plain_text 0.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,973 @@
1
+ # -*- coding: utf-8 -*-
2
+
3
+ require "plain_text/util"
4
+
5
+ module PlainText
6
+
7
+ # Class to represent a Chapter-like entity like an Array
8
+ #
9
+ # An instance of this class contains always an even number of elements,
10
+ # either another {Part} instance or {Paragraph}-type String-like instance,
11
+ # followed by a {Boundary}-type String-like instance. The first element is
12
+ # always a former and the last element is always a latter.
13
+ #
14
+ # Essentially, the instance of this class holds the order information between sub-{Part}-s (< Array)
15
+ # and/or {Paragraph}-s (< String) and {Boundary}-s (< String).
16
+ #
17
+ # An example instance looks like this:
18
+ #
19
+ # Part (
20
+ # (0) Paragraph::Empty,
21
+ # (1) Boundary::General,
22
+ # (2) Part::ArticleHeader(
23
+ # (0) Paragraph::Title,
24
+ # (1) Boundary::Empty
25
+ # ),
26
+ # (3) Boundary::TitleMain,
27
+ # (4) Part::ArticleMain(
28
+ # (0) Part::ArticleSection(
29
+ # (0) Paragraph::Title,
30
+ # (1) Boundary::General,
31
+ # (2) Paragraph::General,
32
+ # (3) Boundary::General,
33
+ # (4) Part::ArticleSubSection(...),
34
+ # (5) Boundary::General,
35
+ # (6) Paragraph::General,
36
+ # (7) Boundary::Empty
37
+ # ),
38
+ # (1) Boundary::General,
39
+ # (2) Paragraph::General,
40
+ # (3) Boundary::Empty
41
+ # ),
42
+ # (5) Boundary::General
43
+ # )
44
+ #
45
+ # A Section (Part) always has an even number of elements: pairs of ({Part}|{Paragraph}) and {Boundary} in this order.
46
+ #
47
+ # Note some standard destructive Array operations, most notably +#delete+, +#delete_if+, +#reject!+,
48
+ # +#select!+, +#filter!+, +#keep_if+, +#flatten!+, +#uniq!+ may alter the content in a way
49
+ # it breaks the self-inconsistency of the object.
50
+ # Use it at your own risk, if you wish (or don't).
51
+ #
52
+ # An instance of this class is always *non-equal* to that of the standard Array class.
53
+ # To compare it at the Array level, convert a {Part} class instance into Array with #to_a first and compare them.
54
+ #
55
+ # For CRUD of elements (contents) of an instance, the following methods are most basic:
56
+ #
57
+ # * Create:
58
+ # * Insert/Append: {#insert} to insert. If the specified index is #size}, it means "append". For primitive operations, specify +primitive: true+ to skip various checks performed to guarantee the self-consistency as an instance of this class.
59
+ # * {#<<} is disabled.
60
+ # * Read:
61
+ # * Read: #to_a gives the standard Array, and then you can do whatever manipulation allowed for Array. For example, if you delete an element in the returned Array, that does not affect the original {Part} instance. However, it is a shallow copy, and hence if you alter an element of it destructively (such as String#replace), the original instance, too, is affected.
62
+ # * The methods {#[]} (or its alias {#slice}) have some restrictions, such as, an odd number of elements cannot be retrieved, so as to conform the returned value is a valid instance of this class.
63
+ # * Update:
64
+ # * Replace: {#[]=} has some restrictions, such as, if multiple elements are replaced, they have to be pairs of Paragraph and Boundary. To skip all the checks, do {#insert} with +primitive: true+
65
+ # * Delete:
66
+ # * Delete: {#slice!} to delete. For primitive operations, specify +primitive: true+ to skip various checks performed to guarantee the self-consistency as an instance of this class.
67
+ # * +#delete_at+ is disabled. +#delete+, +#delete_if+, +#reject!+, +#select!+, +#filter!+, +#keep_if+ (and +#drop_while+ and +#take_whie+ in recent Ruby) remain enabled, but if you use them, make sure to use them carefully at your own risk, as no self-consistency checks would be performed automatically.
68
+ #
69
+ # @author Masa Sakano (Wise Babel Ltd)
70
+ #
71
+ # @todo methods
72
+ # * flatten
73
+ # * SAFE level for command-line tools?
74
+ #
75
+ class Part < Array
76
+
77
+ include PlainText::Util
78
+
79
+ # Error messages
80
+ ERR_MSGS = {
81
+ even_num: 'even number of elements must be specified.',
82
+ use_to_a: 'To handle it as an Array, use to_a first.',
83
+ }
84
+
85
+ # @param arin [Array] of [Paragraph1, Boundary1, Para2, Bd2, ...] or Part/Paragraph if boundaries is given
86
+ # @param boundaries [Array] of Boundary
87
+ # @option recursive: [Boolean] if true (Default), normalize recursively.
88
+ # @option compact: [Boolean] if true (Default), pairs of nil paragraph and boundary are removed. Otherwise, nil is converted to an empty string.
89
+ # @option compacter: [Boolean] if true (Default), pairs of nil or empty paragraph and boundary are removed.
90
+ # @return [self]
91
+ def initialize(arin, boundaries=nil, recursive: true, compact: true, compacter: true)
92
+ if !boundaries
93
+ super(arin)
94
+ else
95
+
96
+ armain = []
97
+ arin.each_with_index do |ea_e, i|
98
+ armain << ea_e
99
+ armain << (boundaries[i] || Boundary.new(''))
100
+ end
101
+ super armain
102
+ end
103
+ normalize!(recursive: recursive, compact: compact, compacter: compacter)
104
+ end
105
+
106
+ # Parses a given string (or {Part}) and returns this class of instance.
