Class: HTMLDiff::DiffBuilder

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

Constant Summary collapse

VALID_METHODS = 
[:replace, :insert, :delete, :equal]

Instance Method Summary collapse

Constructor Details

#initialize(old_version, new_version, dual = false) ⇒ DiffBuilder

Returns a new instance of DiffBuilder.



18
19
20
21
22
23
24
25
26
27
 
# File 'lib/htmldiff.rb', line 18

def initialize(old_version, new_version, dual = false)
  @old_version, @new_version = old_version, new_version
  @dual = dual
  if @dual
    @new_content = []
    @old_content = []
  else
    @content = []
  end
end

Instance Method Details

#add_matching_words_left(match_in_old, match_in_new, match_size, start_in_old, start_in_new) ⇒ Object 



150
151
152
153
154
155
156
157
158
159
 
# File 'lib/htmldiff.rb', line 150

def add_matching_words_left(match_in_old, match_in_new, match_size, start_in_old, start_in_new)
  while match_in_old > start_in_old and 
        match_in_new > start_in_new and 
        @old_words[match_in_old - 1] == @new_words[match_in_new - 1]
    match_in_old -= 1
    match_in_new -= 1
    match_size += 1
  end
  [match_in_old, match_in_new, match_size]
end

#add_matching_words_right(match_in_old, match_in_new, match_size, end_in_old, end_in_new) ⇒ Object 



161
162
163
164
165
166
167
168
 
# File 'lib/htmldiff.rb', line 161

def add_matching_words_right(match_in_old, match_in_new, match_size, end_in_old, end_in_new)
  while match_in_old + match_size < end_in_old and 
        match_in_new + match_size < end_in_new and
        @old_words[match_in_old + match_size] == @new_words[match_in_new + match_size]
    match_size += 1
  end
  [match_in_old, match_in_new, match_size]
end

#add_special_attribute(opening_tag, type) ⇒ Object 



227
228
229
 
# File 'lib/htmldiff.rb', line 227

def add_special_attribute opening_tag, type
  opening_tag.sub('>', ' difftype="' + type + '">')
end

#build ⇒ Object 



29
30
31
32
33
34
 
# File 'lib/htmldiff.rb', line 29

def build
  split_inputs_to_words
  index_new_words
  operations.each { |op| perform_operation(op) }
  return @dual ? [ @old_content.join, @new_content.join ] : @content.join 
end

#closing_tag?(item) ⇒ Boolean

Returns:

  • (Boolean) 


204
205
206
 
# File 'lib/htmldiff.rb', line 204

def closing_tag?(item)
  item =~ %r!^\s*</[^>]+>\s*$!
end

#convert_html_to_list_of_words(x, use_brackets = false) ⇒ Object 



283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
 
# File 'lib/htmldiff.rb', line 283

def convert_html_to_list_of_words(x, use_brackets = false)
  mode = :char
  current_word  = ''
  words = []
  
  explode(x).each do |char|
    case mode
    when :tag
      if end_of_tag? char
        current_word << (use_brackets ? ']' : '>')
        words << current_word
        current_word = ''
        if whitespace?(char) 
          mode = :whitespace 
        else
          mode = :char
        end
      else
        current_word << char
      end
    when :char
      if start_of_tag? char
        words << current_word unless current_word.empty?
        current_word = (use_brackets ? '[' : '<')
        mode = :tag
      elsif /\s/.match char
        words << current_word unless current_word.empty?
        current_word = char
        mode = :whitespace
      else
        current_word << char
      end
    when :whitespace
      if start_of_tag? char
        words << current_word unless current_word.empty?
        current_word = (use_brackets ? '[' : '<')
        mode = :tag
      elsif /\s/.match char
        current_word << char
      else
        words << current_word unless current_word.empty?
        current_word = char
        mode = :char
      end
    else 
      raise "Unknown mode #{mode.inspect}"
    end
  end
  words << current_word unless current_word.empty?
  words
end

#delete(operation, tagclass = 'diffdel') ⇒ Object 



186
187
188
 
# File 'lib/htmldiff.rb', line 186

def delete(operation, tagclass = 'diffdel')
   insert_tag('del', tagclass, @old_words[operation.start_in_old...operation.end_in_old], @dual ? @old_content : @content)
end

#end_of_tag?(char) ⇒ Boolean

Returns:

  • (Boolean) 


