Python code coverage for Lib/difflib.py

#	count	content
1	n/a	"""
2	n/a	Module difflib -- helpers for computing deltas between objects.
3	n/a
4	n/a	Function get_close_matches(word, possibilities, n=3, cutoff=0.6):
5	n/a	Use SequenceMatcher to return list of the best "good enough" matches.
6	n/a
7	n/a	Function context_diff(a, b):
8	n/a	For two lists of strings, return a delta in context diff format.
9	n/a
10	n/a	Function ndiff(a, b):
11	n/a	Return a delta: the difference between `a` and `b` (lists of strings).
12	n/a
13	n/a	Function restore(delta, which):
14	n/a	Return one of the two sequences that generated an ndiff delta.
15	n/a
16	n/a	Function unified_diff(a, b):
17	n/a	For two lists of strings, return a delta in unified diff format.
18	n/a
19	n/a	Class SequenceMatcher:
20	n/a	A flexible class for comparing pairs of sequences of any type.
21	n/a
22	n/a	Class Differ:
23	n/a	For producing human-readable deltas from sequences of lines of text.
24	n/a
25	n/a	Class HtmlDiff:
26	n/a	For producing HTML side by side comparison with change highlights.
27	n/a	"""
28	n/a
29	n/a	__all__ = ['get_close_matches', 'ndiff', 'restore', 'SequenceMatcher',
30	n/a	'Differ','IS_CHARACTER_JUNK', 'IS_LINE_JUNK', 'context_diff',
31	n/a	'unified_diff', 'diff_bytes', 'HtmlDiff', 'Match']
32	n/a
33	n/a	from heapq import nlargest as _nlargest
34	n/a	from collections import namedtuple as _namedtuple
35	n/a
36	n/a	Match = _namedtuple('Match', 'a b size')
37	n/a
38	n/a	def _calculate_ratio(matches, length):
39	n/a	if length:
40	n/a	return 2.0 * matches / length
41	n/a	return 1.0
42	n/a
43	n/a	class SequenceMatcher:
44	n/a
45	n/a	"""
46	n/a	SequenceMatcher is a flexible class for comparing pairs of sequences of
47	n/a	any type, so long as the sequence elements are hashable. The basic
48	n/a	algorithm predates, and is a little fancier than, an algorithm
49	n/a	published in the late 1980's by Ratcliff and Obershelp under the
50	n/a	hyperbolic name "gestalt pattern matching". The basic idea is to find
51	n/a	the longest contiguous matching subsequence that contains no "junk"
52	n/a	elements (R-O doesn't address junk). The same idea is then applied
53	n/a	recursively to the pieces of the sequences to the left and to the right
54	n/a	of the matching subsequence. This does not yield minimal edit
55	n/a	sequences, but does tend to yield matches that "look right" to people.
56	n/a
57	n/a	SequenceMatcher tries to compute a "human-friendly diff" between two
58	n/a	sequences. Unlike e.g. UNIX(tm) diff, the fundamental notion is the
59	n/a	longest contiguous & junk-free matching subsequence. That's what
60	n/a	catches peoples' eyes. The Windows(tm) windiff has another interesting
61	n/a	notion, pairing up elements that appear uniquely in each sequence.
62	n/a	That, and the method here, appear to yield more intuitive difference
63	n/a	reports than does diff. This method appears to be the least vulnerable
64	n/a	to synching up on blocks of "junk lines", though (like blank lines in
65	n/a	ordinary text files, or maybe "<P>" lines in HTML files). That may be
66	n/a	because this is the only method of the 3 that has a concept of
67	n/a	"junk" <wink>.