107
+ #
108
+ # @param inprm [String, Array, Part]
109
+ # @option rule: [PlainText::ParseRule]
110
+ # @return [PlainText::Part]
111
+ def self.parse(inprm, rule: PlainText::ParseRule::RuleConsecutiveLbs)
112
+ arin = rule.apply(inprm)
113
+ self.new(arin)
114
+ end
115
+
116
+ ####################################################
117
+ # Instance methods
118
+ ####################################################
119
+
120
+ ##########
121
+ # Unique instance methods (not existing in Array)
122
+ ##########
123
+
124
+ # Returns an array of boundary parts (odd-number-index parts), consisting of Boundaries
125
+ #
126
+ # @return [Array<Boundary>]
127
+ # @see #parts
128
+ def boundaries
129
+ select.with_index { |_, i| i.odd? } rescue select.each_with_index { |_, i| i.odd? } # Rescue for Ruby 2.1 or earlier
130
+ end
131
+
132
+ # returns all the Boundaries immediately before the index and at it as an Array
133
+ #
134
+ # See {#squash_boundary_at!} to squash them.
135
+ #
136
+ # @param index [Integer]
137
+ # @return [Array, nil] nil if a too large index is specified.
138
+ def boundary_extended_at(index)
139
+ (i_pos = get_valid_ipos_for_boundary(index)) || return
140
+ arret = []
141
+ prt = self[i_pos-1]
142
+ arret = prt.public_send(__method__, -1) if prt.class.method_defined? __method__
143
+ arret << self[index]
144
+ end
145
+
146
+ # Returns a dup-ped instance with all the Arrays and Strings dup-ped.
147
+ #
148
+ # @return [Part]
149
+ def deepcopy
150
+ dup.map!{ |i| i.class.method_defined?(:deepcopy) ? i.deepcopy : i.dup }
151
+ end
152
+
153
+ # each method for boundaries only, providing also the index (always an odd number) to the block.
154
+ #
155
+ # For just looping over the elements of {#boundaries}, do simply
156
+ #
157
+ # boundaries.each do |ec|
158
+ # end
159
+ #
160
+ # The indices provided in this method are for the main Array,
161
+ # and hence different from {#boundaries}.each_with_index
162
+ #
163
+ # @param (see #map_boundaries_with_index)
164
+ # @return as self
165
+ def each_boundaries_with_index(**kwd, &bl)
166
+ map_boundaries_core(map: false, with_index: true, **kwd, &bl)
167
+ end
168
+
169
+ # each method for parts only, providing also the index (always an even number) to the block.
170
+ #
171
+ # For just looping over the elements of {#parts}, do simply
172
+ #
173
+ # parts.each do |ec|
174
+ # end
175
+ #
176
+ # The indices provided in this method are for the main Array,
177
+ # and hence different from {#parts}.each_with_index
178
+ #
179
+ # @param (see #map_parts_with_index)
180
+ # @return as self
181
+ def each_parts_with_index(**kwd, &bl)
182
+ map_parts_core(map: false, with_index: false, **kwd, &bl)
183
+ end
184
+
185
+ # The first significant (=non-empty) element.
186
+ #
187
+ # If the returned value is non-nil and destructively altered, self changes.
188
+ #
189
+ # @return [Integer, nil] if self.empty? nil is returned.
190
+ def first_significant_element
191
+ (i = first_significant_index) || return
192
+ self[i]
193
+ end
194
+
195
+ # Index of the first significant (=non-empty) element.
196
+ #
197
+ # If every element is empty, the last index is returned.
198
+ #
199
+ # @return [Integer, nil] if self.empty? nil is returned.
200
+ def first_significant_index
201
+ return nil if empty?
202
+ each_index do |i|
203
+ return i if self[i] && !self[i].empty? # self for sanity
204
+ end
205
+ return size-1
206
+ end
207
+
208
+ # True if the index should be semantically for Paragraph?
209
+ #
210
+ # @param i [Integer] index for the array of self
211
+ # @option skip_check: [Boolean] if true (Default: false), skip conversion of the negative index to positive.
212
+ # @see #parts
213
+ def index_para?(i, skip_check: false)
214
+ skip_check ? i.even? : positive_array_index_checked(i, self).even?
215
+ end
216
+
217
+ # The last significant (=non-empty) element.
218
+ #
219
+ # If the returned value is non-nil and destructively altered, self changes.
220
+ #
221
+ # @return [Integer, nil] if self.empty? nil is returned.
222
+ def last_significant_element
223
+ (i = last_significant_index) || return
224
+ self[i]
225
+ end
226
+
227
+ # Index of the last significant (=non-empty) element.
228
+ #
229
+ # If every element is empty, 0 is returned.
230
+ #
231
+ # @return [Integer, nil] if self.empty? nil is returned.
232
+ def last_significant_index
233
+ return nil if empty?
234
+ (0..(size-1)).to_a.reverse.each do |i|
235
+ return i if self[i] && !self[i].empty? # self for sanity
236
+ end
237
+ return 0
238
+ end
239
+
240
+ # map method for boundaries only, returning a copied self.
241
+ #
242
+ # If recursive is true (Default), any Boundaries in the descendant Parts are also handled.
