Module: Diff::LCS::Internals

Defined in:
lib/diff/lcs/internals.rb

Overview

:nodoc:

Class Method Summary collapse

Class Method Details

.analyze_patchset(patchset, depth = 0) ⇒ Object

This method will analyze the provided patchset to provide a single-pass normalization (conversion of the array form of Diff::LCS::Change objects to the object form of same) and detection of whether the patchset represents changes to be made.


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
130
131
132
133
134
# File 'lib/diff/lcs/internals.rb', line 98

def analyze_patchset(patchset, depth = 0)
  fail 'Patchset too complex' if depth > 1

  has_changes = false
  new_patchset = []

  # Format:
  # [ # patchset
  #   # hunk (change)
  #   [ # hunk
  #     # change
  #   ]
  # ]

  patchset.each do |hunk|
    case hunk
    when Diff::LCS::Change
      has_changes ||= !hunk.unchanged?
      new_patchset << hunk
    when Array
      # Detect if the 'hunk' is actually an array-format change object.
      if Diff::LCS::Change.valid_action? hunk[0]
        hunk = Diff::LCS::Change.from_a(hunk)
        has_changes ||= !hunk.unchanged?
        new_patchset << hunk
      else
        with_changes, hunk = analyze_patchset(hunk, depth + 1)
        has_changes ||= with_changes
        new_patchset.concat(hunk)
      end
    else
      fail ArgumentError, "Cannot normalise a hunk of class #{hunk.class}."
    end
  end

  [has_changes, new_patchset]
end

.intuit_diff_direction(src, patchset, limit = nil) ⇒ Object

Examine the patchset and the source to see in which direction the patch should be applied.

WARNING: By default, this examines the whole patch, so this could take some time. This also works better with Diff::LCS::ContextChange or Diff::LCS::Change as its source, as an array will cause the creation of one of the above.


143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
# File 'lib/diff/lcs/internals.rb', line 143

def intuit_diff_direction(src, patchset, limit = nil)
  string = src.kind_of?(String)
  count = left_match = left_miss = right_match = right_miss = 0

  patchset.each do |change|
    count += 1

    case change
    when Diff::LCS::ContextChange
      le = string ? src[change.old_position, 1] : src[change.old_position]
      re = string ? src[change.new_position, 1] : src[change.new_position]

      case change.action
      when '-' # Remove details from the old string
        if le == change.old_element
          left_match += 1
        else
          left_miss += 1
        end
      when '+'
        if re == change.new_element
          right_match += 1
        else
          right_miss += 1
        end
      when '='
        left_miss += 1 if le != change.old_element
        right_miss += 1 if re != change.new_element
      when '!'
        if le == change.old_element
          left_match += 1
        elsif re == change.new_element
          right_match += 1
        else
          left_miss += 1
          right_miss += 1
        end
      end
    when Diff::LCS::Change
      # With a simplistic change, we can't tell the difference between
      # the left and right on '!' actions, so we ignore those. On '='
      # actions, if there's a miss, we miss both left and right.
      element = string ? src[change.position, 1] : src[change.position]

      case change.action
      when '-'
        if element == change.element
          left_match += 1
        else
          left_miss += 1
        end
      when '+'
        if element == change.element
          right_match += 1
        else
          right_miss += 1
        end
      when '='
        if element != change.element
          left_miss += 1
          right_miss += 1
        end
      end
    end

    break if !limit.nil? && (count > limit)
  end

  no_left = left_match.zero? && left_miss.positive?
  no_right = right_match.zero? && right_miss.positive?

  case [no_left, no_right]
  when [false, true]
    :patch
  when [true, false]
    :unpatch
  else
    case left_match <=> right_match
    when 1
      if left_miss.zero?
        :patch
      else
        :unpatch
      end
    when -1
      if right_miss.zero?
        :unpatch
      else
        :patch
      end
    else
      fail "The provided patchset does not appear to apply to the provided \
enumerable as either source or destination value."
    end
  end
end

.inverse_vector(a, vector) ⇒ Object

If vector maps the matching elements of another collection onto this Enumerable, compute the inverse of vector that maps this Enumerable onto the collection. (Currently unused.)