68	n/a
69	n/a	Example, comparing two strings, and considering blanks to be "junk":
70	n/a
71	n/a	>>> s = SequenceMatcher(lambda x: x == " ",
72	n/a	... "private Thread currentThread;",
73	n/a	... "private volatile Thread currentThread;")
74	n/a	>>>
75	n/a
76	n/a	.ratio() returns a float in [0, 1], measuring the "similarity" of the
77	n/a	sequences. As a rule of thumb, a .ratio() value over 0.6 means the
78	n/a	sequences are close matches:
79	n/a
80	n/a	>>> print(round(s.ratio(), 3))
81	n/a	0.866
82	n/a	>>>
83	n/a
84	n/a	If you're only interested in where the sequences match,
85	n/a	.get_matching_blocks() is handy:
86	n/a
87	n/a	>>> for block in s.get_matching_blocks():
88	n/a	... print("a[%d] and b[%d] match for %d elements" % block)
89	n/a	a[0] and b[0] match for 8 elements
90	n/a	a[8] and b[17] match for 21 elements
91	n/a	a[29] and b[38] match for 0 elements
92	n/a
93	n/a	Note that the last tuple returned by .get_matching_blocks() is always a
94	n/a	dummy, (len(a), len(b), 0), and this is the only case in which the last
95	n/a	tuple element (number of elements matched) is 0.
96	n/a
97	n/a	If you want to know how to change the first sequence into the second,
98	n/a	use .get_opcodes():
99	n/a
100	n/a	>>> for opcode in s.get_opcodes():
101	n/a	... print("%6s a[%d:%d] b[%d:%d]" % opcode)
102	n/a	equal a[0:8] b[0:8]
103	n/a	insert a[8:8] b[8:17]
104	n/a	equal a[8:29] b[17:38]
105	n/a
106	n/a	See the Differ class for a fancy human-friendly file differencer, which
107	n/a	uses SequenceMatcher both to compare sequences of lines, and to compare
108	n/a	sequences of characters within similar (near-matching) lines.
109	n/a
110	n/a	See also function get_close_matches() in this module, which shows how
111	n/a	simple code building on SequenceMatcher can be used to do useful work.
112	n/a
113	n/a	Timing: Basic R-O is cubic time worst case and quadratic time expected
114	n/a	case. SequenceMatcher is quadratic time for the worst case and has
115	n/a	expected-case behavior dependent in a complicated way on how many
116	n/a	elements the sequences have in common; best case time is linear.
117	n/a
118	n/a	Methods:
119	n/a
120	n/a	__init__(isjunk=None, a='', b='')
121	n/a	Construct a SequenceMatcher.
122	n/a
123	n/a	set_seqs(a, b)
124	n/a	Set the two sequences to be compared.
125	n/a
126	n/a	set_seq1(a)
127	n/a	Set the first sequence to be compared.
128	n/a
129	n/a	set_seq2(b)
130	n/a	Set the second sequence to be compared.
131	n/a
132	n/a	find_longest_match(alo, ahi, blo, bhi)
133	n/a	Find longest matching block in a[alo:ahi] and b[blo:bhi].
134	n/a
135	n/a	get_matching_blocks()
136	n/a	Return list of triples describing matching subsequences.
137	n/a
138	n/a	get_opcodes()
139	n/a	Return list of 5-tuples describing how to turn a into b.
140	n/a
141	n/a	ratio()
142	n/a	Return a measure of the sequences' similarity (float in [0,1]).
143	n/a
144	n/a	quick_ratio()
145	n/a	Return an upper bound on .ratio() relatively quickly.
146	n/a
147	n/a	real_quick_ratio()
148	n/a	Return an upper bound on ratio() very quickly.
149	n/a	"""
150	n/a
151	n/a	def __init__(self, isjunk=None, a='', b='', autojunk=True):
152	n/a	"""Construct a SequenceMatcher.
153	n/a
154	n/a	Optional arg isjunk is None (the default), or a one-argument
155	n/a	function that takes a sequence element and returns true iff the
156	n/a	element is junk. None is equivalent to passing "lambda x: 0", i.e.
157	n/a	no elements are considered to be junk. For example, pass
158	n/a	lambda x: x in " \\t"
159	n/a	if you're comparing lines as sequences of characters, and don't
160	n/a	want to synch up on blanks or hard tabs.
161	n/a
162	n/a	Optional arg a is the first of two sequences to be compared. By
163	n/a	default, an empty string. The elements of a must be hashable. See
164	n/a	also .set_seqs() and .set_seq1().
165	n/a
166	n/a	Optional arg b is the second of two sequences to be compared. By
167	n/a	default, an empty string. The elements of b must be hashable. See
168	n/a	also .set_seqs() and .set_seq2().