243
+ #
244
+ # If a Boundary is set nil or empty, along with the preceding Paragraph,
245
+ # the pair is removed from the returned instance in Default (:compact and :compacter options
246
+ # - see {#initialize} for detail)
247
+ #
248
+ # @option recursive: [Boolean] if true (Default), map is performed recursively.
249
+ # @return as self
250
+ # @see #initialize for the other options (:compact and :compacter)
251
+ def map_boundaries(**kwd, &bl)
252
+ map_boundaries_core(with_index: false, **kwd, &bl)
253
+ end
254
+
255
+ # map method for boundaries only, providing also the index (always an odd number) to the block, returning a copied self.
256
+ #
257
+ # @param (see #map_boundaries)
258
+ # @return as self
259
+ def map_boundaries_with_index(**kwd, &bl)
260
+ map_boundaries_core(with_index: true, **kwd, &bl)
261
+ end
262
+
263
+ # map method for parts only, returning a copied self.
264
+ #
265
+ # If recursive is true (Default), any Paragraphs in the descendant Parts are also handled.
266
+ #
267
+ # If a Paragraph is set nil or empty, along with the following Boundary,
268
+ # the pair is removed from the returned instance in Default (:compact and :compacter options
269
+ # - see {#initialize} for detail)
270
+ #
271
+ # @option recursive: [Boolean] if true (Default), map is performed recursively.
272
+ # @return as self
273
+ # @see #initialize for the other options (:compact and :compacter)
274
+ def map_parts(**kwd, &bl)
275
+ map_parts_core(with_index: false, **kwd, &bl)
276
+ end
277
+
278
+ # map method for parts only, providing also the index (always an even number) to the block, returning a copied self.
279
+ #
280
+ # @param (see #map_parts)
281
+ # @return as self
282
+ def map_parts_with_index(**kwd, &bl)
283
+ map_parts_core(with_index: false, **kwd, &bl)
284
+ end
285
+
286
+ # Normalize the content, making sure it has an even number of elements
287
+ #
288
+ # The even and odd number elements are, if bare Strings or Array, converted into
289
+ # Paeagraph and Boundary, or Part, respectively. If not, Exception is raised.
290
+ # Note nil is conveted into either an empty Paragraph or Boundary.
291
+ #
292
+ # @option recursive: [Boolean] if true (Default), normalize recursively.
293
+ # @option ignore_array_boundary: [Boolean] if true (Default), even if a Boundary element (odd-numbered index) is an Array, ignore it.
294
+ # @option compact: [Boolean] if true (Default), pairs of nil paragraph and boundary are removed. Otherwise, nil is converted to an empty string.
295
+ # @option compacter: [Boolean] if true (Default), pairs of nil or empty paragraph and boundary are removed.
296
+ # @return [self]
297
+ def normalize!(recursive: true, ignore_array_boundary: true, compact: true, compacter: true)
298
+ # Trim pairs of consecutive Paragraph and Boundary of nil
299
+ size_parity = (size.even? ? 0 : 1)
300
+ if (compact || compacter) && (size > 0+size_parity)
301
+ ((size-2-size_parity)..0).each do |i|
302
+ # Loop over every Paragraph
303
+ next if i.odd?
304
+ slice! i, 2 if compact && !self[i] && !self[i+1]
305
+ slice! i, 2 if compacter && (!self[i] || self[i].empty?) && (!self[i+1] || self[i+1].empty?)
306
+ end
307
+ end
308
+
309
+ i = -1
310
+ map!{ |ea|
311
+ i += 1
312
+ normalize_core(ea, i, recursive: recursive)
313
+ }
314
+ insert_original_b4_part(size, Boundary.new('')) if size.odd?
315
+ self
316
+ end
317
+
318
+ # Non-destructive version of {#normalize!}
319
+ #
320
+ # @option recursive: [Boolean] if true (Default), normalize recursively.
321
+ # @option ignore_array_boundary: [Boolean] if true (Default), even if a Boundary element (odd-numbered index) is an Array, ignore it.
322
+ # @option compact: [Boolean] if true (Default), pairs of nil paragraph and boundary are removed. Otherwise, nil is converted to an empty string.
323
+ # @option compacter: [Boolean] if true (Default), pairs of nil or empty paragraph and boundary are removed.
324
+ # @return as self
325
+ # @see #normalize!
326
+ def normalize(recursive: true, ignore_array_boundary: true, compact: true, compacter: true)
327
+ # Trims pairs of consecutive Paragraph and Boundary of nil
328
+ arall = to_a
329
+ size_parity = (size.even? ? 0 : 1)
330
+ if (compact || compacter) && (size > 0+size_parity)
331
+ ((size-2-size_parity)..0).each do |i|
332
+ # Loop over every Paragraph
333
+ next if i.odd?
334
+ arall.slice! i, 2 if compact && !self[i] && !self[i+1]
335
+ arall.slice! i, 2 if compacter && (!self[i] || self[i].empty?) && (!self[i+1] || self[i+1].empty?)
