|
250
|
1 |
#!/usr/bin/python2.4 |
|
|
2 |
|
|
|
3 |
"""Diff Match and Patch |
|
|
4 |
|
|
|
5 |
Copyright 2006 Google Inc. |
|
|
6 |
http://code.google.com/p/google-diff-match-patch/ |
|
|
7 |
|
|
|
8 |
Licensed under the Apache License, Version 2.0 (the "License"); |
|
|
9 |
you may not use this file except in compliance with the License. |
|
|
10 |
You may obtain a copy of the License at |
|
|
11 |
|
|
|
12 |
http://www.apache.org/licenses/LICENSE-2.0 |
|
|
13 |
|
|
|
14 |
Unless required by applicable law or agreed to in writing, software |
|
|
15 |
distributed under the License is distributed on an "AS IS" BASIS, |
|
|
16 |
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
|
|
17 |
See the License for the specific language governing permissions and |
|
|
18 |
limitations under the License. |
|
|
19 |
""" |
|
|
20 |
|
|
|
21 |
"""Functions for diff, match and patch. |
|
|
22 |
|
|
|
23 |
Computes the difference between two texts to create a patch. |
|
|
24 |
Applies the patch onto another text, allowing for errors. |
|
|
25 |
""" |
|
|
26 |
|
|
|
27 |
__author__ = 'fraser@google.com (Neil Fraser)' |
|
|
28 |
|
|
|
29 |
import math |
|
|
30 |
import time |
|
|
31 |
import urllib |
|
|
32 |
import re |
|
|
33 |
|
|
|
34 |
class diff_match_patch: |
|
|
35 |
"""Class containing the diff, match and patch methods. |
|
|
36 |
|
|
|
37 |
Also contains the behaviour settings. |
|
|
38 |
""" |
|
|
39 |
|
|
|
40 |
def __init__(self): |
|
|
41 |
"""Inits a diff_match_patch object with default settings. |
|
|
42 |
Redefine these in your program to override the defaults. |
|
|
43 |
""" |
|
|
44 |
|
|
|
45 |
# Number of seconds to map a diff before giving up (0 for infinity). |
|
|
46 |
self.Diff_Timeout = 1.0 |
|
|
47 |
# Cost of an empty edit operation in terms of edit characters. |
|
|
48 |
self.Diff_EditCost = 4 |
|
|
49 |
# The size beyond which the double-ended diff activates. |
|
|
50 |
# Double-ending is twice as fast, but less accurate. |
|
|
51 |
self.Diff_DualThreshold = 32 |
|
|
52 |
# At what point is no match declared (0.0 = perfection, 1.0 = very loose). |
|
|
53 |
self.Match_Threshold = 0.5 |
|
|
54 |
# How far to search for a match (0 = exact location, 1000+ = broad match). |
|
|
55 |
# A match this many characters away from the expected location will add |
|
|
56 |
# 1.0 to the score (0.0 is a perfect match). |
|
|
57 |
self.Match_Distance = 1000 |
|
|
58 |
# When deleting a large block of text (over ~64 characters), how close does |
|
|
59 |
# the contents have to match the expected contents. (0.0 = perfection, |
|
|
60 |
# 1.0 = very loose). Note that Match_Threshold controls how closely the |
|
|
61 |
# end points of a delete need to match. |
|
|
62 |
self.Patch_DeleteThreshold = 0.5 |
|
|
63 |
# Chunk size for context length. |
|
|
64 |
self.Patch_Margin = 4 |
|
|
65 |
|
|
|
66 |
# How many bits in a number? |
|
|
67 |
# Python has no maximum, thus to disable patch splitting set to 0. |
|
|
68 |
# However to avoid long patches in certain pathological cases, use 32. |
|
|
69 |
# Multiple short patches (using native ints) are much faster than long ones. |
|
|
70 |
self.Match_MaxBits = 32 |
|
|
71 |
|
|
|
72 |
# DIFF FUNCTIONS |
|
|
73 |
|
|
|
74 |
# The data structure representing a diff is an array of tuples: |
|
|
75 |
# [(DIFF_DELETE, "Hello"), (DIFF_INSERT, "Goodbye"), (DIFF_EQUAL, " world.")] |
|
|
76 |
# which means: delete "Hello", add "Goodbye" and keep " world." |
|
|
77 |
DIFF_DELETE = -1 |
|
|
78 |
DIFF_INSERT = 1 |
|
|
79 |
DIFF_EQUAL = 0 |
|
|
80 |
|
|
|
81 |
def diff_main(self, text1, text2, checklines=True): |
|
|
82 |
"""Find the differences between two texts. Simplifies the problem by |
|
|
83 |
stripping any common prefix or suffix off the texts before diffing. |
|
|
84 |
|
|
|
85 |
Args: |
|
|
86 |
text1: Old string to be diffed. |
|
|
87 |
text2: New string to be diffed. |
|
|
88 |
checklines: Optional speedup flag. If present and false, then don't run |
|
|
89 |
a line-level diff first to identify the changed areas. |
|
|
90 |
Defaults to true, which does a faster, slightly less optimal diff. |
|
|
91 |
|
|
|
92 |
Returns: |
|
|
93 |
Array of changes. |
|
|
94 |
""" |
|
|
95 |
|
|
|
96 |
# Check for equality (speedup) |
|
|
97 |
if text1 == text2: |
|
|
98 |
return [(self.DIFF_EQUAL, text1)] |
|
|
99 |
|
|
|
100 |
# Trim off common prefix (speedup) |
|
|
101 |
commonlength = self.diff_commonPrefix(text1, text2) |
|
|
102 |
commonprefix = text1[:commonlength] |
|
|
103 |
text1 = text1[commonlength:] |
|
|
104 |
text2 = text2[commonlength:] |
|
|
105 |
|
|
|
106 |
# Trim off common suffix (speedup) |
|
|
107 |
commonlength = self.diff_commonSuffix(text1, text2) |
|
|
108 |
if commonlength == 0: |
|
|
109 |
commonsuffix = '' |
|
|
110 |
else: |
|
|
111 |
commonsuffix = text1[-commonlength:] |
|
|
112 |
text1 = text1[:-commonlength] |
|
|
113 |
text2 = text2[:-commonlength] |
|
|
114 |
|
|
|
115 |
# Compute the diff on the middle block |
|
|
116 |
diffs = self.diff_compute(text1, text2, checklines) |
|
|
117 |
|
|
|
118 |
# Restore the prefix and suffix |
|
|
119 |
if commonprefix: |
|
|
120 |
diffs[:0] = [(self.DIFF_EQUAL, commonprefix)] |
|
|
121 |
if commonsuffix: |
|
|
122 |
diffs.append((self.DIFF_EQUAL, commonsuffix)) |
|
|
123 |
self.diff_cleanupMerge(diffs) |
|
|
124 |
return diffs |
|
|
125 |
|
|
|
126 |
def diff_compute(self, text1, text2, checklines): |
|
|
127 |
"""Find the differences between two texts. Assumes that the texts do not |
|
|
128 |
have any common prefix or suffix. |
|
|
129 |
|
|
|
130 |
Args: |
|
|
131 |
text1: Old string to be diffed. |
|
|
132 |
text2: New string to be diffed. |
|
|
133 |
checklines: Speedup flag. If false, then don't run a line-level diff |
|
|
134 |
first to identify the changed areas. |
|
|
135 |
If true, then run a faster, slightly less optimal diff. |
|
|
136 |
|
|
|
137 |
Returns: |
|
|
138 |
Array of changes. |
|
|
139 |
""" |
|
|
140 |
if not text1: |
|
|
141 |
# Just add some text (speedup) |
|
|
142 |
return [(self.DIFF_INSERT, text2)] |
|
|
143 |
|
|
|
144 |
if not text2: |
|
|
145 |
# Just delete some text (speedup) |
|
|
146 |
return [(self.DIFF_DELETE, text1)] |
|
|
147 |
|
|
|
148 |
if len(text1) > len(text2): |
|
|
149 |
(longtext, shorttext) = (text1, text2) |
|
|
150 |
else: |
|
|
151 |
(shorttext, longtext) = (text1, text2) |
|
|
152 |
i = longtext.find(shorttext) |
|
|
153 |
if i != -1: |
|
|
154 |
# Shorter text is inside the longer text (speedup) |
|
|
155 |
diffs = [(self.DIFF_INSERT, longtext[:i]), (self.DIFF_EQUAL, shorttext), |
|
|
156 |
(self.DIFF_INSERT, longtext[i + len(shorttext):])] |
|
|
157 |
# Swap insertions for deletions if diff is reversed. |
|
|
158 |
if len(text1) > len(text2): |
|
|
159 |
diffs[0] = (self.DIFF_DELETE, diffs[0][1]) |
|
|
160 |
diffs[2] = (self.DIFF_DELETE, diffs[2][1]) |
|
|
161 |
return diffs |
|
|
162 |
longtext = shorttext = None # Garbage collect. |
|
|
163 |
|
|
|
164 |
# Check to see if the problem can be split in two. |
|
|
165 |
hm = self.diff_halfMatch(text1, text2) |
|
|
166 |
if hm: |
|
|
167 |
# A half-match was found, sort out the return data. |
|
|
168 |
(text1_a, text1_b, text2_a, text2_b, mid_common) = hm |
|
|
169 |
# Send both pairs off for separate processing. |
|
|
170 |
diffs_a = self.diff_main(text1_a, text2_a, checklines) |
|
|
171 |
diffs_b = self.diff_main(text1_b, text2_b, checklines) |
|
|
172 |
# Merge the results. |
|
|
173 |
return diffs_a + [(self.DIFF_EQUAL, mid_common)] + diffs_b |
|
|
174 |
|
|
|
175 |
# Perform a real diff. |
|
|
176 |
if checklines and (len(text1) < 100 or len(text2) < 100): |
|
|
177 |
checklines = False # Too trivial for the overhead. |
|
|
178 |
if checklines: |
|
|
179 |
# Scan the text on a line-by-line basis first. |
|
|
180 |
(text1, text2, linearray) = self.diff_linesToChars(text1, text2) |
|
|
181 |
|
|
|
182 |
diffs = self.diff_map(text1, text2) |
|
|
183 |
if not diffs: # No acceptable result. |
|
|
184 |
diffs = [(self.DIFF_DELETE, text1), (self.DIFF_INSERT, text2)] |
|
|
185 |
if checklines: |
|
|
186 |
# Convert the diff back to original text. |
|
|
187 |
self.diff_charsToLines(diffs, linearray) |
|
|
188 |
# Eliminate freak matches (e.g. blank lines) |
|
|
189 |
self.diff_cleanupSemantic(diffs) |
|
|
190 |
|
|
|
191 |
# Rediff any replacement blocks, this time character-by-character. |
|
|
192 |
# Add a dummy entry at the end. |
|
|
193 |
diffs.append((self.DIFF_EQUAL, '')) |
|
|
194 |
pointer = 0 |
|
|
195 |
count_delete = 0 |
|
|
196 |
count_insert = 0 |
|
|
197 |
text_delete = '' |
|
|
198 |
text_insert = '' |
|
|
199 |
while pointer < len(diffs): |
|
|
200 |
if diffs[pointer][0] == self.DIFF_INSERT: |
|
|
201 |
count_insert += 1 |
|
|
202 |
text_insert += diffs[pointer][1] |
|
|
203 |
elif diffs[pointer][0] == self.DIFF_DELETE: |
|
|
204 |
count_delete += 1 |
|
|
205 |
text_delete += diffs[pointer][1] |
|
|
206 |
elif diffs[pointer][0] == self.DIFF_EQUAL: |
|
|
207 |
# Upon reaching an equality, check for prior redundancies. |
|
|
208 |
if count_delete >= 1 and count_insert >= 1: |
|
|
209 |
# Delete the offending records and add the merged ones. |
|
|
210 |
a = self.diff_main(text_delete, text_insert, False) |
|
|
211 |
diffs[pointer - count_delete - count_insert : pointer] = a |
|
|
212 |
pointer = pointer - count_delete - count_insert + len(a) |
|
|
213 |
count_insert = 0 |
|
|
214 |
count_delete = 0 |
|
|
215 |
text_delete = '' |
|
|
216 |