271
272
273
 
# File 'lib/htmldiff.rb', line 271

def end_of_tag?(char)
  char == '>'
end

#equal(operation) ⇒ Object 



190
191
192
193
194
195
196
197
198
 
# File 'lib/htmldiff.rb', line 190

def equal(operation)
  # no tags to insert, simply copy the matching words from one of the versions
  if @dual
    @old_content += @old_words[operation.start_in_old...operation.end_in_old]
    @new_content += @new_words[operation.start_in_new...operation.end_in_new]
  else
    @content += @new_words[operation.start_in_new...operation.end_in_new]
  end
end

#explode(sequence) ⇒ Object 



267
268
269
 
# File 'lib/htmldiff.rb', line 267

def explode(sequence)
  sequence.is_a?(String) ? sequence.split(//) : sequence
end

#extract_consecutive_words(words, &condition) ⇒ Object 



212
213
214
215
216
217
218
219
220
221
222
223
224
225
 
# File 'lib/htmldiff.rb', line 212

def extract_consecutive_words(words, &condition)
  index_of_first_tag = nil
  words.each_with_index do |word, i| 
    if !condition.call(word)
      index_of_first_tag = i
      break
    end
  end
  if index_of_first_tag
    return words.slice!(0...index_of_first_tag)
  else
    return words.slice!(0..words.length)
  end
end

#find_match(start_in_old, end_in_old, start_in_new, end_in_new) ⇒ Object 



114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'lib/htmldiff.rb', line 114

def find_match(start_in_old, end_in_old, start_in_new, end_in_new)

  best_match_in_old = start_in_old
  best_match_in_new = start_in_new
  best_match_size = 0
  
  match_length_at = Hash.new { |h, index| h[index] = 0 }
  
  start_in_old.upto(end_in_old - 1) do |index_in_old|

    new_match_length_at = Hash.new { |h, index| h[index] = 0 }

    @word_indices[@old_words[index_in_old]].each do |index_in_new|
      next  if index_in_new < start_in_new
      break if index_in_new >= end_in_new

      new_match_length = match_length_at[index_in_new - 1] + 1
      new_match_length_at[index_in_new] = new_match_length

      if new_match_length > best_match_size
        best_match_in_old = index_in_old - new_match_length + 1
        best_match_in_new = index_in_new - new_match_length + 1
        best_match_size = new_match_length
      end
    end
    match_length_at = new_match_length_at
  end

#      best_match_in_old, best_match_in_new, best_match_size = add_matching_words_left(
#          best_match_in_old, best_match_in_new, best_match_size, start_in_old, start_in_new)
#      best_match_in_old, best_match_in_new, match_size = add_matching_words_right(
#          best_match_in_old, best_match_in_new, best_match_size, end_in_old, end_in_new)

  return (best_match_size != 0 ? Match.new(best_match_in_old, best_match_in_new, best_match_size) : nil)
end

#index_new_words ⇒ Object 



41
42
43
44
 
# File 'lib/htmldiff.rb', line 41

def index_new_words
  @word_indices = Hash.new { |h, word| h[word] = [] }
  @new_words.each_with_index { |word, i| @word_indices[word] << i }
end

#insert(operation, tagclass = 'diffins') ⇒ Object 



182
183
184
 
# File 'lib/htmldiff.rb', line 182

def insert(operation, tagclass = 'diffins')
  insert_tag('ins', tagclass, @new_words[operation.start_in_new...operation.end_in_new], @dual ? @new_content : @content)
end

#insert_tag(tagname, cssclass, words, content) ⇒ Object

This method encloses words within a specified tag (ins or del), and adds this into @content, with a twist: if there are words contain tags, it actually creates multiple ins or del, so that they don’t include any ins or del. This handles cases like old: ‘<p>a</p>’ new: ‘<p>ab</p><p>c</b>’ diff result: ‘<p>a<ins>b</ins></p><p><ins>c</ins></p>’ this still doesn’t guarantee valid HTML (hint: think about diffing a text containing ins or del tags), but handles correctly more cases than the earlier version.

P.S.: Spare a thought for people who write HTML browsers. They live in this … every day.