336
+ end
337
+ end
338
+
339
+ i = -1
340
+ self.class.new(
341
+ arall.map{ |ea|
342
+ i += 1
343
+ normalize_core(ea, i, recursive: recursive)
344
+ } + (arall.size.odd? ? [Boundary.new('')] : [])
345
+ )
346
+ end
347
+
348
+
349
+ # Returns an array of substantial parts (even-number-index parts), consisting of Part and/or Paragraph
350
+ #
351
+ # @return [Array<Part, Paragraph>]
352
+ # @see #boundaries
353
+ def parts
354
+ select.with_index { |_, i| i.even? } rescue select.each_with_index { |_, i| i.even? } # Rescue for Ruby 2.1 or earlier
355
+ # ret.freeze
356
+ end
357
+
358
+ # Reparses self or a part of it.
359
+ #
360
+ # @param str [String]
361
+ # @option rule: [PlainText::ParseRule] (PlainText::ParseRule::RuleConsecutiveLbs)
362
+ # @option name: [String, Symbol, Integer, nil] Identifier of rule, if need to specify.
363
+ # @option range: [Range, nil] Range of indices of self to reparse. In Default, the entire self.
364
+ # @return [self]
365
+ def reparse!(rule: PlainText::ParseRule::RuleConsecutiveLbs, name: nil, range: (0..-1))
366
+ insert range.begin, self.class.parse((range ? self[range] : self), rule: rule, name: name)
367
+ self
368
+ end
369
+
370
+ # Non-destructive version of {reparse!}
371
+ #
372
+ # @param (see #reparse!)
373
+ # @return [PlainText::Part]
374
+ def reparse(**kwd)
375
+ ret = self.dup
376
+ ret.reparse!(**kwd)
377
+ ret
378
+ end
379
+
380
+
381
+ # Emptifies all the Boundaries immediately before the Boundary at the index and squashes it to the one at it.
382
+ #
383
+ # See {#boundary_extended_at} to view them.
384
+ #
385
+ # @param index [Integer]
386
+ # @return [Boundary, nil] nil if a too large index is specified.
387
+ def squash_boundary_at!(index)
388
+ (i_pos = get_valid_ipos_for_boundary(index)) || return
389
+ prt = self[i_pos-1]
390
+ m = :emptify_last_boundaries!
391
+ self[i_pos] << prt.public_send(m) if prt.class.method_defined? m
392
+ self[i_pos]
393
+ end
394
+
395
+
396
+ # Wrapper of {#squash_boundary_at!} to loop over the whole {Part}
397
+ #
398
+ # @return [self]
399
+ def squash_boundaryies!
400
+ each_boundaries_with_index do |ec, i|
401
+ squash_boundary_at!(i)
402
+ end
403
+ self
404
+ end
405
+
406
+
407
+ # Boundary sub-class name only
408
+ #
409
+ # Make sure your class is a child class of Part
410
+ # Otherwise this method would not be inherited, obviously.
411
+ #
412
+ # @example
413
+ # class PlainText::Part
414
+ # class Section < self
415
+ # class Subsection < self; end # It must be a child class!
416
+ # end
417
+ # end
418
+ # ss = PlainText::Part::Section::Subsection.new ["abc"]
419
+ # ss.subclass_name # => "Section::Subsection"
420
+ #
421
+ # @return [String]
422
+ # @see PlainText::Part#subclass_name
423
+ def subclass_name
424
+ printf "__method__=(%s)\n", __method__
425
+ self.class.name.split(/\A#{Regexp.quote method(__method__).owner.name}::/)[1] || ''
426
+ end
427
+
428
+ ##########
429
+ # Overwriting instance methods of the parent Object or Array class
430
+ ##########
431
+
432
+ # Original equal and plus operators of Array
433
+ hsmethod = {
434
+ :equal_original_b4_part => :==,
435
+ :substitute_original_b4_part => :[]=,
436
+ :insert_original_b4_part => :insert,
437
+ :delete_at_original_b4_part => :delete_at,
438
+ :slice_original_b4_part => :slice,
439
+ :slice_original_b4_part! => :slice!,
440
+ }
441
+
442
+ hsmethod.each_pair do |k, ea_orig|
443
+ if self.method_defined?(k)
444
+ # To Developer: If you see this message, switch the DEBUG flag on (-d option) and run it.
445
+ warn sprintf("WARNING: Method %s#%s has been already defined, which should not be. Contact the code developer. Line %d in %s%s", self.name, k.to_s, __FILE__, __LINE__, ($DEBUG ? "\n"+caller_locations.join("\n").map{|i| " "+i} : ""))
446
+ else
447
+ alias_method k, ea_orig
448
+ end
449
+ end
450
+
451
+ alias_method :substit, :substitute_original_b4_part
452
+
453
+ ########## Most basic methods (Object) ##########
454
+
455
+ # @return [String]
456
+ def inspect
457
+ self.class.name + super
458
+ end
459
+
460
+ # # clone
461
+ # #
462
+ # # Redefines Array#clone so the instance variables are also cloned.
463
+ # #
464
+ # # @return [self]
465
+ # def clone
466
+ # copied = super
467
+ # val = (@sep.clone rescue @sep) # rescue in case of immutable.
468
+ # copied.instance_eval{ @sep = val }
469
+ # copied
470
+ # end
471
+
472
+ # Equal operator
473
+ #
474
+ # Unless both are kind of Part instances, false is returned.