text_insert = '' |
|
|
217 |
|
|
|
218 |
pointer += 1 |
|
|
219 |
|
|
|
220 |
diffs.pop() # Remove the dummy entry at the end. |
|
|
221 |
return diffs |
|
|
222 |
|
|
|
223 |
def diff_linesToChars(self, text1, text2): |
|
|
224 |
"""Split two texts into an array of strings. Reduce the texts to a string |
|
|
225 |
of hashes where each Unicode character represents one line. |
|
|
226 |
|
|
|
227 |
Args: |
|
|
228 |
text1: First string. |
|
|
229 |
text2: Second string. |
|
|
230 |
|
|
|
231 |
Returns: |
|
|
232 |
Three element tuple, containing the encoded text1, the encoded text2 and |
|
|
233 |
the array of unique strings. The zeroth element of the array of unique |
|
|
234 |
strings is intentionally blank. |
|
|
235 |
""" |
|
|
236 |
lineArray = [] # e.g. lineArray[4] == "Hello\n" |
|
|
237 |
lineHash = {} # e.g. lineHash["Hello\n"] == 4 |
|
|
238 |
|
|
|
239 |
# "\x00" is a valid character, but various debuggers don't like it. |
|
|
240 |
# So we'll insert a junk entry to avoid generating a null character. |
|
|
241 |
lineArray.append('') |
|
|
242 |
|
|
|
243 |
def diff_linesToCharsMunge(text): |
|
|
244 |
"""Split a text into an array of strings. Reduce the texts to a string |
|
|
245 |
of hashes where each Unicode character represents one line. |
|
|
246 |
Modifies linearray and linehash through being a closure. |
|
|
247 |
|
|
|
248 |
Args: |
|
|
249 |
text: String to encode. |
|
|
250 |
|
|
|
251 |
Returns: |
|
|
252 |
Encoded string. |
|
|
253 |
""" |
|
|
254 |
chars = [] |
|
|
255 |
# Walk the text, pulling out a substring for each line. |
|
|
256 |
# text.split('\n') would would temporarily double our memory footprint. |
|
|
257 |
# Modifying text would create many large strings to garbage collect. |
|
|
258 |
lineStart = 0 |
|
|
259 |
lineEnd = -1 |
|
|
260 |
while lineEnd < len(text) - 1: |
|
|
261 |
lineEnd = text.find('\n', lineStart) |
|
|
262 |
if lineEnd == -1: |
|
|
263 |
lineEnd = len(text) - 1 |
|
|
264 |
line = text[lineStart:lineEnd + 1] |
|
|
265 |
lineStart = lineEnd + 1 |
|
|
266 |
|
|
|
267 |
if line in lineHash: |
|
|
268 |
chars.append(unichr(lineHash[line])) |
|
|
269 |
else: |
|
|
270 |
lineArray.append(line) |
|
|
271 |
lineHash[line] = len(lineArray) - 1 |
|
|
272 |
chars.append(unichr(len(lineArray) - 1)) |
|
|
273 |
return "".join(chars) |
|
|
274 |
|
|
|
275 |
chars1 = diff_linesToCharsMunge(text1) |
|
|
276 |
chars2 = diff_linesToCharsMunge(text2) |
|
|
277 |
return (chars1, chars2, lineArray) |
|
|
278 |
|
|
|
279 |
def diff_charsToLines(self, diffs, lineArray): |
|
|
280 |
"""Rehydrate the text in a diff from a string of line hashes to real lines |
|
|
281 |
of text. |
|
|
282 |
|
|
|
283 |
Args: |
|
|
284 |
diffs: Array of diff tuples. |
|
|
285 |
lineArray: Array of unique strings. |
|
|
286 |
""" |
|
|
287 |
for x in xrange(len(diffs)): |
|
|
288 |
text = [] |
|
|
289 |
for char in diffs[x][1]: |
|
|
290 |
text.append(lineArray[ord(char)]) |
|
|
291 |
diffs[x] = (diffs[x][0], "".join(text)) |
|
|
292 |
|
|
|
293 |
def diff_map(self, text1, text2): |
|
|
294 |
"""Explore the intersection points between the two texts. |
|
|
295 |
|
|
|
296 |
Args: |
|
|
297 |
text1: Old string to be diffed. |
|
|
298 |
text2: New string to be diffed. |
|
|
299 |
|
|
|
300 |
Returns: |
|
|
301 |
Array of diff tuples or None if no diff available. |
|
|
302 |
""" |
|
|
303 |
|
|
|
304 |
# Unlike in most languages, Python counts time in seconds. |
|
|
305 |
s_end = time.time() + self.Diff_Timeout # Don't run for too long. |
|
|
306 |
# Cache the text lengths to prevent multiple calls. |
|
|
307 |
text1_length = len(text1) |
|
|
308 |
text2_length = len(text2) |
|
|
309 |
max_d = text1_length + text2_length - 1 |
|
|
310 |
doubleEnd = self.Diff_DualThreshold * 2 < max_d |
|
|
311 |
v_map1 = [] |
|
|
312 |
v_map2 = [] |
|
|
313 |
v1 = {} |
|
|
314 |
v2 = {} |
|
|
315 |
v1[1] = 0 |
|
|
316 |
v2[1] = 0 |
|
|
317 |
footsteps = {} |
|
|
318 |
done = False |
|
|
319 |
# If the total number of characters is odd, then the front path will |
|
|
320 |
# collide with the reverse path. |
|
|
321 |
front = (text1_length + text2_length) % 2 |
|
|
322 |
for d in xrange(max_d): |
|
|
323 |
# Bail out if timeout reached. |
|
|
324 |
if self.Diff_Timeout > 0 and time.time() > s_end: |
|
|
325 |
return None |
|
|
326 |
|
|
|
327 |
# Walk the front path one step. |
|
|
328 |
v_map1.append({}) |
|
|
329 |
for k in xrange(-d, d + 1, 2): |
|
|
330 |
if k == -d or k != d and v1[k - 1] < v1[k + 1]: |
|
|
331 |
x = v1[k + 1] |
|
|
332 |
else: |
|
|
333 |
x = v1[k - 1] + 1 |
|
|
334 |
y = x - k |
|
|
335 |
if doubleEnd: |
|
|
336 |
footstep = (x, y) |
|
|
337 |
if front and footstep in footsteps: |
|
|
338 |
done = True |
|
|
339 |
if not front: |
|
|
340 |
footsteps[footstep] = d |
|
|
341 |
|
|
|
342 |
while (not done and x < text1_length and y < text2_length and |
|
|
343 |
text1[x] == text2[y]): |
|
|
344 |
x += 1 |
|
|
345 |
y += 1 |
|
|
346 |
if doubleEnd: |
|
|
347 |
footstep = (x, y) |
|
|
348 |
if front and footstep in footsteps: |
|
|
349 |
done = True |
|
|
350 |
if not front: |
|
|
351 |
footsteps[footstep] = d |
|
|
352 |
|
|
|
353 |
v1[k] = x |
|
|
354 |
v_map1[d][(x, y)] = True |
|
|
355 |
if x == text1_length and y == text2_length: |
|
|
356 |
# Reached the end in single-path mode. |
|
|
357 |
return self.diff_path1(v_map1, text1, text2) |
|
|
358 |
elif done: |
|
|
359 |
# Front path ran over reverse path. |
|
|
360 |
v_map2 = v_map2[:footsteps[footstep] + 1] |
|
|
361 |
a = self.diff_path1(v_map1, text1[:x], text2[:y]) |
|
|
362 |
b = self.diff_path2(v_map2, text1[x:], text2[y:]) |
|
|
363 |
return a + b |
|
|
364 |
|
|
|
365 |
if doubleEnd: |
|
|
366 |
# Walk the reverse path one step. |
|
|
367 |
v_map2.append({}) |
|
|
368 |
for k in xrange(-d, d + 1, 2): |
|
|
369 |
if k == -d or k != d and v2[k - 1] < v2[k + 1]: |
|
|
370 |
x = v2[k + 1] |
|
|
371 |
else: |
|
|
372 |
x = v2[k - 1] + 1 |
|
|
373 |
y = x - k |
|
|
374 |
footstep = (text1_length - x, text2_length - y) |
|
|
375 |
if not front and footstep in footsteps: |
|
|
376 |
done = True |
|
|
377 |
if front: |
|
|
378 |
footsteps[footstep] = d |
|
|
379 |
while (not done and x < text1_length and y < text2_length and |
|
|
380 |
text1[-x - 1] == text2[-y - 1]): |
|
|
381 |
x += 1 |
|
|
382 |
y += 1 |
|
|
383 |
footstep = (text1_length - x, text2_length - y) |
|
|
384 |
if not front and footstep in footsteps: |
|
|
385 |
done = True |
|
|
386 |
if front: |
|
|
387 |
footsteps[footstep] = d |
|
|
388 |
|
|
|
389 |
v2[k] = x |
|
|
390 |
v_map2[d][(x, y)] = True |
|
|
391 |
if done: |
|
|
392 |
# Reverse path ran over front path. |
|
|
393 |
v_map1 = v_map1[:footsteps[footstep] + 1] |
|
|
394 |
a = self.diff_path1(v_map1, text1[:text1_length - x], |
|
|
395 |
text2[:text2_length - y]) |
|
|
396 |
b = self.diff_path2(v_map2, text1[text1_length - x:], |
|
|
397 |
text2[text2_length - y:]) |
|
|
398 |
return a + b |
|
|
399 |
|
|
|
400 |
# Number of diffs equals number of characters, no commonality at all. |
|
|
401 |
return None |
|
|
402 |
|
|
|
403 |
def diff_path1(self, v_map, text1, text2): |
|
|
404 |
"""Work from the middle back to the start to determine the path. |
|
|
405 |
|
|
|
406 |
Args: |
|
|
407 |
v_map: Array of paths. |
|
|
408 |
text1: Old string fragment to be diffed. |
|
|
409 |
text2: New string fragment to be diffed. |
|
|
410 |
|
|
|
411 |
Returns: |
|
|
412 |
Array of diff tuples. |
|
|
413 |
""" |
|
|
414 |
path = [] |
|
|
415 |
x = len(text1) |
|
|
416 |
y = len(text2) |
|
|
417 |
last_op = None |
|
|
418 |
for d in xrange(len(v_map) - 2, -1, -1): |
|
|
419 |
while True: |
|
|
420 |
if (x - 1, y) in v_map[d]: |
|
|
421 |
x -= 1 |
|
|
422 |
if last_op == self.DIFF_DELETE: |
|
|
423 |
path[0] = (self.DIFF_DELETE, text1[x] + path[0][1]) |
|
|
424 |
else: |
|
|
425 |
path[:0] = [(self.DIFF_DELETE, text1[x])] |
|
|
426 |
last_op = self.DIFF_DELETE |
|
|
427 |
break |
|
|
428 |
elif (x, y - 1) in v_map[d]: |
|
|
429 |
y -= 1 |
|
|
430 |
if last_op == self.DIFF_INSERT: |
|
|
431 |
path[0] = (self.DIFF_INSERT, text2[y] + path[0][1]) |
|
|
432 |
else: |
|
|
433 |
path[:0] = [(self.DIFF_INSERT, text2[y])] |
|
|
434 |
last_op = self.DIFF_INSERT |
|
|
435 |
break |
|
|
436 |
else: |
|
|
437 |
x -= 1 |
|
|
438 |
y -= 1 |
|
|
439 |
assert text1[x] == text2[y], ("No diagonal. " + |
|
|
440 |
"Can't happen. (diff_path1)") |
|
|
441 |
if last_op == self.DIFF_EQUAL: |
|
|
442 |
path[0] = (self.DIFF_EQUAL, text1[x] + path[0][1]) |
|
|
443 |
else: |
|
|
444 |
path[:0] = [(self.DIFF_EQUAL, text1[x])] |
|
|
445 |
last_op = self.DIFF_EQUAL |
|
|
446 |
return path |
|
|
447 |
|
|
|
448 |
def diff_path2(self, v_map, text1, text2): |
|
|
449 |
"""Work from the middle back to the end to determine the path. |
|
|
450 |
|
|
|
451 |
Args: |
|
|
452 |
v_map: Array of paths. |
|
|
453 |
text1: Old string fragment to be diffed. |
|
|
454 |
text2: New string fragment to be diffed. |
|
|
455 |
|
|
|
456 |
Returns: |
|
|
457 |
Array of diff tuples. |
|
|
458 |
""" |
|
|
459 |
path = [] |
|
|
460 |
x = len(text1) |
|
|
461 |
y = len(text2) |
|
|
462 |
last_op = None |
|
|
463 |
for d in xrange(len(v_map) - 2, -1, -1): |
|
|
464 |
while True: |
|
|
465 |
if (x - 1, y) in v_map[d]: |
|
|
466 |
x -= 1 |
|
|
467 |
if last_op == self.DIFF_DELETE: |
|
|
468 |
path[-1] = (self.DIFF_DELETE, path[-1][1] + text1[-x - 1]) |
|
|
469 |
else: |
|
|
470 |
path.append((self.DIFF_DELETE, text1[-x - 1])) |
|
|
471 |