242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
 
# File 'lib/htmldiff.rb', line 242

def insert_tag(tagname, cssclass, words, content)
  loop do
    break if words.empty?
    non_tags = extract_consecutive_words(words) { |word| not tag?(word) }
    content << wrap_text(non_tags.join, tagname, cssclass) unless non_tags.empty?

    break if words.empty?
    loop do
      opening = opening_tag? words.first
      closing = closing_tag? words.first
      break unless opening || closing
      if opening
        content << add_special_attribute(words.shift, cssclass)
      else
        content << words.shift
      end
    end
    #content.push(*extract_consecutive_words(words) { |word| tag?(word) })
  end
end

#matching_blocks ⇒ Object 



93
94
95
96
97
 
# File 'lib/htmldiff.rb', line 93

def matching_blocks
  matching_blocks = []
  recursively_find_matching_blocks(0, @old_words.size, 0, @new_words.size, matching_blocks)
  matching_blocks
end

#opening_tag?(item) ⇒ Boolean

Returns:

  • (Boolean) 


200
201
202
 
# File 'lib/htmldiff.rb', line 200

def opening_tag?(item)
  item =~ %r!^\s*<[^>]+>\s*$!
end

#operations ⇒ Object 



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
 
# File 'lib/htmldiff.rb', line 46

def operations
  position_in_old = position_in_new = 0
  operations = []
  
  matches = matching_blocks
  # an empty match at the end forces the loop below to handle the unmatched tails
  # I'm sure it can be done more gracefully, but not at 23:52
  matches << Match.new(@old_words.length, @new_words.length, 0)
  
  matches.each_with_index do |match, i|
    match_starts_at_current_position_in_old = (position_in_old == match.start_in_old)
    match_starts_at_current_position_in_new = (position_in_new == match.start_in_new)
    
    action_upto_match_positions = 
      case [match_starts_at_current_position_in_old, match_starts_at_current_position_in_new]
      when [false, false]
        :replace
      when [true, false]
        :insert
      when [false, true]
        :delete
      else
        # this happens if the first few words are same in both versions
        :none
      end

    if action_upto_match_positions != :none
      operation_upto_match_positions = 
          Operation.new(action_upto_match_positions, 
              position_in_old, match.start_in_old, 
              position_in_new, match.start_in_new)
      operations << operation_upto_match_positions
    end
    if match.size != 0
      match_operation = Operation.new(:equal, 
          match.start_in_old, match.end_in_old, 
          match.start_in_new, match.end_in_new)
      operations << match_operation
    end

    position_in_old = match.end_in_old
    position_in_new = match.end_in_new
  end
  
  operations
end

#perform_operation(operation) ⇒ Object 



172
173
174
175
 
# File 'lib/htmldiff.rb', line 172

def perform_operation(operation)
  @operation = operation
  self.send operation.action, operation
end

#recursively_find_matching_blocks(start_in_old, end_in_old, start_in_new, end_in_new, matching_blocks) ⇒ Object 



99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/htmldiff.rb', line 99

def recursively_find_matching_blocks(start_in_old, end_in_old, start_in_new, end_in_new, matching_blocks)
  match = find_match(start_in_old, end_in_old, start_in_new, end_in_new)
  if match
    if start_in_old < match.start_in_old and start_in_new < match.start_in_new
      recursively_find_matching_blocks(
          start_in_old, match.start_in_old, start_in_new, match.start_in_new, matching_blocks) 
    end
    matching_blocks << match
    if match.end_in_old < end_in_old and match.end_in_new < end_in_new
      recursively_find_matching_blocks(
          match.end_in_old, end_in_old, match.end_in_new, end_in_new, matching_blocks)
    end
  end
end

#replace(operation) ⇒ Object 



177
178
179
180
 
# File 'lib/htmldiff.rb', line 177

def replace(operation)
  delete(operation, 'diffmod')
  insert(operation, 'diffmod')
end

#split_inputs_to_words ⇒ Object 



36
37
38
39
 
# File 'lib/htmldiff.rb', line 36

def split_inputs_to_words
  @old_words = convert_html_to_list_of_words(explode(@old_version))
  @new_words = convert_html_to_list_of_words(explode(@new_version))
end

#start_of_tag?(char) ⇒ Boolean

Returns:

  • (Boolean) 


275
276
277
 
# File 'lib/htmldiff.rb', line 275

def start_of_tag?(char)
  char == '<'
end

#tag?(item) ⇒ Boolean

Returns:

  • (Boolean) 


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

def tag?(item)
  opening_tag?(item) or closing_tag?(item)
end

#whitespace?(char) ⇒ Boolean

Returns:

  • (Boolean) 


279
280
281
 
# File 'lib/htmldiff.rb', line 279

def whitespace?(char)
  char =~ /\s/
end

#wrap_text(text, tagname, cssclass) ⇒ Object 



263
264
265
 
# File 'lib/htmldiff.rb', line 263

def wrap_text(text, tagname, cssclass)
  %(<#{tagname} class="#{cssclass}">#{text}</#{tagname}>)
end