475
+ # If you want to make comparison in the Array level, do
476
+ # p1.to_a == a1.to_a
477
+ #
478
+ # @param other [Object]
479
+ def ==(other)
480
+ return false if !other.class.method_defined?(:to_ary)
481
+ %i(parts boundaries).each do |ea_m| # %i(...) defined in Ruby 2.0 and later
482
+ return false if !other.class.method_defined?(ea_m) || (self.public_send(ea_m) != other.public_send(ea_m)) # public_send() defined in Ruby 2.0 (1.9?) and later
483
+ end
484
+ super
485
+ end
486
+
487
+
488
+ # # Multiplication operator
489
+ # #
490
+ # # @param other [Integer, String]
491
+ # # @return as self
492
+ # def *(other)
493
+ # super
494
+ # end
495
+
496
+
497
+ # Plus operator
498
+ #
499
+ # @param other [Object]
500
+ # @return as self
501
+ def +(other)
502
+ # ## The following is strict, but a bit redundant.
503
+ # # is_part = true # Whether "other" is a Part class instance.
504
+ # # %i(to_ary parts boundaries).each do |ea_m| # %i(...) defined in Ruby 2.0 and later
505
+ # # is_part &&= other.class.method_defined?(ea_m)
506
+ # # end
507
+
508
+ # begin
509
+ # other_even_odd =
510
+ # ([other.parts, other.boundaries] rescue even_odd_arrays(self, size_even: true, filler: ""))
511
+ # rescue NoMethodError
512
+ # raise TypeError, sprintf("no implicit conversion of %s into %s", other.class.name, self.class.name)
513
+ # end
514
+
515
+ # # eg., if self is PlainText::Part::Section, the returned object is the same.
516
+ # ret = self.class.new(self.parts+other_even_odd[0], self.boundaries+other_even_odd[1])
517
+ ret = self.class.new super
518
+ ret.normalize!
519
+ end
520
+
521
+
522
+ # Minus operator
523
+ #
524
+ # @param other [Object]
525
+ # @return as self
526
+ def -(other)
527
+ ret = self.class.new super
528
+ ret.normalize!
529
+ end
530
+
531
+ # Array#<< is now undefined
532
+ # (because the instances of this class must take always an even number of elements).
533
+ undef_method(:<<) if method_defined?(:<<)
534
+
535
+ # Array#delete_at is now undefined
536
+ # (because the instances of this class must have always an even number of elements).
537
+ undef_method(:delete_at) if method_defined?(:delete_at)
538
+
539
+ # Returns a partial Part-Array (or Object, if a single Integer is specified)
540
+ #
541
+ # Because the returned object is this class of instance (when a pair of Integer or Range
542
+ # is specified), only an even number of elements, starting from an even number of index,
543
+ # is allowed.
544
+ #
545
+ # @param arg1 [Integer, Range]
546
+ # @option arg2 [Integer, NilClass]
547
+ # @return [Object]
548
+ def [](arg1, *rest)
549
+ arg2 = rest[0]
550
+ return super(arg1) if !arg2 && !arg1.class.method_defined?(:exclude_end?)
551
+
552
+ check_bracket_args_type_error(arg1, arg2) # Args are now either (Int, Int) or (Range)
553
+
554
+ if arg2
555
+ size2ret = size2extract(arg1, arg2, ignore_error: true) # maybe nil (if the index is too small).
556
+ raise ArgumentError, ERR_MSGS[:even_num]+" "+ERR_MSGS[:use_to_a] if size2ret.odd?
557
+ begin
558
+ raise ArgumentError, "odd index is not allowed as the starting index for #{self.class.name}. It must be even. "+ERR_MSGS[:use_to_a] if positive_array_index_checked(arg1, self).odd?
559
+ rescue TypeError, IndexError
560
+ # handled by super
561
+ end
562
+ return super
563
+ end
564
+
565
+ begin
566
+ rang = normalize_index_range(arg1)
567
+ rescue IndexError #=> err
568
+ return nil
569
+ # raise RangeError, err.message
570
+ end
571
+
572
+ raise RangeError if rang.begin < 0 || rang.end < 0
573
+
574
+ # The end is smaller than the begin in the positive index. Empty instance of this class is returned.
575
+ if rang.end < rang.begin
576
+ return super
577
+ end
578
+
579
+ raise RangeError, "odd index is not allowed as the starting Range for #{sefl.class.name}. It must be even. "+ERR_MSGS[:use_to_a] if rang.begin.odd?
580
+ size2ret = size2extract(rang, skip_renormalize: true)
581
+ raise ArgumentError, ERR_MSGS[:even_num]+" "+ERR_MSGS[:use_to_a] if size2ret.odd?
582
+ super
583
+ end
584
+
585
+
586
+ # Replaces some of the Array content.
587
+ #
588
+ # @param arg1 [Integer, Range]
589
+ # @option arg2 [Integer, NilClass]
590
+ # @return [Object]
591
+ def []=(arg1, *rest)
592
+ if rest.size == 1
593
+ arg2, val = [nil, rest[-1]]
594
+ else
595
+ arg2, val = rest
596
+ end
597
+
598
+ # Simple substitution to a single element
599
+ return super(arg1, val) if !arg2 && !arg1.class.method_defined?(:exclude_end?)
600
+
601
+ check_bracket_args_type_error(arg1, arg2) # Args are now either (Int, Int) or (Range)
602
+
603
+ # raise TypeError, "object to replace must be Array type with an even number of elements." if !val.class.method_defined?(:to_ary) || val.size.odd?
604
+
605
+ vals = (val.to_ary rescue [val])
606
+ if arg2
607
+ size2delete = size2extract(arg1, arg2, ignore_error: true) # maybe nil (if the index is too small).