last_op = self.DIFF_DELETE |
|
|
472 |
break |
|
|
473 |
elif (x, y - 1) in v_map[d]: |
|
|
474 |
y -= 1 |
|
|
475 |
if last_op == self.DIFF_INSERT: |
|
|
476 |
path[-1] = (self.DIFF_INSERT, path[-1][1] + text2[-y - 1]) |
|
|
477 |
else: |
|
|
478 |
path.append((self.DIFF_INSERT, text2[-y - 1])) |
|
|
479 |
last_op = self.DIFF_INSERT |
|
|
480 |
break |
|
|
481 |
else: |
|
|
482 |
x -= 1 |
|
|
483 |
y -= 1 |
|
|
484 |
assert text1[-x - 1] == text2[-y - 1], ("No diagonal. " + |
|
|
485 |
"Can't happen. (diff_path2)") |
|
|
486 |
if last_op == self.DIFF_EQUAL: |
|
|
487 |
path[-1] = (self.DIFF_EQUAL, path[-1][1] + text1[-x - 1]) |
|
|
488 |
else: |
|
|
489 |
path.append((self.DIFF_EQUAL, text1[-x - 1])) |
|
|
490 |
last_op = self.DIFF_EQUAL |
|
|
491 |
return path |
|
|
492 |
|
|
|
493 |
def diff_commonPrefix(self, text1, text2): |
|
|
494 |
"""Determine the common prefix of two strings. |
|
|
495 |
|
|
|
496 |
Args: |
|
|
497 |
text1: First string. |
|
|
498 |
text2: Second string. |
|
|
499 |
|
|
|
500 |
Returns: |
|
|
501 |
The number of characters common to the start of each string. |
|
|
502 |
""" |
|
|
503 |
# Quick check for common null cases. |
|
|
504 |
if not text1 or not text2 or text1[0] != text2[0]: |
|
|
505 |
return 0 |
|
|
506 |
# Binary search. |
|
|
507 |
# Performance analysis: http://neil.fraser.name/news/2007/10/09/ |
|
|
508 |
pointermin = 0 |
|
|
509 |
pointermax = min(len(text1), len(text2)) |
|
|
510 |
pointermid = pointermax |
|
|
511 |
pointerstart = 0 |
|
|
512 |
while pointermin < pointermid: |
|
|
513 |
if text1[pointerstart:pointermid] == text2[pointerstart:pointermid]: |
|
|
514 |
pointermin = pointermid |
|
|
515 |
pointerstart = pointermin |
|
|
516 |
else: |
|
|
517 |
pointermax = pointermid |
|
|
518 |
pointermid = int((pointermax - pointermin) / 2 + pointermin) |
|
|
519 |
return pointermid |
|
|
520 |
|
|
|
521 |
def diff_commonSuffix(self, text1, text2): |
|
|
522 |
"""Determine the common suffix of two strings. |
|
|
523 |
|
|
|
524 |
Args: |
|
|
525 |
text1: First string. |
|
|
526 |
text2: Second string. |
|
|
527 |
|
|
|
528 |
Returns: |
|
|
529 |
The number of characters common to the end of each string. |
|
|
530 |
""" |
|
|
531 |
# Quick check for common null cases. |
|
|
532 |
if not text1 or not text2 or text1[-1] != text2[-1]: |
|
|
533 |
return 0 |
|
|
534 |
# Binary search. |
|
|
535 |
# Performance analysis: http://neil.fraser.name/news/2007/10/09/ |
|
|
536 |
pointermin = 0 |
|
|
537 |
pointermax = min(len(text1), len(text2)) |
|
|
538 |
pointermid = pointermax |
|
|
539 |
pointerend = 0 |
|
|
540 |
while pointermin < pointermid: |
|
|
541 |
if (text1[-pointermid:len(text1) - pointerend] == |
|
|
542 |
text2[-pointermid:len(text2) - pointerend]): |
|
|
543 |
pointermin = pointermid |
|
|
544 |
pointerend = pointermin |
|
|
545 |
else: |
|
|
546 |
pointermax = pointermid |
|
|
547 |
pointermid = int((pointermax - pointermin) / 2 + pointermin) |
|
|
548 |
return pointermid |
|
|
549 |
|
|
|
550 |
def diff_halfMatch(self, text1, text2): |
|
|
551 |
"""Do the two texts share a substring which is at least half the length of |
|
|
552 |
the longer text? |
|
|
553 |
|
|
|
554 |
Args: |
|
|
555 |
text1: First string. |
|
|
556 |
text2: Second string. |
|
|
557 |
|
|
|
558 |
Returns: |
|
|
559 |
Five element Array, containing the prefix of text1, the suffix of text1, |
|
|
560 |
the prefix of text2, the suffix of text2 and the common middle. Or None |
|
|
561 |
if there was no match. |
|
|
562 |
""" |
|
|
563 |
if len(text1) > len(text2): |
|
|
564 |
(longtext, shorttext) = (text1, text2) |
|
|
565 |
else: |
|
|
566 |
(shorttext, longtext) = (text1, text2) |
|
|
567 |
if len(longtext) < 10 or len(shorttext) < 1: |
|
|
568 |
return None # Pointless. |
|
|
569 |
|
|
|
570 |
def diff_halfMatchI(longtext, shorttext, i): |
|
|
571 |
"""Does a substring of shorttext exist within longtext such that the |
|
|
572 |
substring is at least half the length of longtext? |
|
|
573 |
Closure, but does not reference any external variables. |
|
|
574 |
|
|
|
575 |
Args: |
|
|
576 |
longtext: Longer string. |
|
|
577 |
shorttext: Shorter string. |
|
|
578 |
i: Start index of quarter length substring within longtext. |
|
|
579 |
|
|
|
580 |
Returns: |
|
|
581 |
Five element Array, containing the prefix of longtext, the suffix of |
|
|
582 |
longtext, the prefix of shorttext, the suffix of shorttext and the |
|
|
583 |
common middle. Or None if there was no match. |
|
|
584 |
""" |
|
|
585 |
seed = longtext[i:i + len(longtext) / 4] |
|
|
586 |
best_common = '' |
|
|
587 |
j = shorttext.find(seed) |
|
|
588 |
while j != -1: |
|
|
589 |
prefixLength = self.diff_commonPrefix(longtext[i:], shorttext[j:]) |
|
|
590 |
suffixLength = self.diff_commonSuffix(longtext[:i], shorttext[:j]) |
|
|
591 |
if len(best_common) < suffixLength + prefixLength: |
|
|
592 |
best_common = (shorttext[j - suffixLength:j] + |
|
|
593 |
shorttext[j:j + prefixLength]) |
|
|
594 |
best_longtext_a = longtext[:i - suffixLength] |
|
|
595 |
best_longtext_b = longtext[i + prefixLength:] |
|
|
596 |
best_shorttext_a = shorttext[:j - suffixLength] |
|
|
597 |
best_shorttext_b = shorttext[j + prefixLength:] |
|
|
598 |
j = shorttext.find(seed, j + 1) |
|
|
599 |
|
|
|
600 |
if len(best_common) >= len(longtext) / 2: |
|
|
601 |
return (best_longtext_a, best_longtext_b, |
|
|
602 |
best_shorttext_a, best_shorttext_b, best_common) |
|
|
603 |
else: |
|
|
604 |
return None |
|
|
605 |
|
|
|
606 |
# First check if the second quarter is the seed for a half-match. |
|
|
607 |
hm1 = diff_halfMatchI(longtext, shorttext, (len(longtext) + 3) / 4) |
|
|
608 |
# Check again based on the third quarter. |
|
|
609 |
hm2 = diff_halfMatchI(longtext, shorttext, (len(longtext) + 1) / 2) |
|
|
610 |
if not hm1 and not hm2: |
|
|
611 |
return None |
|
|
612 |
elif not hm2: |
|
|
613 |
hm = hm1 |
|
|
614 |
elif not hm1: |
|
|
615 |
hm = hm2 |
|
|
616 |
else: |
|
|
617 |
# Both matched. Select the longest. |
|
|
618 |
if len(hm1[4]) > len(hm2[4]): |
|
|
619 |
hm = hm1 |
|
|
620 |
else: |
|
|
621 |
hm = hm2 |
|
|
622 |
|
|
|
623 |
# A half-match was found, sort out the return data. |
|
|
624 |
if len(text1) > len(text2): |
|
|
625 |
(text1_a, text1_b, text2_a, text2_b, mid_common) = hm |
|
|
626 |
else: |
|
|
627 |
(text2_a, text2_b, text1_a, text1_b, mid_common) = hm |
|
|
628 |
return (text1_a, text1_b, text2_a, text2_b, mid_common) |
|
|
629 |
|
|
|
630 |
def diff_cleanupSemantic(self, diffs): |
|
|
631 |
"""Reduce the number of edits by eliminating semantically trivial |
|
|
632 |
equalities. |
|
|
633 |
|
|
|
634 |
Args: |
|
|
635 |
diffs: Array of diff tuples. |
|
|
636 |
""" |
|
|
637 |
changes = False |
|
|
638 |
equalities = [] # Stack of indices where equalities are found. |
|
|
639 |
lastequality = None # Always equal to equalities[-1][1] |
|
|
640 |
pointer = 0 # Index of current position. |
|
|
641 |
length_changes1 = 0 # Number of chars that changed prior to the equality. |
|
|
642 |
length_changes2 = 0 # Number of chars that changed after the equality. |
|
|
643 |
while pointer < len(diffs): |
|
|
644 |
if diffs[pointer][0] == self.DIFF_EQUAL: # equality found |
|
|
645 |
equalities.append(pointer) |
|
|
646 |
length_changes1 = length_changes2 |
|
|
647 |
length_changes2 = 0 |
|
|
648 |
lastequality = diffs[pointer][1] |
|
|
649 |
else: # an insertion or deletion |
|
|
650 |
length_changes2 += len(diffs[pointer][1]) |
|
|
651 |
if (lastequality != None and (len(lastequality) <= length_changes1) and |
|
|
652 |
(len(lastequality) <= length_changes2)): |
|
|
653 |
# Duplicate record |
|
|
654 |
diffs.insert(equalities[-1], (self.DIFF_DELETE, lastequality)) |
|
|
655 |
# Change second copy to insert. |
|
|
656 |
diffs[equalities[-1] + 1] = (self.DIFF_INSERT, |
|
|
657 |
diffs[equalities[-1] + 1][1]) |
|
|
658 |
# Throw away the equality we just deleted. |
|
|
659 |
equalities.pop() |
|
|
660 |
# Throw away the previous equality (it needs to be reevaluated). |
|
|
661 |
if len(equalities) != 0: |
|
|
662 |
equalities.pop() |
|
|
663 |
if len(equalities): |
|
|
664 |
pointer = equalities[-1] |
|
|
665 |
else: |
|
|
666 |
pointer = -1 |
|
|
667 |
length_changes1 = 0 # Reset the counters. |
|
|
668 |
length_changes2 = 0 |
|
|
669 |
lastequality = None |
|
|
670 |
changes = True |
|
|
671 |
pointer += 1 |
|
|
672 |
|
|
|
673 |
if changes: |
|
|
674 |
self.diff_cleanupMerge(diffs) |
|
|
675 |
|
|
|
676 |
self.diff_cleanupSemanticLossless(diffs) |
|
|
677 |
|
|
|
678 |
def diff_cleanupSemanticLossless(self, diffs): |
|
|
679 |
"""Look for single edits surrounded on both sides by equalities |
|
|
680 |
which can be shifted sideways to align the edit to a word boundary. |
|
|
681 |
e.g: The c<ins>at c</ins>ame. -> The <ins>cat </ins>came. |
|
|
682 |
|
|
|
683 |
Args: |
|
|
684 |
diffs: Array of diff tuples. |
|
|
685 |
""" |
|
|
686 |
|
|
|
687 |
def diff_cleanupSemanticScore(one, two): |
|
|
688 |
"""Given two strings, compute a score representing whether the |
|
|
689 |
internal boundary falls on logical boundaries. |
|
|
690 |
Scores range from 5 (best) to 0 (worst). |
|
|
691 |
Closure, but does not reference any external variables. |
|
|
692 |
|
|
|
693 |
Args: |
|
|
694 |
one: First string. |
|
|
695 |
two: Second string. |
|
|
696 |
|
|
|
697 |
Returns: |
|
|
698 |
The score. |
|
|
699 |
""" |
|
|
700 |
if not one or not two: |
|
|
701 |
# Edges are the best. |
|
|
702 |
return 5 |
