!state[:namedGroups].include?(name), "cannot use #{name} again for a capture group") assert(name.scan(NamedGroupRegex).any?, "name '#{name}' is not valid (only letters, numbers, and underscores)") state[:namedGroups].push name end |
#unicode ⇒ Object
287 288 289 290 291 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 287 def unicode n = clone n.state[:flags][:u] = true n end |
#whitespace_char ⇒ Object
309 310 311 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 309 def whitespace_char match_element(t[:whitespaceChar]) end |
#word ⇒ Object
325 326 327 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 325 def word match_element(t[:word]) end |
#word_boundary ⇒ Object
333 334 335 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 333 def word_boundary match_element(t[:wordBoundary]) end |
#zero_or_more ⇒ Object
434 435 436 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 434 def zero_or_more quantifier_element('zeroOrMore') end |
#zero_or_more_lazy ⇒ Object
438 439 440 |
# File 'lib/super-expressive-ruby/super-expressive-ruby.rb', line 438 def zero_or_more_lazy quantifier_element('zeroOrMoreLazy') end |