608
+ raise ArgumentError, "odd-even parity of size of array to replace must be identical to that to slice." if size2delete && ((size2delete % 2) != (vals.size % 2))
609
+ return super
610
+ end
611
+
612
+ begin
613
+ rang = normalize_index_range(arg1)
614
+ rescue IndexError => err
615
+ raise RangeError, err.message
616
+ end
617
+
618
+ raise RangeError if rang.begin < 0 || rang.end < 0
619
+
620
+ # The end is smaller than the begin in the positive index. It is the same as insert (default in Ruby), except it returns the replaced Object (which may not be an Array).
621
+ if rang.end < rang.begin
622
+ insert(arg1, *vals)
623
+ return val
624
+ end
625
+
626
+ size2delete = size2extract(rang, skip_renormalize: true)
627
+ raise ArgumentError, "odd-even parity of size of array to replace must be identical to that to slice." if size2delete && ((size2delete % 2) != (vals.size % 2))
628
+ ret = super
629
+
630
+ # The result may not be in an even number anymore. Correct it.
631
+ push Boundary.new("") if size.odd?
632
+
633
+ # Original method may fill some part of the array with String or even nil.
634
+ normalize!
635
+ ret
636
+ end
637
+
638
+ # Array#insert
639
+ #
640
+ # The most basic method to add/insert elements to self. Called from {#[]=} and {#push}, for example.
641
+ #
642
+ # If ind is greater than size, a number of "", as opposed to nil, are inserted.
643
+ #
644
+ # @param ind [Index]
645
+ # @param rest [Array] This must have an even number of arguments, unless ind is larger than the array size and an odd number.
646
+ # @option primitive: [String] if true (Def: false), the original {#insert_original_b4_part} is called.
647
+ # @return [self]
648
+ def insert(ind, *rest, primitive: false)
649
+ return insert_original_b4_part(ind, *rest) if primitive
650
+ ipos = positive_array_index_checked(ind, self)
651
+ if rest.size.even? && (ipos > size - 1) && ipos.even? # ipos.even? is equivalent to index_para?(ipos), i.e., "is the index for Paragraph?"
652
+ raise ArgumentError, sprintf("number of arguments (%d) must be odd for index %s.", rest.size, ind)
653
+ elsif rest.size.odd? && (ipos <= size - 1)
654
+ raise ArgumentError, sprintf("number of arguments (%d) must be even.", rest.size)
655
+ end
656
+
657
+ if ipos >= size
658
+ rest = Array.new(ipos - size).map{|i| ""} + rest
659
+ ipos = size
660
+ end
661
+
662
+ super(ipos, rest)
663
+ end
664
+
665
+
666
+ # Delete elements and return the deleted content or nil if nothing is deleted.
667
+ #
668
+ # The number of elements to be deleted must be even.
669
+ #
670
+ # @param arg1 [Integer, Range]
671
+ # @option arg2 [Integer, NilClass]
672
+ # @option primitive: [Boolean] if true (Def: false), the original {#insert_original_b4_part} is called.
673
+ # @return as self or NilClass
674
+ def slice!(arg1, *rest, primitive: false)
675
+ return slice_original_b4_part!(arg1, *rest) if primitive
676
+
677
+ arg2 = rest[0]
678
+
679
+ # Simple substitution to a single element
680
+ raise ArgumentError, ERR_MSGS[:even_num] if !arg2 && !arg1.class.method_defined?(:exclude_end?)
681
+
682
+ check_bracket_args_type_error(arg1, arg2) # Args are now either (Int, Int) or (Range)
683
+
684
+ if arg2
685
+ size2delete = size2extract(arg1, arg2, ignore_error: true) # maybe nil (if the index is too small).
686
+ # raise ArgumentError, ERR_MSGS[:even_num] if arg2.to_int.odd?
687
+ raise ArgumentError, ERR_MSGS[:even_num] if size2delete && size2delete.odd?
688
+ raise ArgumentError, "odd index is not allowed as the starting Range for #{self.class.name}. It must be even." if arg1.odd? # Because the returned value is this class of instance.
689
+ return super(arg1, *rest)
690
+ end
691
+
692
+ begin
693
+ rang = normalize_index_range(arg1)
694
+ rescue IndexError => err
695
+ raise RangeError, err.message
696
+ end
697
+
698
+ raise RangeError if rang.begin < 0 || rang.end < 0
699
+
700
+ return super(arg1, *rest) if (rang.begin > rang.end) # nil or [] is returned
701
+
702
+ size2delete = size2extract(rang, skip_renormalize: true)
703
+ raise ArgumentError, ERR_MSGS[:even_num] if size2delete && size2delete.odd?
704
+ raise ArgumentError, "odd index is not allowed as the starting Range for #{self.class.name}. It must be even." if rang.begin.odd? # Because the returned value is this class of instance.
705
+ super(arg1, *rest)
706
+ end
707
+
708
+
709
+ ########## Other methods of Array ##########
710
+
711
+ # Array#compact!
712
+ #
713
+ # If changed, re-{#normalize!} it.
714
+ #
715
+ # @return [self, NilClass]
716
+ def compact!