|
|
703 |
|
|
|
704 |
# Each port of this function behaves slightly differently due to |
|
|
705 |
# subtle differences in each language's definition of things like |
|
|
706 |
# 'whitespace'. Since this function's purpose is largely cosmetic, |
|
|
707 |
# the choice has been made to use each language's native features |
|
|
708 |
# rather than force total conformity. |
|
|
709 |
score = 0 |
|
|
710 |
# One point for non-alphanumeric. |
|
|
711 |
if not one[-1].isalnum() or not two[0].isalnum(): |
|
|
712 |
score += 1 |
|
|
713 |
# Two points for whitespace. |
|
|
714 |
if one[-1].isspace() or two[0].isspace(): |
|
|
715 |
score += 1 |
|
|
716 |
# Three points for line breaks. |
|
|
717 |
if (one[-1] == "\r" or one[-1] == "\n" or |
|
|
718 |
two[0] == "\r" or two[0] == "\n"): |
|
|
719 |
score += 1 |
|
|
720 |
# Four points for blank lines. |
|
|
721 |
if (re.search("\\n\\r?\\n$", one) or |
|
|
722 |
re.match("^\\r?\\n\\r?\\n", two)): |
|
|
723 |
score += 1 |
|
|
724 |
return score |
|
|
725 |
|
|
|
726 |
pointer = 1 |
|
|
727 |
# Intentionally ignore the first and last element (don't need checking). |
|
|
728 |
while pointer < len(diffs) - 1: |
|
|
729 |
if (diffs[pointer - 1][0] == self.DIFF_EQUAL and |
|
|
730 |
diffs[pointer + 1][0] == self.DIFF_EQUAL): |
|
|
731 |
# This is a single edit surrounded by equalities. |
|
|
732 |
equality1 = diffs[pointer - 1][1] |
|
|
733 |
edit = diffs[pointer][1] |
|
|
734 |
equality2 = diffs[pointer + 1][1] |
|
|
735 |
|
|
|
736 |
# First, shift the edit as far left as possible. |
|
|
737 |
commonOffset = self.diff_commonSuffix(equality1, edit) |
|
|
738 |
if commonOffset: |
|
|
739 |
commonString = edit[-commonOffset:] |
|
|
740 |
equality1 = equality1[:-commonOffset] |
|
|
741 |
edit = commonString + edit[:-commonOffset] |
|
|
742 |
equality2 = commonString + equality2 |
|
|
743 |
|
|
|
744 |
# Second, step character by character right, looking for the best fit. |
|
|
745 |
bestEquality1 = equality1 |
|
|
746 |
bestEdit = edit |
|
|
747 |
bestEquality2 = equality2 |
|
|
748 |
bestScore = (diff_cleanupSemanticScore(equality1, edit) + |
|
|
749 |
diff_cleanupSemanticScore(edit, equality2)) |
|
|
750 |
while edit and equality2 and edit[0] == equality2[0]: |
|
|
751 |
equality1 += edit[0] |
|
|
752 |
edit = edit[1:] + equality2[0] |
|
|
753 |
equality2 = equality2[1:] |
|
|
754 |
score = (diff_cleanupSemanticScore(equality1, edit) + |
|
|
755 |
diff_cleanupSemanticScore(edit, equality2)) |
|
|
756 |
# The >= encourages trailing rather than leading whitespace on edits. |
|
|
757 |
if score >= bestScore: |
|
|
758 |
bestScore = score |
|
|
759 |
bestEquality1 = equality1 |
|
|
760 |
bestEdit = edit |
|
|
761 |
bestEquality2 = equality2 |
|
|
762 |
|
|
|
763 |
if diffs[pointer - 1][1] != bestEquality1: |
|
|
764 |
# We have an improvement, save it back to the diff. |
|
|
765 |
if bestEquality1: |
|
|
766 |
diffs[pointer - 1] = (diffs[pointer - 1][0], bestEquality1) |
|
|
767 |
else: |
|
|
768 |
del diffs[pointer - 1] |
|
|
769 |
pointer -= 1 |
|
|
770 |
diffs[pointer] = (diffs[pointer][0], bestEdit) |
|
|
771 |
if bestEquality2: |
|
|
772 |
diffs[pointer + 1] = (diffs[pointer + 1][0], bestEquality2) |
|
|
773 |
else: |
|
|
774 |
del diffs[pointer + 1] |
|
|
775 |
pointer -= 1 |
|
|
776 |
pointer += 1 |
|
|
777 |
|
|
|
778 |
def diff_cleanupEfficiency(self, diffs): |
|
|
779 |
"""Reduce the number of edits by eliminating operationally trivial |
|
|
780 |
equalities. |
|
|
781 |
|
|
|
782 |
Args: |
|
|
783 |
diffs: Array of diff tuples. |
|
|
784 |
""" |
|
|
785 |
changes = False |
|
|
786 |
equalities = [] # Stack of indices where equalities are found. |
|
|
787 |
lastequality = '' # Always equal to equalities[-1][1] |
|
|
788 |
pointer = 0 # Index of current position. |
|
|
789 |
pre_ins = False # Is there an insertion operation before the last equality. |
|
|
790 |
pre_del = False # Is there a deletion operation before the last equality. |
|
|
791 |
post_ins = False # Is there an insertion operation after the last equality. |
|
|
792 |
post_del = False # Is there a deletion operation after the last equality. |
|
|
793 |
while pointer < len(diffs): |
|
|
794 |
if diffs[pointer][0] == self.DIFF_EQUAL: # equality found |
|
|
795 |
if (len(diffs[pointer][1]) < self.Diff_EditCost and |
|
|
796 |
(post_ins or post_del)): |
|
|
797 |
# Candidate found. |
|
|
798 |
equalities.append(pointer) |
|
|
799 |
pre_ins = post_ins |
|
|
800 |
pre_del = post_del |
|
|
801 |
lastequality = diffs[pointer][1] |
|
|
802 |
else: |
|
|
803 |
# Not a candidate, and can never become one. |
|
|
804 |
equalities = [] |
|
|
805 |
lastequality = '' |
|
|
806 |
|
|
|
807 |
post_ins = post_del = False |
|
|
808 |
else: # an insertion or deletion |
|
|
809 |
if diffs[pointer][0] == self.DIFF_DELETE: |
|
|
810 |
post_del = True |
|
|
811 |
else: |
|
|
812 |
post_ins = True |
|
|
813 |
|
|
|
814 |
# Five types to be split: |
|
|
815 |
# <ins>A</ins><del>B</del>XY<ins>C</ins><del>D</del> |
|
|
816 |
# <ins>A</ins>X<ins>C</ins><del>D</del> |
|
|
817 |
# <ins>A</ins><del>B</del>X<ins>C</ins> |
|
|
818 |
# <ins>A</del>X<ins>C</ins><del>D</del> |
|
|
819 |
# <ins>A</ins><del>B</del>X<del>C</del> |
|
|
820 |
|
|
|
821 |
if lastequality and ((pre_ins and pre_del and post_ins and post_del) or |
|
|
822 |
((len(lastequality) < self.Diff_EditCost / 2) and |
|
|
823 |
(pre_ins + pre_del + post_ins + post_del) == 3)): |
|
|
824 |
# Duplicate record |
|
|
825 |
diffs.insert(equalities[-1], (self.DIFF_DELETE, lastequality)) |
|
|
826 |
# Change second copy to insert. |
|
|
827 |
diffs[equalities[-1] + 1] = (self.DIFF_INSERT, |
|
|
828 |
diffs[equalities[-1] + 1][1]) |
|
|
829 |
equalities.pop() # Throw away the equality we just deleted |
|
|
830 |
lastequality = '' |
|
|
831 |
if pre_ins and pre_del: |
|
|
832 |
# No changes made which could affect previous entry, keep going. |
|
|
833 |
post_ins = post_del = True |
|
|
834 |
equalities = [] |
|
|
835 |
else: |
|
|
836 |
if len(equalities): |
|
|
837 |
equalities.pop() # Throw away the previous equality |
|
|
838 |
if len(equalities): |
|
|
839 |
pointer = equalities[-1] |
|
|
840 |
else: |
|
|
841 |
pointer = -1 |
|
|
842 |
post_ins = post_del = False |
|
|
843 |
changes = True |
|
|
844 |
pointer += 1 |
|
|
845 |
|
|
|
846 |
if changes: |
|
|
847 |
self.diff_cleanupMerge(diffs) |
|
|
848 |
|
|
|
849 |
def diff_cleanupMerge(self, diffs): |
|
|
850 |
"""Reorder and merge like edit sections. Merge equalities. |
|
|
851 |
Any edit section can move as long as it doesn't cross an equality. |
|
|
852 |
|
|
|
853 |
Args: |
|
|
854 |
diffs: Array of diff tuples. |
|
|
855 |
""" |
|
|
856 |
diffs.append((self.DIFF_EQUAL, '')) # Add a dummy entry at the end. |
|
|
857 |
pointer = 0 |
|
|
858 |
count_delete = 0 |
|
|
859 |
count_insert = 0 |
|
|
860 |
text_delete = '' |
|
|
861 |
text_insert = '' |
|
|
862 |
while pointer < len(diffs): |
|
|
863 |
if diffs[pointer][0] == self.DIFF_INSERT: |
|
|
864 |
count_insert += 1 |
|
|
865 |
text_insert += diffs[pointer][1] |
|
|
866 |
pointer += 1 |
|
|
867 |
elif diffs[pointer][0] == self.DIFF_DELETE: |
|
|
868 |
count_delete += 1 |
|
|
869 |
text_delete += diffs[pointer][1] |
|
|
870 |
pointer += 1 |
|
|
871 |
elif diffs[pointer][0] == self.DIFF_EQUAL: |
|
|
872 |
# Upon reaching an equality, check for prior redundancies. |
|
|
873 |
if count_delete != 0 or count_insert != 0: |
|
|
874 |
if count_delete != 0 and count_insert != 0: |
|
|
875 |
# Factor out any common prefixies. |
|
|
876 |
commonlength = self.diff_commonPrefix(text_insert, text_delete) |
|
|
877 |
if commonlength != 0: |
|
|
878 |
x = pointer - count_delete - count_insert - 1 |
|
|
879 |
if x >= 0 and diffs[x][0] == self.DIFF_EQUAL: |
|
|
880 |
diffs[x] = (diffs[x][0], diffs[x][1] + |
|
|
881 |
text_insert[:commonlength]) |
|
|
882 |
else: |
|
|
883 |
diffs.insert(0, (self.DIFF_EQUAL, text_insert[:commonlength])) |
|
|
884 |
pointer += 1 |
|
|
885 |
text_insert = text_insert[commonlength:] |
|
|
886 |
text_delete = text_delete[commonlength:] |
|
|
887 |
# Factor out any common suffixies. |
|
|
888 |
commonlength = self.diff_commonSuffix(text_insert, text_delete) |
|
|
889 |
if commonlength != 0: |
|
|
890 |
diffs[pointer] = (diffs[pointer][0], text_insert[-commonlength:] + |
|
|
891 |
diffs[pointer][1]) |
|
|
892 |
text_insert = text_insert[:-commonlength] |
|
|
893 |
text_delete = text_delete[:-commonlength] |
|
|
894 |
# Delete the offending records and add the merged ones. |
|
|
895 |
if count_delete == 0: |
|
|
896 |
diffs[pointer - count_insert : pointer] = [ |
|
|
897 |
(self.DIFF_INSERT, text_insert)] |
|
|
898 |
elif count_insert == 0: |
|
|
899 |
diffs[pointer - count_delete : pointer] = [ |
|
|
900 |
(self.DIFF_DELETE, text_delete)] |
|
|
901 |
else: |
|
|
902 |
diffs[pointer - count_delete - count_insert : pointer] = [ |
|
|
903 |
(self.DIFF_DELETE, text_delete), |
|
|
904 |
(self.DIFF_INSERT, text_insert)] |
|
|
905 |
pointer = pointer - count_delete - count_insert + 1 |
|
|
906 |
if count_delete != 0: |
|
|
907 |
pointer += 1 |
|
|
908 |
if count_insert != 0: |
|
|
909 |
pointer += 1 |
|
|
910 |
elif pointer != 0 and diffs[pointer - 1][0] == self.DIFF_EQUAL: |
|
|
911 |
# Merge this equality with the previous one. |
|
|
912 |
diffs[pointer - 1] = (diffs[pointer - 1][0], |
|
|
913 |
diffs[pointer - 1][1] + diffs[pointer][1]) |
|
|
914 |
del diffs[pointer] |
|
|
915 |
else: |
|
|
916 |
pointer += 1 |
|
|