169	n/a
170	n/a	Optional arg autojunk should be set to False to disable the
171	n/a	"automatic junk heuristic" that treats popular elements as junk
172	n/a	(see module documentation for more information).
173	n/a	"""
174	n/a
175	n/a	# Members:
176	n/a	# a
177	n/a	# first sequence
178	n/a	# b
179	n/a	# second sequence; differences are computed as "what do
180	n/a	# we need to do to 'a' to change it into 'b'?"
181	n/a	# b2j
182	n/a	# for x in b, b2j[x] is a list of the indices (into b)
183	n/a	# at which x appears; junk and popular elements do not appear
184	n/a	# fullbcount
185	n/a	# for x in b, fullbcount[x] == the number of times x
186	n/a	# appears in b; only materialized if really needed (used
187	n/a	# only for computing quick_ratio())
188	n/a	# matching_blocks
189	n/a	# a list of (i, j, k) triples, where a[i:i+k] == b[j:j+k];
190	n/a	# ascending & non-overlapping in i and in j; terminated by
191	n/a	# a dummy (len(a), len(b), 0) sentinel
192	n/a	# opcodes
193	n/a	# a list of (tag, i1, i2, j1, j2) tuples, where tag is
194	n/a	# one of
195	n/a	# 'replace' a[i1:i2] should be replaced by b[j1:j2]
196	n/a	# 'delete' a[i1:i2] should be deleted
197	n/a	# 'insert' b[j1:j2] should be inserted
198	n/a	# 'equal' a[i1:i2] == b[j1:j2]
199	n/a	# isjunk
200	n/a	# a user-supplied function taking a sequence element and
201	n/a	# returning true iff the element is "junk" -- this has
202	n/a	# subtle but helpful effects on the algorithm, which I'll
203	n/a	# get around to writing up someday <0.9 wink>.
204	n/a	# DON'T USE! Only __chain_b uses this. Use "in self.bjunk".
205	n/a	# bjunk
206	n/a	# the items in b for which isjunk is True.
207	n/a	# bpopular
208	n/a	# nonjunk items in b treated as junk by the heuristic (if used).
209	n/a
210	n/a	self.isjunk = isjunk
211	n/a	self.a = self.b = None
212	n/a	self.autojunk = autojunk
213	n/a	self.set_seqs(a, b)
214	n/a
215	n/a	def set_seqs(self, a, b):
216	n/a	"""Set the two sequences to be compared.
217	n/a
218	n/a	>>> s = SequenceMatcher()
219	n/a	>>> s.set_seqs("abcd", "bcde")
220	n/a	>>> s.ratio()
221	n/a	0.75
222	n/a	"""
223	n/a
224	n/a	self.set_seq1(a)
225	n/a	self.set_seq2(b)
226	n/a
227	n/a	def set_seq1(self, a):
228	n/a	"""Set the first sequence to be compared.
229	n/a
230	n/a	The second sequence to be compared is not changed.
231	n/a
232	n/a	>>> s = SequenceMatcher(None, "abcd", "bcde")
233	n/a	>>> s.ratio()
234	n/a	0.75
235	n/a	>>> s.set_seq1("bcde")
236	n/a	>>> s.ratio()
237	n/a	1.0
238	n/a	>>>
239	n/a
240	n/a	SequenceMatcher computes and caches detailed information about the
241	n/a	second sequence, so if you want to compare one sequence S against
242	n/a	many sequences, use .set_seq2(S) once and call .set_seq1(x)
243	n/a	repeatedly for each of the other sequences.
244	n/a
245	n/a	See also set_seqs() and set_seq2().
246	n/a	"""
247	n/a
248	n/a	if a is self.a:
249	n/a	return
250	n/a	self.a = a
251	n/a	self.matching_blocks = self.opcodes = None
252	n/a
253	n/a	def set_seq2(self, b):
254	n/a	"""Set the second sequence to be compared.
255	n/a
256	n/a	The first sequence to be compared is not changed.