282
283
284
285
286
287
288
# File 'lib/diff/lcs/internals.rb', line 282

def inverse_vector(a, vector)
  inverse = a.dup
  (0...vector.size).each do |i|
    inverse[vector[i]] = i unless vector[i].nil?
  end
  inverse
end

.lcs(a, b) ⇒ Object

Compute the longest common subsequence between the sequenced Enumerables a and b. The result is an array whose contents is such that

result = Diff::LCS::Internals.lcs(a, b)
result.each_with_index do |e, i|
  assert_equal(a[i], b[e]) unless e.nil?
end

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
# File 'lib/diff/lcs/internals.rb', line 41

def lcs(a, b)
  a_start = b_start = 0
  a_finish = a.size - 1
  b_finish = b.size - 1
  vector = []

  # Prune off any common elements at the beginning...
  while (a_start <= a_finish) and (b_start <= b_finish) and (a[a_start] == b[b_start])
    vector[a_start] = b_start
    a_start += 1
    b_start += 1
  end
  b_start = a_start

  # Now the end...
  while (a_start <= a_finish) and (b_start <= b_finish) and (a[a_finish] == b[b_finish])
    vector[a_finish] = b_finish
    a_finish -= 1
    b_finish -= 1
  end

  # Now, compute the equivalence classes of positions of elements.
  b_matches = position_hash(b, b_start..b_finish)

  thresh = []
  links  = []
  string = a.kind_of?(String)

  (a_start..a_finish).each do |i|
    ai = string ? a[i, 1] : a[i]
    bm = b_matches[ai]
    k = nil
    bm.reverse_each do |j|
      if k and (thresh[k] > j) and (thresh[k - 1] < j)
        thresh[k] = j
      else
        k = replace_next_larger(thresh, j, k)
      end
      links[k] = [k.positive? ? links[k - 1] : nil, i, j] unless k.nil?
    end
  end

  unless thresh.empty?
    link = links[thresh.size - 1]
    until link.nil?
      vector[link[1]] = link[2]
      link = link[0]
    end
  end

  vector
end

.position_hash(enum, interval) ⇒ Object

Returns a hash mapping each element of an Enumerable to the set of positions it occupies in the Enumerable, optionally restricted to the elements specified in the range of indexes specified by interval.


294
295
296
297
298
299
300
301
302
# File 'lib/diff/lcs/internals.rb', line 294

def position_hash(enum, interval)
  string = enum.kind_of?(String)
  hash = Hash.new { |h, k| h[k] = [] }
  interval.each do |i|
    k = string ? enum[i, 1] : enum[i]
    hash[k] << i
  end
  hash
end

.replace_next_larger(enum, value, last_index = nil) ⇒ Object

Find the place at which value would normally be inserted into the Enumerable. If that place is already occupied by value, do nothing and return nil. If the place does not exist (i.e., it is off the end of the Enumerable), add it to the end. Otherwise, replace the element at that point with value. It is assumed that the Enumerable's values are numeric.

This operation preserves the sort order.


248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
# File 'lib/diff/lcs/internals.rb', line 248

def replace_next_larger(enum, value, last_index = nil)
  # Off the end?
  if enum.empty? or (value > enum[-1])
    enum << value
    return enum.size - 1
  end

  # Binary search for the insertion point
  last_index ||= enum.size
  first_index = 0
  while first_index <= last_index
    i = (first_index + last_index) >> 1

    found = enum[i]

    return nil if value == found

    if value > found
      first_index = i + 1
    else
      last_index = i - 1
    end
  end

  # The insertion point is in first_index; overwrite the next larger
  # value.
  enum[first_index] = value
  first_index
end