917 |
|
|
|
918 |
count_insert = 0 |
|
|
919 |
count_delete = 0 |
|
|
920 |
text_delete = '' |
|
|
921 |
text_insert = '' |
|
|
922 |
|
|
|
923 |
if diffs[-1][1] == '': |
|
|
924 |
diffs.pop() # Remove the dummy entry at the end. |
|
|
925 |
|
|
|
926 |
# Second pass: look for single edits surrounded on both sides by equalities |
|
|
927 |
# which can be shifted sideways to eliminate an equality. |
|
|
928 |
# e.g: A<ins>BA</ins>C -> <ins>AB</ins>AC |
|
|
929 |
changes = False |
|
|
930 |
pointer = 1 |
|
|
931 |
# Intentionally ignore the first and last element (don't need checking). |
|
|
932 |
while pointer < len(diffs) - 1: |
|
|
933 |
if (diffs[pointer - 1][0] == self.DIFF_EQUAL and |
|
|
934 |
diffs[pointer + 1][0] == self.DIFF_EQUAL): |
|
|
935 |
# This is a single edit surrounded by equalities. |
|
|
936 |
if diffs[pointer][1].endswith(diffs[pointer - 1][1]): |
|
|
937 |
# Shift the edit over the previous equality. |
|
|
938 |
diffs[pointer] = (diffs[pointer][0], |
|
|
939 |
diffs[pointer - 1][1] + |
|
|
940 |
diffs[pointer][1][:-len(diffs[pointer - 1][1])]) |
|
|
941 |
diffs[pointer + 1] = (diffs[pointer + 1][0], |
|
|
942 |
diffs[pointer - 1][1] + diffs[pointer + 1][1]) |
|
|
943 |
del diffs[pointer - 1] |
|
|
944 |
changes = True |
|
|
945 |
elif diffs[pointer][1].startswith(diffs[pointer + 1][1]): |
|
|
946 |
# Shift the edit over the next equality. |
|
|
947 |
diffs[pointer - 1] = (diffs[pointer - 1][0], |
|
|
948 |
diffs[pointer - 1][1] + diffs[pointer + 1][1]) |
|
|
949 |
diffs[pointer] = (diffs[pointer][0], |
|
|
950 |
diffs[pointer][1][len(diffs[pointer + 1][1]):] + |
|
|
951 |
diffs[pointer + 1][1]) |
|
|
952 |
del diffs[pointer + 1] |
|
|
953 |
changes = True |
|
|
954 |
pointer += 1 |
|
|
955 |
|
|
|
956 |
# If shifts were made, the diff needs reordering and another shift sweep. |
|
|
957 |
if changes: |
|
|
958 |
self.diff_cleanupMerge(diffs) |
|
|
959 |
|
|
|
960 |
def diff_xIndex(self, diffs, loc): |
|
|
961 |
"""loc is a location in text1, compute and return the equivalent location |
|
|
962 |
in text2. e.g. "The cat" vs "The big cat", 1->1, 5->8 |
|
|
963 |
|
|
|
964 |
Args: |
|
|
965 |
diffs: Array of diff tuples. |
|
|
966 |
loc: Location within text1. |
|
|
967 |
|
|
|
968 |
Returns: |
|
|
969 |
Location within text2. |
|
|
970 |
""" |
|
|
971 |
chars1 = 0 |
|
|
972 |
chars2 = 0 |
|
|
973 |
last_chars1 = 0 |
|
|
974 |
last_chars2 = 0 |
|
|
975 |
for x in xrange(len(diffs)): |
|
|
976 |
(op, text) = diffs[x] |
|
|
977 |
if op != self.DIFF_INSERT: # Equality or deletion. |
|
|
978 |
chars1 += len(text) |
|
|
979 |
if op != self.DIFF_DELETE: # Equality or insertion. |
|
|
980 |
chars2 += len(text) |
|
|
981 |
if chars1 > loc: # Overshot the location. |
|
|
982 |
break |
|
|
983 |
last_chars1 = chars1 |
|
|
984 |
last_chars2 = chars2 |
|
|
985 |
|
|
|
986 |
if len(diffs) != x and diffs[x][0] == self.DIFF_DELETE: |
|
|
987 |
# The location was deleted. |
|
|
988 |
return last_chars2 |
|
|
989 |
# Add the remaining len(character). |
|
|
990 |
return last_chars2 + (loc - last_chars1) |
|
|
991 |
|
|
|
992 |
def diff_prettyHtml(self, diffs): |
|
|
993 |
"""Convert a diff array into a pretty HTML report. |
|
|
994 |
|
|
|
995 |
Args: |
|
|
996 |
diffs: Array of diff tuples. |
|
|
997 |
|
|
|
998 |
Returns: |
|
|
999 |
HTML representation. |
|
|
1000 |
""" |
|
|
1001 |
html = [] |
|
|
1002 |
i = 0 |
|
|
1003 |
for (op, data) in diffs: |
|
|
1004 |
text = (data.replace("&", "&").replace("<", "<") |
|
|
1005 |
.replace(">", ">").replace("\n", "¶<BR>")) |
|
|
1006 |
if op == self.DIFF_INSERT: |
|
|
1007 |
html.append("<INS STYLE=\"background:#E6FFE6;\" TITLE=\"i=%i\">%s</INS>" |
|
|
1008 |
% (i, text)) |
|
|
1009 |
elif op == self.DIFF_DELETE: |
|
|
1010 |
html.append("<DEL STYLE=\"background:#FFE6E6;\" TITLE=\"i=%i\">%s</DEL>" |
|
|
1011 |
% (i, text)) |
|
|
1012 |
elif op == self.DIFF_EQUAL: |
|
|
1013 |
html.append("<SPAN TITLE=\"i=%i\">%s</SPAN>" % (i, text)) |
|
|
1014 |
if op != self.DIFF_DELETE: |
|
|
1015 |
i += len(data) |
|
|
1016 |
return "".join(html) |
|
|
1017 |
|
|
|
1018 |
def diff_text1(self, diffs): |
|
|
1019 |
"""Compute and return the source text (all equalities and deletions). |
|
|
1020 |
|
|
|
1021 |
Args: |
|
|
1022 |
diffs: Array of diff tuples. |
|
|
1023 |
|
|
|
1024 |
Returns: |
|
|
1025 |
Source text. |
|
|
1026 |
""" |
|
|
1027 |
text = [] |
|
|
1028 |
for (op, data) in diffs: |
|
|
1029 |
if op != self.DIFF_INSERT: |
|
|
1030 |
text.append(data) |
|
|
1031 |
return "".join(text) |
|
|
1032 |
|
|
|
1033 |
def diff_text2(self, diffs): |
|
|
1034 |
"""Compute and return the destination text (all equalities and insertions). |
|
|
1035 |
|
|
|
1036 |
Args: |
|
|
1037 |
diffs: Array of diff tuples. |
|
|
1038 |
|
|
|
1039 |
Returns: |
|
|
1040 |
Destination text. |
|
|
1041 |
""" |
|
|
1042 |
text = [] |
|
|
1043 |
for (op, data) in diffs: |
|
|
1044 |
if op != self.DIFF_DELETE: |
|
|
1045 |
text.append(data) |
|
|
1046 |
return "".join(text) |
|
|
1047 |
|
|
|
1048 |
def diff_levenshtein(self, diffs): |
|
|
1049 |
"""Compute the Levenshtein distance; the number of inserted, deleted or |
|
|
1050 |
substituted characters. |
|
|
1051 |
|
|
|
1052 |
Args: |
|
|
1053 |
diffs: Array of diff tuples. |
|
|
1054 |
|
|
|
1055 |
Returns: |
|
|
1056 |
Number of changes. |
|
|
1057 |
""" |
|
|
1058 |
levenshtein = 0 |
|
|
1059 |
insertions = 0 |
|
|
1060 |
deletions = 0 |
|
|
1061 |
for (op, data) in diffs: |
|
|
1062 |
if op == self.DIFF_INSERT: |
|
|
1063 |
insertions += len(data) |
|
|
1064 |
elif op == self.DIFF_DELETE: |
|
|
1065 |
deletions += len(data) |
|
|
1066 |
elif op == self.DIFF_EQUAL: |
|
|
1067 |
# A deletion and an insertion is one substitution. |
|
|
1068 |
levenshtein += max(insertions, deletions) |
|
|
1069 |
insertions = 0 |
|
|
1070 |
deletions = 0 |
|
|
1071 |
levenshtein += max(insertions, deletions) |
|
|
1072 |
return levenshtein |
|
|
1073 |
|
|
|
1074 |
def diff_toDelta(self, diffs): |
|
|
1075 |
"""Crush the diff into an encoded string which describes the operations |
|
|
1076 |
required to transform text1 into text2. |
|
|
1077 |
E.g. =3\t-2\t+ing -> Keep 3 chars, delete 2 chars, insert 'ing'. |
|
|
1078 |
Operations are tab-separated. Inserted text is escaped using %xx notation. |
|
|
1079 |
|
|
|
1080 |
Args: |
|
|
1081 |
diffs: Array of diff tuples. |
|
|
1082 |
|
|
|
1083 |
Returns: |
|
|
1084 |
Delta text. |
|
|
1085 |
""" |
|
|
1086 |
text = [] |
|
|
1087 |
for (op, data) in diffs: |
|
|
1088 |
if op == self.DIFF_INSERT: |
|
|
1089 |
# High ascii will raise UnicodeDecodeError. Use Unicode instead. |
|
|
1090 |
data = data.encode("utf-8") |
|
|
1091 |
text.append("+" + urllib.quote(data, "!~*'();/?:@&=+$,# ")) |
|
|
1092 |
elif op == self.DIFF_DELETE: |
|
|
1093 |
text.append("-%d" % len(data)) |
|
|
1094 |
elif op == self.DIFF_EQUAL: |
|
|
1095 |
text.append("=%d" % len(data)) |
|
|
1096 |
return "\t".join(text) |
|
|
1097 |
|
|
|
1098 |
def diff_fromDelta(self, text1, delta): |
|
|
1099 |
"""Given the original text1, and an encoded string which describes the |
|
|
1100 |
operations required to transform text1 into text2, compute the full diff. |
|
|
1101 |
|
|
|
1102 |
Args: |
|
|
1103 |
text1: Source string for the diff. |
|
|
1104 |
delta: Delta text. |
|
|
1105 |
|
|
|
1106 |
Returns: |
|
|
1107 |
Array of diff tuples. |
|
|
1108 |
|
|
|
1109 |
Raises: |
|
|
1110 |
ValueError: If invalid input. |
|
|
1111 |
""" |
|
|
1112 |
if type(delta) == unicode: |
|
|
1113 |
# Deltas should be composed of a subset of ascii chars, Unicode not |
|
|
1114 |
# required. If this encode raises UnicodeEncodeError, delta is invalid. |
|
|
1115 |
delta = delta.encode("ascii") |
|
|
1116 |
diffs = [] |
|
|
1117 |
pointer = 0 # Cursor in text1 |
|
|
1118 |
tokens = delta.split("\t") |
|
|
1119 |
for token in tokens: |
|
|
1120 |
if token == "": |
|
|
1121 |
# Blank tokens are ok (from a trailing \t). |
|
|
1122 |
continue |
|
|
1123 |
# Each token begins with a one character parameter which specifies the |
|
|
1124 |
# operation of this token (delete, insert, equality). |
|
|
1125 |
param = token[1:] |
|
|
1126 |
if token[0] == "+": |
|
|
1127 |
param = urllib.unquote(param).decode("utf-8") |
|
|
1128 |
diffs.append((self.DIFF_INSERT, param)) |
|
|
1129 |
elif token[0] == "-" or token[0] == "=": |
|
|
1130 |
try: |
|
|
1131 |
n = int(param) |
|
|
1132 |
except ValueError: |
|
|
1133 |
raise ValueError, "Invalid number in diff_fromDelta: " + param |
|
|
1134 |
if n < 0: |
|
|
1135 |
raise ValueError, "Negative number in diff_fromDelta: " + param |
|
|
1136 |
text = text1[pointer : pointer + n] |
|
|
1137 |
pointer += n |
|
|
1138 |
if token[0] == "=": |
|
|
1139 |
diffs.append((self.DIFF_EQUAL, text)) |
|
|
1140 |
else: |
|
|
1141 |
diffs.append((self.DIFF_DELETE, text)) |
|
|
1142 |
else: |
|
|
1143 |
# Anything else is an error. |
|
|
1144 |
raise ValueError, ("Invalid diff operation in diff_fromDelta: " + |
|
|
1145 |
token[0]) |
|
|
1146 |
if pointer != len(text1): |
|
|
1147 |
raise ValueError, ( |
|
|
1148 |
"Delta length (%d) does not equal source text length (%d)." % |
|
|
1149 |
(pointer, len(text1))) |
|
|
1150 |
return diffs |
|
|
1151 |
|
|
|
1152 |
# MATCH FUNCTIONS |
|
|
1153 |
|
|
|
1154 |
def match_main(self, text, pattern, loc): |
|
|
1155 |
"""Locate the best instance of 'pattern' in 'text' near 'loc'. |
|
|
1156 |
|
|
|
1157 |
Args: |
|
|
1158 |
text: The text to search. |
|
|
1159 |