257	n/a
258	n/a	>>> s = SequenceMatcher(None, "abcd", "bcde")
259	n/a	>>> s.ratio()
260	n/a	0.75
261	n/a	>>> s.set_seq2("abcd")
262	n/a	>>> s.ratio()
263	n/a	1.0
264	n/a	>>>
265	n/a
266	n/a	SequenceMatcher computes and caches detailed information about the
267	n/a	second sequence, so if you want to compare one sequence S against
268	n/a	many sequences, use .set_seq2(S) once and call .set_seq1(x)
269	n/a	repeatedly for each of the other sequences.
270	n/a
271	n/a	See also set_seqs() and set_seq1().
272	n/a	"""
273	n/a
274	n/a	if b is self.b:
275	n/a	return
276	n/a	self.b = b
277	n/a	self.matching_blocks = self.opcodes = None
278	n/a	self.fullbcount = None
279	n/a	self.__chain_b()
280	n/a
281	n/a	# For each element x in b, set b2j[x] to a list of the indices in
282	n/a	# b where x appears; the indices are in increasing order; note that
283	n/a	# the number of times x appears in b is len(b2j[x]) ...
284	n/a	# when self.isjunk is defined, junk elements don't show up in this
285	n/a	# map at all, which stops the central find_longest_match method
286	n/a	# from starting any matching block at a junk element ...
287	n/a	# b2j also does not contain entries for "popular" elements, meaning
288	n/a	# elements that account for more than 1 + 1% of the total elements, and
289	n/a	# when the sequence is reasonably large (>= 200 elements); this can
290	n/a	# be viewed as an adaptive notion of semi-junk, and yields an enormous
291	n/a	# speedup when, e.g., comparing program files with hundreds of
292	n/a	# instances of "return NULL;" ...
293	n/a	# note that this is only called when b changes; so for cross-product
294	n/a	# kinds of matches, it's best to call set_seq2 once, then set_seq1
295	n/a	# repeatedly
296	n/a
297	n/a	def __chain_b(self):
298	n/a	# Because isjunk is a user-defined (not C) function, and we test
299	n/a	# for junk a LOT, it's important to minimize the number of calls.
300	n/a	# Before the tricks described here, __chain_b was by far the most
301	n/a	# time-consuming routine in the whole module! If anyone sees
302	n/a	# Jim Roskind, thank him again for profile.py -- I never would
303	n/a	# have guessed that.
304	n/a	# The first trick is to build b2j ignoring the possibility
305	n/a	# of junk. I.e., we don't call isjunk at all yet. Throwing
306	n/a	# out the junk later is much cheaper than building b2j "right"
307	n/a	# from the start.
308	n/a	b = self.b
309	n/a	self.b2j = b2j = {}
310	n/a
311	n/a	for i, elt in enumerate(b):
312	n/a	indices = b2j.setdefault(elt, [])
313	n/a	indices.append(i)
314	n/a
315	n/a	# Purge junk elements
316	n/a	self.bjunk = junk = set()
317	n/a	isjunk = self.isjunk
318	n/a	if isjunk:
319	n/a	for elt in b2j.keys():
320	n/a	if isjunk(elt):
321	n/a	junk.add(elt)
322	n/a	for elt in junk: # separate loop avoids separate list of keys
323	n/a	del b2j[elt]
324	n/a
325	n/a	# Purge popular elements that are not junk
326	n/a	self.bpopular = popular = set()
327	n/a	n = len(b)
328	n/a	if self.autojunk and n >= 200:
329	n/a	ntest = n // 100 + 1
330	n/a	for elt, idxs in b2j.items():
331	n/a	if len(idxs) > ntest:
332	n/a	popular.add(elt)
333	n/a	for elt in popular: # ditto; as fast for 1% deletion
334	n/a	del b2j[elt]
335	n/a
336	n/a	def find_longest_match(self, alo, ahi, blo, bhi):
337	n/a	"""Find longest matching block in a[alo:ahi] and b[blo:bhi].
338	n/a
339	n/a	If isjunk is not defined:
340	n/a
341	n/a	Return (i,j,k) such that a[i:i+k] is equal to b[j:j+k], where
342	n/a	alo <= i <= i+k <= ahi
343	n/a	blo <= j <= j+k <= bhi
344	n/a	and for all (i',j',k') meeting those conditions,
345	n/a	k >= k'
346	n/a	i <= i'
347	n/a	and if i == i', j <= j'
348	n/a
349	n/a	In other words, of all maximal matching blocks, return one that
350	n/a	starts earliest in a, and of all those maximal matching blocks that
351	n/a	start earliest in a, return the one that starts earliest in b.