717
+ ret = super
718
+ ret ? ret.normalize!(recursive: false) : ret
719
+ end
720
+
721
+ # Array#concat
722
+ #
723
+ # @see #insert
724
+ #
725
+ # @param *rest [Array<Array>]
726
+ # @return [self]
727
+ def concat(*rest)
728
+ insert(size, *(rest.sum([])))
729
+ end
730
+
731
+ # Array#push
732
+ #
733
+ # @see #concat
734
+ #
735
+ # @param ary [Array]
736
+ # @return [self]
737
+ def push(*rest)
738
+ concat(rest)
739
+ end
740
+
741
+ # {#append} is an alias to {#push}
742
+ alias :append :push
743
+
744
+ # {#slice} is an alias to {#[]}
745
+ alias :slice :[]
746
+
747
+
748
+ ##########
749
+ # Private instance methods
750
+ ##########
751
+
752
+ private
753
+
754
+ # Checking whether index-type arguments conform
755
+ #
756
+ # After this, it is guaranteed the arguments are either (Integer, Integer) or (Range, nil).
757
+ #
758
+ # @param arg1 [Integer, Range] Starting index or Range. Maybe including negative values.
759
+ # @option arg2 [Integer, NilClass] Size.
760
+ # @return [NilClass]
761
+ # @raise [TypeError] if not conforms.
762
+ def check_bracket_args_type_error(arg1, arg2=nil)
763
+ if arg2
764
+ raise TypeError, sprintf("no implicit conversion of #{arg2.class} into Integer") if !arg2.class.method_defined?(:to_int)
765
+ else
766
+ raise TypeError if !arg1.class.method_defined?(:exclude_end?)
767
+ end
768
+ end
769
+ private :check_bracket_args_type_error
770
+
771
+
772
+ # Emptifies all the Boundaries immediately before the index and squashes it to the one at it.
773
+ #
774
+ # @return [Boundary] all the descendants' last Boundaries merged.
775
+ def emptify_last_boundaries!
776
+ return Boundary::Empty.dup if size == 0
777
+ ret = ""
778
+ ret << prt.public_send(__method__) if prt.class.method_defined? __method__
779
+ ret << self[-1]
780
+ self[-1] = Boundary::Empty.dup
781
+ ret
782
+ end
783
+ private :emptify_last_boundaries!
784
+
785
+
786
+ # Returns a positive Integer index guaranteed to be 1 or greater and smaller than the size.
787
+ #
788
+ # @param index [Integer]
789
+ # @return [Integer, nil] nil if a too large index is specified.
790
+ def get_valid_ipos_for_boundary(index)
791
+ i_pos = positive_array_index_checked(index, self)
792
+ raise ArgumentError, "Index #{index} specified was for Part/Paragraph, which should be for Boundary." if index_para?(i_pos, skip_check: true)
793
+ (i_pos > size - 1) ? nil : i_pos
794
+ end
795
+ private :get_valid_ipos_for_boundary
796
+
797
+
798
+ # Core routine for {#map_boundaries} and similar.
799
+ #
800
+ # @option map opts: [Boolean] if true (Default), map is performed. Else just each.
801
+ # @option with_index: [Boolean] if true (Default: false), yield with also index
802
+ # @option recursive: [Boolean] if true (Default), map is performed recursively.
803
+ # @return as self if map: is true, else void
804
+ def map_boundaries_core(map: true, with_index: false, recursive: true, **kwd, &bl)
805
+ ind = -1
806
+ arnew = map{ |ec|
807
+ ind += 1
808
+ if recursive && index_para?(ind, skip_check: true) && ec.class.method_defined?(__method__)
809
+ ec.public_send(__method__, recursive: true, **kwd, &bl)
810
+ elsif !index_para?(ind, skip_check: true)
811
+ with_index ? yield(ec, ind) : yield(ec)
812
+ else
813
+ ec
814
+ end
815
+ }
816
+ self.class.new arnew, recursive: recursive, **kwd if map
817
+ end
818
+ private :map_boundaries_core
819
+
820
+ # Core routine for {#map_parts}
821
+ #
822
+ # @option map: [Boolean] if true (Default), map is performed. Else just each.
823
+ # @option with_index: [Boolean] if true (Default: false), yield with also index
824
+ # @option recursive: [Boolean] if true (Default), map is performed recursively.
825
+ # @return as self
826
+ # @see #initialize for the other options (:compact and :compacter)
827
+ def map_parts_core(map: true, with_index: false, recursive: true, **kwd, &bl)
828
+ ind = -1
829
+ new_parts = parts.map{ |ec|
830
+ ind += 1
831
+ if recursive && ec.class.method_defined?(__method__)
832
+ ec.public_send(__method__, recursive: true, **kwd, &bl)
833
+ else
834
+ with_index ? yield(ec, ind) : yield(ec)
835
+ end
836
+ }
837
+ self.class.new new_parts, boundaries, recursive: recursive, **kwd if map
838
+ end
839
+ private :map_parts_core
840
+
841
+ # Core routine for {#normalize!} and {#normalize}
842
+ #
843
+ # @param ea [Array, String, NilClass] the element to evaluate
844
+ # @param i [Integer] Main array index
845
+ # @option recursive: [Boolean] if true (Default), normalize recursively.
846
+ # @option ignore_array_boundary: [Boolean] if true (Default), even if a Boundary element (odd-numbered index) is an Array, ignore it.
847
+ # @return [Part, Paragraph, Boundary]
848
+ def normalize_core(ea, i, recursive: true, ignore_array_boundary: true)
849
+ if ea.class.method_defined?(:to_ary)
850
+ if index_para?(i, skip_check: true) || ignore_array_boundary
851
+ (/\APlainText::/ =~ ea.class.name && defined?(ea.normalize)) ? (recursive ? ea.normalize : ea) : self.class.new(ea, recursive: recursive)
852
+ else
853
+ raise "Index ({#i}) is an Array or its child, but it should be Boundary or String."