pattern: The pattern to search for. |
|
|
1160 |
loc: The location to search around. |
|
|
1161 |
|
|
|
1162 |
Returns: |
|
|
1163 |
Best match index or -1. |
|
|
1164 |
""" |
|
|
1165 |
loc = max(0, min(loc, len(text))) |
|
|
1166 |
if text == pattern: |
|
|
1167 |
# Shortcut (potentially not guaranteed by the algorithm) |
|
|
1168 |
return 0 |
|
|
1169 |
elif not text: |
|
|
1170 |
# Nothing to match. |
|
|
1171 |
return -1 |
|
|
1172 |
elif text[loc:loc + len(pattern)] == pattern: |
|
|
1173 |
# Perfect match at the perfect spot! (Includes case of null pattern) |
|
|
1174 |
return loc |
|
|
1175 |
else: |
|
|
1176 |
# Do a fuzzy compare. |
|
|
1177 |
match = self.match_bitap(text, pattern, loc) |
|
|
1178 |
return match |
|
|
1179 |
|
|
|
1180 |
def match_bitap(self, text, pattern, loc): |
|
|
1181 |
"""Locate the best instance of 'pattern' in 'text' near 'loc' using the |
|
|
1182 |
Bitap algorithm. |
|
|
1183 |
|
|
|
1184 |
Args: |
|
|
1185 |
text: The text to search. |
|
|
1186 |
pattern: The pattern to search for. |
|
|
1187 |
loc: The location to search around. |
|
|
1188 |
|
|
|
1189 |
Returns: |
|
|
1190 |
Best match index or -1. |
|
|
1191 |
""" |
|
|
1192 |
# Python doesn't have a maxint limit, so ignore this check. |
|
|
1193 |
#if self.Match_MaxBits != 0 and len(pattern) > self.Match_MaxBits: |
|
|
1194 |
# raise ValueError("Pattern too long for this application.") |
|
|
1195 |
|
|
|
1196 |
# Initialise the alphabet. |
|
|
1197 |
s = self.match_alphabet(pattern) |
|
|
1198 |
|
|
|
1199 |
def match_bitapScore(e, x): |
|
|
1200 |
"""Compute and return the score for a match with e errors and x location. |
|
|
1201 |
Accesses loc and pattern through being a closure. |
|
|
1202 |
|
|
|
1203 |
Args: |
|
|
1204 |
e: Number of errors in match. |
|
|
1205 |
x: Location of match. |
|
|
1206 |
|
|
|
1207 |
Returns: |
|
|
1208 |
Overall score for match (0.0 = good, 1.0 = bad). |
|
|
1209 |
""" |
|
|
1210 |
accuracy = float(e) / len(pattern) |
|
|
1211 |
proximity = abs(loc - x) |
|
|
1212 |
if not self.Match_Distance: |
|
|
1213 |
# Dodge divide by zero error. |
|
|
1214 |
return proximity and 1.0 or accuracy |
|
|
1215 |
return accuracy + (proximity / float(self.Match_Distance)) |
|
|
1216 |
|
|
|
1217 |
# Highest score beyond which we give up. |
|
|
1218 |
score_threshold = self.Match_Threshold |
|
|
1219 |
# Is there a nearby exact match? (speedup) |
|
|
1220 |
best_loc = text.find(pattern, loc) |
|
|
1221 |
if best_loc != -1: |
|
|
1222 |
score_threshold = min(match_bitapScore(0, best_loc), score_threshold) |
|
|
1223 |
# What about in the other direction? (speedup) |
|
|
1224 |
best_loc = text.rfind(pattern, loc + len(pattern)) |
|
|
1225 |
if best_loc != -1: |
|
|
1226 |
score_threshold = min(match_bitapScore(0, best_loc), score_threshold) |
|
|
1227 |
|
|
|
1228 |
# Initialise the bit arrays. |
|
|
1229 |
matchmask = 1 << (len(pattern) - 1) |
|
|
1230 |
best_loc = -1 |
|
|
1231 |
|
|
|
1232 |
bin_max = len(pattern) + len(text) |
|
|
1233 |
# Empty initialization added to appease pychecker. |
|
|
1234 |
last_rd = None |
|
|
1235 |
for d in xrange(len(pattern)): |
|
|
1236 |
# Scan for the best match each iteration allows for one more error. |
|
|
1237 |
# Run a binary search to determine how far from 'loc' we can stray at |
|
|
1238 |
# this error level. |
|
|
1239 |
bin_min = 0 |
|
|
1240 |
bin_mid = bin_max |
|
|
1241 |
while bin_min < bin_mid: |
|
|
1242 |
if match_bitapScore(d, loc + bin_mid) <= score_threshold: |
|
|
1243 |
bin_min = bin_mid |
|
|
1244 |
else: |
|
|
1245 |
bin_max = bin_mid |
|
|
1246 |
bin_mid = (bin_max - bin_min) / 2 + bin_min |
|
|
1247 |
|
|
|
1248 |
# Use the result from this iteration as the maximum for the next. |
|
|
1249 |
bin_max = bin_mid |
|
|
1250 |
start = max(1, loc - bin_mid + 1) |
|
|
1251 |
finish = min(loc + bin_mid, len(text)) + len(pattern) |
|
|
1252 |
|
|
|
1253 |
rd = range(finish + 1) |
|
|
1254 |
rd.append((1 << d) - 1) |
|
|
1255 |
for j in xrange(finish, start - 1, -1): |
|
|
1256 |
if len(text) <= j - 1: |
|
|
1257 |
# Out of range. |
|
|
1258 |
charMatch = 0 |
|
|
1259 |
else: |
|
|
1260 |
charMatch = s.get(text[j - 1], 0) |
|
|
1261 |
if d == 0: # First pass: exact match. |
|
|
1262 |
rd[j] = ((rd[j + 1] << 1) | 1) & charMatch |
|
|
1263 |
else: # Subsequent passes: fuzzy match. |
|
|
1264 |
rd[j] = ((rd[j + 1] << 1) | 1) & charMatch | ( |
|
|
1265 |
((last_rd[j + 1] | last_rd[j]) << 1) | 1) | last_rd[j + 1] |
|
|
1266 |
if rd[j] & matchmask: |
|
|
1267 |
score = match_bitapScore(d, j - 1) |
|
|
1268 |
# This match will almost certainly be better than any existing match. |
|
|
1269 |
# But check anyway. |
|
|
1270 |
if score <= score_threshold: |
|
|
1271 |
# Told you so. |
|
|
1272 |
score_threshold = score |
|
|
1273 |
best_loc = j - 1 |
|
|
1274 |
if best_loc > loc: |
|
|
1275 |
# When passing loc, don't exceed our current distance from loc. |
|
|
1276 |
start = max(1, 2 * loc - best_loc) |
|
|
1277 |
else: |
|
|
1278 |
# Already passed loc, downhill from here on in. |
|
|
1279 |
break |
|
|
1280 |
# No hope for a (better) match at greater error levels. |
|
|
1281 |
if match_bitapScore(d + 1, loc) > score_threshold: |
|
|
1282 |
break |
|
|
1283 |
last_rd = rd |
|
|
1284 |
return best_loc |
|
|
1285 |
|
|
|
1286 |
def match_alphabet(self, pattern): |
|
|
1287 |
"""Initialise the alphabet for the Bitap algorithm. |
|
|
1288 |
|
|
|
1289 |
Args: |
|
|
1290 |
pattern: The text to encode. |
|
|
1291 |
|
|
|
1292 |
Returns: |
|
|
1293 |
Hash of character locations. |
|
|
1294 |
""" |
|
|
1295 |
s = {} |
|
|
1296 |
for char in pattern: |
|
|
1297 |
s[char] = 0 |
|
|
1298 |
for i in xrange(len(pattern)): |
|
|
1299 |
s[pattern[i]] |= 1 << (len(pattern) - i - 1) |
|
|
1300 |
return s |
|
|
1301 |
|
|
|
1302 |
# PATCH FUNCTIONS |
|
|
1303 |
|
|
|
1304 |
def patch_addContext(self, patch, text): |
|
|
1305 |
"""Increase the context until it is unique, |
|
|
1306 |
but don't let the pattern expand beyond Match_MaxBits. |
|
|
1307 |
|
|
|
1308 |
Args: |
|
|
1309 |
patch: The patch to grow. |
|
|
1310 |
text: Source text. |
|
|
1311 |
""" |
|
|
1312 |
if len(text) == 0: |
|
|
1313 |
return |
|
|
1314 |
pattern = text[patch.start2 : patch.start2 + patch.length1] |
|
|
1315 |
padding = 0 |
|
|
1316 |
|
|
|
1317 |
# Look for the first and last matches of pattern in text. If two different |
|
|
1318 |
# matches are found, increase the pattern length. |
|
|
1319 |
while (text.find(pattern) != text.rfind(pattern) and (self.Match_MaxBits == |
|
|
1320 |
0 or len(pattern) < self.Match_MaxBits - self.Patch_Margin - |
|
|
1321 |
self.Patch_Margin)): |
|
|
1322 |
padding += self.Patch_Margin |
|
|
1323 |
pattern = text[max(0, patch.start2 - padding) : |
|
|
1324 |
patch.start2 + patch.length1 + padding] |
|
|
1325 |
# Add one chunk for good luck. |
|
|
1326 |
padding += self.Patch_Margin |
|
|
1327 |
|
|
|
1328 |
# Add the prefix. |
|
|
1329 |
prefix = text[max(0, patch.start2 - padding) : patch.start2] |
|
|
1330 |
if prefix: |
|
|
1331 |
patch.diffs[:0] = [(self.DIFF_EQUAL, prefix)] |
|
|
1332 |
# Add the suffix. |
|
|
1333 |
suffix = text[patch.start2 + patch.length1 : |
|
|
1334 |
patch.start2 + patch.length1 + padding] |
|
|
1335 |
if suffix: |
|
|
1336 |
patch.diffs.append((self.DIFF_EQUAL, suffix)) |
|
|
1337 |
|
|
|
1338 |
# Roll back the start points. |
|
|
1339 |
patch.start1 -= len(prefix) |
|
|
1340 |
patch.start2 -= len(prefix) |
|
|
1341 |
# Extend lengths. |
|
|
1342 |
patch.length1 += len(prefix) + len(suffix) |
|
|
1343 |
patch.length2 += len(prefix) + len(suffix) |
|
|
1344 |
|
|
|
1345 |
def patch_make(self, a, b=None, c=None): |
|
|
1346 |
"""Compute a list of patches to turn text1 into text2. |
|
|
1347 |
Use diffs if provided, otherwise compute it ourselves. |
|
|
1348 |
There are four ways to call this function, depending on what data is |
|
|
1349 |
available to the caller: |
|
|
1350 |
Method 1: |
|
|
1351 |
a = text1, b = text2 |
|
|
1352 |
Method 2: |
|
|
1353 |
a = diffs |
|
|
1354 |
Method 3 (optimal): |
|
|
1355 |
a = text1, b = diffs |
|
|
1356 |
Method 4 (deprecated, use method 3): |
|
|
1357 |
a = text1, b = text2, c = diffs |
|
|
1358 |
|
|
|
1359 |
Args: |
|
|
1360 |
a: text1 (methods 1,3,4) or Array of diff tuples for text1 to |
|
|
1361 |
text2 (method 2). |
|
|
1362 |
b: text2 (methods 1,4) or Array of diff tuples for text1 to |
|
|
1363 |
text2 (method 3) or undefined (method 2). |
|
|
1364 |
c: Array of diff tuples for text1 to text2 (method 4) or |
|
|
1365 |
undefined (methods 1,2,3). |
|
|
1366 |
|
|
|
1367 |
Returns: |
|
|
1368 |
Array of patch objects. |
|
|
1369 |
""" |
|
|
1370 |
text1 = None |
|
|
1371 |
diffs = None |
|
|
1372 |
# Note that texts may arrive as 'str' or 'unicode'. |
|
|
1373 |
if isinstance(a, basestring) and isinstance(b, basestring) and c is None: |
|
|
1374 |
# Method 1: text1, text2 |
|
|
1375 |
# Compute diffs from text1 and text2. |
|
|
1376 |
text1 = a |
|
|
1377 |
diffs = self.diff_main(text1, b, True) |
|
|
1378 |
if len(diffs) > 2: |
|
|
1379 |
self.diff_cleanupSemantic(diffs) |
|
|
1380 |
self.diff_cleanupEfficiency(diffs) |
|
|
1381 |
elif isinstance(a, list) and b is None and c is None: |
|
|
1382 |
# Method 2: diffs |
|
|
1383 |
# Compute text1 from diffs. |
|
|
1384 |
diffs = a |
|
|
1385 |
text1 = self.diff_text1(diffs) |
|
|
1386 |
elif isinstance(a, basestring) and isinstance(b, list) and c is None: |
|
|
1387 |
# Method 3: text1, diffs |
|
|
1388 |
text1 = a |
|
|
1389 |
diffs = b |
|
|
1390 |
elif (isinstance(a, basestring) and isinstance(b, basestring) and |
|
|
1391 |
isinstance(c, list)): |
|
|
1392 |
# Method 4: text1, text2, diffs |
|
|
1393 |
# text2 is not used. |
|
|
1394 |
text1 = a |
|
|
1395 |
diffs = c |
|
|
1396 |
else: |
|
|
1397 |
raise ValueError("Unknown call format to patch_make.") |
|
|
1398 |
|
|
|
1399 |
if not diffs: |
|
|
1400 |
return [] # Get rid of the None case. |
|
|
1401 |
patches = [] |
|
|
1402 |
patch = patch_obj() |
|
|
1403 |
char_count1 = 0 # Number of characters into the text1 string. |
|
|
1404 |
char_count2 = 0 # Number of characters into the text2 string. |
|
|
1405 |
prepatch_text = text1 # Recreate the patches to determine context info. |
|
|
1406 |
postpatch_text = text1 |
|
|
1407 |
for x in xrange(len(diffs)): |
|
|
1408 |
(diff_type, diff_text) = diffs[x] |
|
|
1409 |
if len(patch.diffs) == 0 and diff_type != self.DIFF_EQUAL: |
|
|
1410 |
# A new patch starts here. |
|
|
1411 |
patch.start1 = char_count1 |
|
|
1412 |
patch.start2 = char_count2 |
|
|
1413 |
if diff_type == self.DIFF_INSERT: |
|
|
1414 |
# Insertion |
|
|
1415 |
patch.diffs.append(diffs[x]) |
|
|
1416 |
patch.length2 += len(diff_text) |
|
|
1417 |
postpatch_text = (postpatch_text[:char_count2] + diff_text + |
|
|
1418 |
postpatch_text[char_count2:]) |
|
|
1419 |
elif diff_type == self.DIFF_DELETE: |
|
|
1420 |
# Deletion. |
|
|
1421 |
patch.length1 += len(diff_text) |
|
|
1422 |
patch.diffs.append(diffs[x]) |
|
|
1423 |
postpatch_text = (postpatch_text[:char_count2] + |
|
|
1424 |
postpatch_text[char_count2 + len(diff_text):]) |
|
|
1425 |
elif (diff_type == self.DIFF_EQUAL and |
|
|
1426 |
len(diff_text) <= 2 * self.Patch_Margin and |
|
|
1427 |
len(patch.diffs) != 0 and len(diffs) != x + 1): |
|
|
1428 |
# Small equality inside a patch. |
|
|
1429 |
patch.diffs.append(diffs[x]) |
|
|
1430 |
patch.length1 += len(diff_text) |
|
|
1431 |
patch.length2 += len(diff_text) |
|
|
1432 |
|
|
|
1433 |
if (diff_type == self.DIFF_EQUAL and |
|
|
1434 |
len(diff_text) >= 2 * self.Patch_Margin): |
|
|
1435 |
# Time for a new patch. |
|
|
1436 |
if len(patch.diffs) != 0: |
|
|
1437 |
self.patch_addContext(patch, prepatch_text) |
|
|
1438 |
patches.append(patch) |
|
|
1439 |
patch = patch_obj() |
|
|
1440 |
# Unlike Unidiff, our patch lists have a rolling context. |
|
|
1441 |
# http://code.google.com/p/google-diff-match-patch/wiki/Unidiff |
|
|
1442 |
# Update prepatch text & pos to reflect the application of the |
|
|
1443 |
# just completed patch. |
|
|
1444 |
prepatch_text = postpatch_text |
|
|
1445 |
char_count1 = char_count2 |
|
|
1446 |
|
|
|
1447 |
# Update the current character count. |
|
|
1448 |
if diff_type != self.DIFF_INSERT: |
|
|
1449 |
char_count1 += len(diff_text) |
|
|
1450 |
if diff_type != self.DIFF_DELETE: |
|
|
1451 |
char_count2 += len(diff_text) |
|
|
1452 |
|
|
|
1453 |
# Pick up the leftover patch if not empty. |
|
|
1454 |
if len(patch.diffs) != 0: |
|
|
1455 |
self.patch_addContext(patch, prepatch_text) |
|
|
1456 |
patches.append(patch) |
|
|
1457 |
return patches |
|
|
1458 |
|
|
|
1459 |
def patch_deepCopy(self, patches): |
|
|
1460 |
"""Given an array of patches, return another array that is identical. |
|
|
1461 |
|
|
|
1462 |
Args: |
|
|
1463 |
patches: Array of patch objects. |
|
|
1464 |
|
|
|
1465 |
Returns: |
|
|
1466 |
Array of patch objects. |
|
|
1467 |
""" |
|
|
1468 |
patchesCopy = [] |
|
|
1469 |
for patch in patches: |
|
|
1470 |
patchCopy = patch_obj() |
|
|
1471 |
# No need to deep copy the tuples since they are immutable. |
|
|
1472 |
patchCopy.diffs = patch.diffs[:] |
|
|
1473 |
patchCopy.start1 = patch.start1 |
|
|
1474 |
patchCopy.start2 = patch.start2 |
|
|
1475 |
patchCopy.length1 = patch.length1 |
|
|
1476 |
patchCopy.length2 = patch.length2 |
|
|
1477 |
patchesCopy.append(patchCopy) |
|
|
1478 |
return patchesCopy |
|
|
1479 |
|
|
|
1480 |
def patch_apply(self, patches, text): |
|
|
1481 |
"""Merge a set of patches onto the text. Return a patched text, as well |
|
|
1482 |
as a list of true/false values indicating which patches were applied. |
|
|
1483 |
|
|
|
1484 |
Args: |
|
|
1485 |
patches: Array of patch objects. |
|
|
1486 |
text: Old text. |
|
|
1487 |
|
|
|
1488 |
Returns: |
|
|
1489 |
Two element Array, containing the new text and an array of boolean values. |
|
|
1490 |
""" |
|
|
1491 |
if not patches: |
|
|
1492 |
return (text, []) |
|
|
1493 |
|
|
|
1494 |
# Deep copy the patches so that no changes are made to originals. |
|
|
1495 |
patches = self.patch_deepCopy(patches) |
|
|
1496 |
|
|
|
1497 |
nullPadding = self.patch_addPadding(patches) |
|
|
1498 |
text = nullPadding + text + nullPadding |
|
|
1499 |
self.patch_splitMax(patches) |
|
|
1500 |
|
|
|
1501 |
# delta keeps track of the offset between the expected and actual location |
|
|
1502 |
# of the previous patch. If there are patches expected at positions 10 and |
|
|
1503 |
# 20, but the first patch was found at 12, delta is 2 and the second patch |
|
|
1504 |
# has an effective expected position of 22. |
|
|
1505 |
delta = 0 |
|
|
1506 |
results = [] |
|
|
1507 |
for patch in patches: |
|
|
1508 |
expected_loc = patch.start2 + delta |
|
|
1509 |
text1 = self.diff_text1(patch.diffs) |
|
|
1510 |
end_loc = -1 |
|
|
1511 |
if len(text1) > self.Match_MaxBits: |
|
|
1512 |
# patch_splitMax will only provide an oversized pattern in the case of |
|
|
1513 |
# a monster delete. |
|
|
1514 |
start_loc = self.match_main(text, text1[:self.Match_MaxBits], |
|
|
1515 |
expected_loc) |
|
|
1516 |
if start_loc != -1: |
|
|
1517 |
end_loc = self.match_main(text, text1[-self.Match_MaxBits:], |
|
|
1518 |
expected_loc + len(text1) - self.Match_MaxBits) |
|
|
1519 |
if end_loc == -1 or start_loc >= end_loc: |
|
|
1520 |
# Can't find valid trailing context. Drop this patch. |
|
|
1521 |
start_loc = -1 |
|
|
1522 |
else: |
|
|
1523 |
start_loc = self.match_main(text, text1, expected_loc) |
|
|
1524 |
if start_loc == -1: |
|
|
1525 |
# No match found. :( |
|
|
1526 |
results.append(False) |
|
|
1527 |
# Subtract the delta for this failed patch from subsequent patches. |
|
|
1528 |
delta -= patch.length2 - patch.length1 |
|
|
1529 |
else: |
|
|
1530 |
# Found a match. :) |
|
|
1531 |
results.append(True) |
|
|
1532 |
delta = start_loc - expected_loc |
|
|
1533 |
if end_loc == -1: |
|
|
1534 |
text2 = text[start_loc : start_loc + len(text1)] |
|
|
1535 |
else: |
|
|
1536 |
text2 = text[start_loc : end_loc + self.Match_MaxBits] |
|
|
1537 |
if text1 == text2: |
|
|
1538 |
# Perfect match, just shove the replacement text in. |
|
|
1539 |
text = (text[:start_loc] + self.diff_text2(patch.diffs) + |
|
|
1540 |
text[start_loc + len(text1):]) |
|
|
1541 |
else: |
|
|
1542 |
# Imperfect match. |
|
|
1543 |
# Run a diff to get a framework of equivalent indices. |
|
|
1544 |
diffs = self.diff_main(text1, text2, False) |
|
|
1545 |
if (len(text1) > self.Match_MaxBits and |
|
|
1546 |
self.diff_levenshtein(diffs) / float(len(text1)) > |
|
|
1547 |
self.Patch_DeleteThreshold): |
|
|
1548 |
# The end points match, but the content is unacceptably bad. |
|
|
1549 |
results[-1] = False |
|
|
1550 |
else: |
|
|
1551 |
self.diff_cleanupSemanticLossless(diffs) |
|
|
1552 |
index1 = 0 |
|
|
1553 |
for (op, data) in patch.diffs: |
|
|
1554 |
if op != self.DIFF_EQUAL: |
|
|
1555 |
index2 = self.diff_xIndex(diffs, index1) |
|
|
1556 |
if op == self.DIFF_INSERT: # Insertion |
|
|
1557 |
text = text[:start_loc + index2] + data + text[start_loc + |
|
|
1558 |
index2:] |
|
|
1559 |
elif op == self.DIFF_DELETE: # Deletion |
|
|
1560 |
text = text[:start_loc + index2] + text[start_loc + |
|
|
1561 |
self.diff_xIndex(diffs, index1 + len(data)):] |
|
|
1562 |
if op != self.DIFF_DELETE: |
|
|
1563 |
index1 += len(data) |
|
|
1564 |
# Strip the padding off. |
|
|
1565 |
text = text[len(nullPadding):-len(nullPadding)] |
|
|
1566 |
return (text, results) |
|
|
1567 |
|
|
|
1568 |
def patch_addPadding(self, patches): |
|
|
1569 |
"""Add some padding on text start and end so that edges can match |
|
|
1570 |
something. Intended to be called only from within patch_apply. |
|
|
1571 |
|
|
|
1572 |
Args: |
|
|
1573 |
patches: Array of patch objects. |
|
|
1574 |
|
|
|
1575 |
Returns: |
|
|
1576 |
The padding string added to each side. |
|
|
1577 |
""" |
|
|
1578 |
paddingLength = self.Patch_Margin |
|
|
1579 |
nullPadding = "" |
|
|
1580 |
for x in xrange(1, paddingLength + 1): |
|
|
1581 |
nullPadding += chr(x) |
|
|
1582 |
|
|
|
1583 |
# Bump all the patches forward. |
|
|
1584 |
for patch in patches: |
|
|
1585 |
patch.start1 += paddingLength |
|
|
1586 |
patch.start2 += paddingLength |
|
|
1587 |
|
|
|
1588 |
# Add some padding on start of first diff. |
|
|
1589 |
patch = patches[0] |
|
|
1590 |
diffs = patch.diffs |
|
|
1591 |
if not diffs or diffs[0][0] != self.DIFF_EQUAL: |
|
|
1592 |
# Add nullPadding equality. |
|
|
1593 |
diffs.insert(0, (self.DIFF_EQUAL, nullPadding)) |
|
|
1594 |
patch.start1 -= paddingLength # Should be 0. |
|
|
1595 |
patch.start2 -= paddingLength # Should be 0. |
|
|
1596 |
patch.length1 += paddingLength |
|
|
1597 |
patch.length2 += paddingLength |
|
|
1598 |
elif paddingLength > len(diffs[0][1]): |
|
|
1599 |
# Grow first equality. |
|
|
1600 |
extraLength = paddingLength - len(diffs[0][1]) |
|
|
1601 |
newText = nullPadding[len(diffs[0][1]):] + diffs[0][1] |
|
|
1602 |