854
+ end
855
+ elsif ea.class.method_defined?(:to_str)
856
+ if /\APlainText::/ =~ ea.class.name
857
+ # Paragraph or Boundary
858
+ ea.unicode_normalize
859
+ else
860
+ if index_para?(i, skip_check: true)
861
+ Paragraph.new(ea.unicode_normalize || "")
862
+ else
863
+ Boundary.new( ea.unicode_normalize || "")
864
+ end
865
+ end
866
+ else
867
+ raise ArgumentError, "Unrecognised elements for #{self.class}: "+ea.inspect
868
+ end
869
+ end
870
+ private :normalize_core
871
+
872
+
873
+ # Returns (inclusive, i.e., not "...") Range of non-negative indices
874
+ #
875
+ # @param rng [Range] It has to be a Range
876
+ # @return [Range]
877
+ # @raise [IndexError] if too negative index is specified.
878
+ def normalize_index_range(rng, **kwd)
879
+ # NOTE to developers: (0..-1).to_a returns [] (!)
880
+ arpair = [rng.begin, rng.end].to_a.map{ |i| positive_array_index_checked(i, self, **kwd) }
881
+ arpair[1] -= 1 if rng.exclude_end?
882
+ (arpair[0]..arpair[1])
883
+ end
884
+ private :normalize_index_range
885
+
886
+
887
+ # Returns the size (the number of elements) to extract
888
+ #
889
+ # taking into account the size of self
890
+ #
891
+ # 1. if (3, 2) is specified when self.size==4, this returns 1.
892
+ # 2. if (3..2) is specified, this returns 0.
893
+ #
894
+ # Make sure to call {#check_bracket_args_type_error} beforehand.
895
+ #
896
+ # @param arg1 [Integer, Range] Starting index or Range. Maybe including negative values.
897
+ # @option arg2 [Integer, NilClass] Size.
898
+ # @option ignore_error: [Boolean] if true (Def: false), nil is returned instead of raising IndexError (when a negative index is too small)
899
+ # @option skip_renormalize: [Boolean] if true (Def: false), the given Range is assumed to be already normalized by {#normalize_index_range}
900
+ # @return [Integer, NilClass] nil if an Error is raised with ignore_error being true
901
+ def size2extract(arg1, arg2=nil, ignore_error: false, skip_renormalize: false, **kwd)
902
+ begin
903
+ if arg1.class.method_defined? :exclude_end?
904
+ rng = arg1
905
+ rng = normalize_index_range(rng, **kwd) if !skip_renormalize
906
+ else
907
+ ipos = positive_array_index_checked(arg1, self)
908
+ rng = (ipos..(ipos+arg2.to_int-1))
909
+ end
910
+ return 0 if rng.begin > size-1
911
+ return 0 if rng.begin > rng.end
912
+ return [rng.end, size-1].min-rng.begin+1
913
+ rescue IndexError
914
+ return nil if ignore_error
915
+ raise
916
+ end
917
+ end
918
+ private :size2extract
919
+
920
+ end # class Part < Array
921
+
922
+ end # module PlainText
923
+
924
+
925
+ ####################################################
926
+ # Modifies Array
927
+ ####################################################
928
+
929
+ class Array
930
+
931
+ # Original equal and plus operators of Array
932
+ hsmethod = {
933
+ :equal_original_b4_part? => :== ,
934
+ # :plus_operator_original_b4_part => :+
935
+ }
936
+
937
+ hsmethod.each_pair do |k, ea_orig|
938
+ if self.method_defined?(k)
939
+ # To Developer: If you see this message, switch the DEBUG flag on (-d option) and run it.
940
+ warn sprintf("WARNING: Method %s#%s has been already defined, which should not be. Contact the code developer. Line %d in %s%s", self.name, k.to_s, __FILE__, __LINE__, ($DEBUG ? "\n"+caller_locations.join("\n").map{|i| " "+i} : ""))
941
+ else
942
+ alias_method k, ea_orig
943
+ end
944
+ end
945
+
946
+ # Equal operator modified to deal with {PlainText::Part}
947
+ #
948
+ # @param other [Object]
949
+ def ==(other)
950
+ return false if !other.class.method_defined?(:to_ary)
951
+ %i(parts boundaries).each do |ea_m| # %i(...) defined in Ruby 2.0 and later
952
+ return equal_original_b4_part?(other) if !other.class.method_defined?(ea_m)
953
+ return false if !self.class.method_defined?(ea_m) || (self.public_send(ea_m) != other.public_send(ea_m)) # public_send() defined in Ruby 2.0 (1.9?) and later
954
+ end
955
+ true
956
+ end
957
+
958
+ end
959
+
960
+ ####################################################
961
+ # require (after the module is defined)
962
+ ####################################################
963
+
964
+ require 'plain_text/part/paragraph'
965
+ require 'plain_text/part/boundary'
966
+ require "plain_text/parse_rule"
967
+
968
+ #######
969
+ # idea
970
+ # * Potentially, a specification of the main array of always an odd number of elements is possible: [Para, Boundary, B, P, B, P], either or both of the last and first of which may be empty.
971
+ # * then joining is very straightforward.
972
+ # * A trouble is, open/close tab structures like HTML/LaTeX-type (+<ul>...</ul>+) are not represented very well.
973
+ #