diffs[0] = (diffs[0][0], newText) |
|
|
1603 |
patch.start1 -= extraLength |
|
|
1604 |
patch.start2 -= extraLength |
|
|
1605 |
patch.length1 += extraLength |
|
|
1606 |
patch.length2 += extraLength |
|
|
1607 |
|
|
|
1608 |
# Add some padding on end of last diff. |
|
|
1609 |
patch = patches[-1] |
|
|
1610 |
diffs = patch.diffs |
|
|
1611 |
if not diffs or diffs[-1][0] != self.DIFF_EQUAL: |
|
|
1612 |
# Add nullPadding equality. |
|
|
1613 |
diffs.append((self.DIFF_EQUAL, nullPadding)) |
|
|
1614 |
patch.length1 += paddingLength |
|
|
1615 |
patch.length2 += paddingLength |
|
|
1616 |
elif paddingLength > len(diffs[-1][1]): |
|
|
1617 |
# Grow last equality. |
|
|
1618 |
extraLength = paddingLength - len(diffs[-1][1]) |
|
|
1619 |
newText = diffs[-1][1] + nullPadding[:extraLength] |
|
|
1620 |
diffs[-1] = (diffs[-1][0], newText) |
|
|
1621 |
patch.length1 += extraLength |
|
|
1622 |
patch.length2 += extraLength |
|
|
1623 |
|
|
|
1624 |
return nullPadding |
|
|
1625 |
|
|
|
1626 |
def patch_splitMax(self, patches): |
|
|
1627 |
"""Look through the patches and break up any which are longer than the |
|
|
1628 |
maximum limit of the match algorithm. |
|
|
1629 |
|
|
|
1630 |
Args: |
|
|
1631 |
patches: Array of patch objects. |
|
|
1632 |
""" |
|
|
1633 |
if self.Match_MaxBits == 0: |
|
|
1634 |
return |
|
|
1635 |
for x in xrange(len(patches)): |
|
|
1636 |
if patches[x].length1 > self.Match_MaxBits: |
|
|
1637 |
bigpatch = patches[x] |
|
|
1638 |
# Remove the big old patch. |
|
|
1639 |
del patches[x] |
|
|
1640 |
x -= 1 |
|
|
1641 |
patch_size = self.Match_MaxBits |
|
|
1642 |
start1 = bigpatch.start1 |
|
|
1643 |
start2 = bigpatch.start2 |
|
|
1644 |
precontext = '' |
|
|
1645 |
while len(bigpatch.diffs) != 0: |
|
|
1646 |
# Create one of several smaller patches. |
|
|
1647 |
patch = patch_obj() |
|
|
1648 |
empty = True |
|
|
1649 |
patch.start1 = start1 - len(precontext) |
|
|
1650 |
patch.start2 = start2 - len(precontext) |
|
|
1651 |
if precontext: |
|
|
1652 |
patch.length1 = patch.length2 = len(precontext) |
|
|
1653 |
patch.diffs.append((self.DIFF_EQUAL, precontext)) |
|
|
1654 |
|
|
|
1655 |
while (len(bigpatch.diffs) != 0 and |
|
|
1656 |
patch.length1 < patch_size - self.Patch_Margin): |
|
|
1657 |
(diff_type, diff_text) = bigpatch.diffs[0] |
|
|
1658 |
if diff_type == self.DIFF_INSERT: |
|
|
1659 |
# Insertions are harmless. |
|
|
1660 |
patch.length2 += len(diff_text) |
|
|
1661 |
start2 += len(diff_text) |
|
|
1662 |
patch.diffs.append(bigpatch.diffs.pop(0)) |
|
|
1663 |
empty = False |
|
|
1664 |
elif (diff_type == self.DIFF_DELETE and len(patch.diffs) == 1 and |
|
|
1665 |
patch.diffs[0][0] == self.DIFF_EQUAL and |
|
|
1666 |
len(diff_text) > 2 * patch_size): |
|
|
1667 |
# This is a large deletion. Let it pass in one chunk. |
|
|
1668 |
patch.length1 += len(diff_text) |
|
|
1669 |
start1 += len(diff_text) |
|
|
1670 |
empty = False |
|
|
1671 |
patch.diffs.append((diff_type, diff_text)) |
|
|
1672 |
del bigpatch.diffs[0] |
|
|
1673 |
else: |
|
|
1674 |
# Deletion or equality. Only take as much as we can stomach. |
|
|
1675 |
diff_text = diff_text[:patch_size - patch.length1 - |
|
|
1676 |
self.Patch_Margin] |
|
|
1677 |
patch.length1 += len(diff_text) |
|
|
1678 |
start1 += len(diff_text) |
|
|
1679 |
if diff_type == self.DIFF_EQUAL: |
|
|
1680 |
patch.length2 += len(diff_text) |
|
|
1681 |
start2 += len(diff_text) |
|
|
1682 |
else: |
|
|
1683 |
empty = False |
|
|
1684 |
|
|
|
1685 |
patch.diffs.append((diff_type, diff_text)) |
|
|
1686 |
if diff_text == bigpatch.diffs[0][1]: |
|
|
1687 |
del bigpatch.diffs[0] |
|
|
1688 |
else: |
|
|
1689 |
bigpatch.diffs[0] = (bigpatch.diffs[0][0], |
|
|
1690 |
bigpatch.diffs[0][1][len(diff_text):]) |
|
|
1691 |
|
|
|
1692 |
# Compute the head context for the next patch. |
|
|
1693 |
precontext = self.diff_text2(patch.diffs) |
|
|
1694 |
precontext = precontext[-self.Patch_Margin:] |
|
|
1695 |
# Append the end context for this patch. |
|
|
1696 |
postcontext = self.diff_text1(bigpatch.diffs)[:self.Patch_Margin] |
|
|
1697 |
if postcontext: |
|
|
1698 |
patch.length1 += len(postcontext) |
|
|
1699 |
patch.length2 += len(postcontext) |
|
|
1700 |
if len(patch.diffs) != 0 and patch.diffs[-1][0] == self.DIFF_EQUAL: |
|
|
1701 |
patch.diffs[-1] = (self.DIFF_EQUAL, patch.diffs[-1][1] + |
|
|
1702 |
postcontext) |
|
|
1703 |
else: |
|
|
1704 |
patch.diffs.append((self.DIFF_EQUAL, postcontext)) |
|
|
1705 |
|
|
|
1706 |
if not empty: |
|
|
1707 |
x += 1 |
|
|
1708 |
patches.insert(x, patch) |
|
|
1709 |
|
|
|
1710 |
def patch_toText(self, patches): |
|
|
1711 |
"""Take a list of patches and return a textual representation. |
|
|
1712 |
|
|
|
1713 |
Args: |
|
|
1714 |
patches: Array of patch objects. |
|
|
1715 |
|
|
|
1716 |
Returns: |
|
|
1717 |
Text representation of patches. |
|
|
1718 |
""" |
|
|
1719 |
text = [] |
|
|
1720 |
for patch in patches: |
|
|
1721 |
text.append(str(patch)) |
|
|
1722 |
return "".join(text) |
|
|
1723 |
|
|
|
1724 |
def patch_fromText(self, textline): |
|
|
1725 |
"""Parse a textual representation of patches and return a list of patch |
|
|
1726 |
objects. |
|
|
1727 |
|
|
|
1728 |
Args: |
|
|
1729 |
textline: Text representation of patches. |
|
|
1730 |
|
|
|
1731 |
Returns: |
|
|
1732 |
Array of patch objects. |
|
|
1733 |
|
|
|
1734 |
Raises: |
|
|
1735 |
ValueError: If invalid input. |
|
|
1736 |
""" |
|
|
1737 |
if type(textline) == unicode: |
|
|
1738 |
# Patches should be composed of a subset of ascii chars, Unicode not |
|
|
1739 |
# required. If this encode raises UnicodeEncodeError, patch is invalid. |
|
|
1740 |
textline = textline.encode("ascii") |
|
|
1741 |
patches = [] |
|
|
1742 |
if not textline: |
|
|
1743 |
return patches |
|
|
1744 |
text = textline.split('\n') |
|
|
1745 |
while len(text) != 0: |
|
|
1746 |
m = re.match("^@@ -(\d+),?(\d*) \+(\d+),?(\d*) @@$", text[0]) |
|
|
1747 |
if not m: |
|
|
1748 |
raise ValueError, "Invalid patch string: " + text[0] |
|
|
1749 |
patch = patch_obj() |
|
|
1750 |
patches.append(patch) |
|
|
1751 |
patch.start1 = int(m.group(1)) |
|
|
1752 |
if m.group(2) == '': |
|
|
1753 |
patch.start1 -= 1 |
|
|
1754 |
patch.length1 = 1 |
|
|
1755 |
elif m.group(2) == '0': |
|
|
1756 |
patch.length1 = 0 |
|
|
1757 |
else: |
|
|
1758 |
patch.start1 -= 1 |
|
|
1759 |
patch.length1 = int(m.group(2)) |
|
|
1760 |
|
|
|
1761 |
patch.start2 = int(m.group(3)) |
|
|
1762 |
if m.group(4) == '': |
|
|
1763 |
patch.start2 -= 1 |
|
|
1764 |
patch.length2 = 1 |
|
|
1765 |
elif m.group(4) == '0': |
|
|
1766 |
patch.length2 = 0 |
|
|
1767 |
else: |
|
|
1768 |
patch.start2 -= 1 |
|
|
1769 |
patch.length2 = int(m.group(4)) |
|
|
1770 |
|
|
|
1771 |
del text[0] |
|
|
1772 |
|
|
|
1773 |
while len(text) != 0: |
|
|
1774 |
if text[0]: |
|
|
1775 |
sign = text[0][0] |
|
|
1776 |
else: |
|
|
1777 |
sign = '' |
|
|
1778 |
line = urllib.unquote(text[0][1:]) |
|
|
1779 |
line = line.decode("utf-8") |
|
|
1780 |
if sign == '+': |
|
|
1781 |
# Insertion. |
|
|
1782 |
patch.diffs.append((self.DIFF_INSERT, line)) |
|
|
1783 |
elif sign == '-': |
|
|
1784 |
# Deletion. |
|
|
1785 |
patch.diffs.append((self.DIFF_DELETE, line)) |
|
|
1786 |
elif sign == ' ': |
|
|
1787 |
# Minor equality. |
|
|
1788 |
patch.diffs.append((self.DIFF_EQUAL, line)) |
|
|
1789 |
elif sign == '@': |
|
|
1790 |
# Start of next patch. |
|
|
1791 |
break |
|
|
1792 |
elif sign == '': |
|
|
1793 |
# Blank line? Whatever. |
|
|
1794 |
pass |
|
|
1795 |
else: |
|
|
1796 |
# WTF? |
|
|
1797 |
raise ValueError, "Invalid patch mode: '%s'\n%s" % (sign, line) |
|
|
1798 |
del text[0] |
|
|
1799 |
return patches |
|
|
1800 |
|
|
|
1801 |
|
|
|
1802 |
class patch_obj: |
|
|
1803 |
"""Class representing one patch operation. |
|
|
1804 |
""" |
|
|
1805 |
|
|
|
1806 |
def __init__(self): |
|
|
1807 |
"""Initializes with an empty list of diffs. |
|
|
1808 |
""" |
|
|
1809 |
self.diffs = [] |
|
|
1810 |
self.start1 = None |
|
|
1811 |
self.start2 = None |
|
|
1812 |
self.length1 = 0 |
|
|
1813 |
self.length2 = 0 |
|
|
1814 |
|
|
|
1815 |
def __str__(self): |
|
|
1816 |
"""Emmulate GNU diff's format. |
|
|
1817 |
Header: @@ -382,8 +481,9 @@ |
|
|
1818 |
Indicies are printed as 1-based, not 0-based. |
|
|
1819 |
|
|
|
1820 |
Returns: |
|
|
1821 |
The GNU diff string. |
|
|
1822 |
""" |
|
|
1823 |
if self.length1 == 0: |
|
|
1824 |
coords1 = str(self.start1) + ",0" |
|
|
1825 |
elif self.length1 == 1: |
|
|
1826 |
coords1 = str(self.start1 + 1) |
|
|
1827 |
else: |
|
|
1828 |
coords1 = str(self.start1 + 1) + "," + str(self.length1) |
|
|
1829 |
if self.length2 == 0: |
|
|
1830 |
coords2 = str(self.start2) + ",0" |
|
|
1831 |
elif self.length2 == 1: |
|
|
1832 |
coords2 = str(self.start2 + 1) |
|
|
1833 |
else: |
|
|
1834 |
coords2 = str(self.start2 + 1) + "," + str(self.length2) |
|
|
1835 |
text = ["@@ -", coords1, " +", coords2, " @@\n"] |
|
|
1836 |
# Escape the body of the patch with %xx notation. |
|
|
1837 |
for (op, data) in self.diffs: |
|
|
1838 |
if op == diff_match_patch.DIFF_INSERT: |
|
|
1839 |
text.append("+") |
|
|
1840 |
elif op == diff_match_patch.DIFF_DELETE: |
|
|
1841 |
text.append("-") |
|
|
1842 |
elif op == diff_match_patch.DIFF_EQUAL: |
|
|
1843 |
text.append(" ") |
|
|
1844 |
# High ascii will raise UnicodeDecodeError. Use Unicode instead. |
|
|
1845 |
data = data.encode("utf-8") |
|
|
1846 |
text.append(urllib.quote(data, "!~*'();/?:@&=+$,# ") + "\n") |
|
|
1847 |
return "".join(text) |