352	n/a
353	n/a	>>> s = SequenceMatcher(None, " abcd", "abcd abcd")
354	n/a	>>> s.find_longest_match(0, 5, 0, 9)
355	n/a	Match(a=0, b=4, size=5)
356	n/a
357	n/a	If isjunk is defined, first the longest matching block is
358	n/a	determined as above, but with the additional restriction that no
359	n/a	junk element appears in the block. Then that block is extended as
360	n/a	far as possible by matching (only) junk elements on both sides. So
361	n/a	the resulting block never matches on junk except as identical junk
362	n/a	happens to be adjacent to an "interesting" match.
363	n/a
364	n/a	Here's the same example as before, but considering blanks to be
365	n/a	junk. That prevents " abcd" from matching the " abcd" at the tail
366	n/a	end of the second sequence directly. Instead only the "abcd" can
367	n/a	match, and matches the leftmost "abcd" in the second sequence:
368	n/a
369	n/a	>>> s = SequenceMatcher(lambda x: x==" ", " abcd", "abcd abcd")
370	n/a	>>> s.find_longest_match(0, 5, 0, 9)
371	n/a	Match(a=1, b=0, size=4)
372	n/a
373	n/a	If no blocks match, return (alo, blo, 0).
374	n/a
375	n/a	>>> s = SequenceMatcher(None, "ab", "c")
376	n/a	>>> s.find_longest_match(0, 2, 0, 1)
377	n/a	Match(a=0, b=0, size=0)
378	n/a	"""
379	n/a
380	n/a	# CAUTION: stripping common prefix or suffix would be incorrect.
381	n/a	# E.g.,
382	n/a	# ab
383	n/a	# acab
384	n/a	# Longest matching block is "ab", but if common prefix is
385	n/a	# stripped, it's "a" (tied with "b"). UNIX(tm) diff does so
386	n/a	# strip, so ends up claiming that ab is changed to acab by
387	n/a	# inserting "ca" in the middle. That's minimal but unintuitive:
388	n/a	# "it's obvious" that someone inserted "ac" at the front.
389	n/a	# Windiff ends up at the same place as diff, but by pairing up
390	n/a	# the unique 'b's and then matching the first two 'a's.
391	n/a
392	n/a	a, b, b2j, isbjunk = self.a, self.b, self.b2j, self.bjunk.__contains__
393	n/a	besti, bestj, bestsize = alo, blo, 0
394	n/a	# find longest junk-free match
395	n/a	# during an iteration of the loop, j2len[j] = length of longest
396	n/a	# junk-free match ending with a[i-1] and b[j]
397	n/a	j2len = {}
398	n/a	nothing = []
399	n/a	for i in range(alo, ahi):
400	n/a	# look at all instances of a[i] in b; note that because
401	n/a	# b2j has no junk keys, the loop is skipped if a[i] is junk
402	n/a	j2lenget = j2len.get
403	n/a	newj2len = {}
404	n/a	for j in b2j.get(a[i], nothing):
405	n/a	# a[i] matches b[j]
406	n/a	if j < blo:
407	n/a	continue
408	n/a	if j >= bhi:
409	n/a	break
410	n/a	k = newj2len[j] = j2lenget(j-1, 0) + 1
411	n/a	if k > bestsize:
412	n/a	besti, bestj, bestsize = i-k+1, j-k+1, k
413	n/a	j2len = newj2len
414	n/a
415	n/a	# Extend the best by non-junk elements on each end. In particular,
416	n/a	# "popular" non-junk elements aren't in b2j, which greatly speeds
417	n/a	# the inner loop above, but also means "the best" match so far
418	n/a	# doesn't contain any junk or popular non-junk elements.
419	n/a	while besti > alo and bestj > blo and \
420	n/a	not isbjunk(b[bestj-1]) and \
421	n/a	a[besti-1] == b[bestj-1]:
422	n/a	besti, bestj, bestsize = besti-1, bestj-1, bestsize+1
423	n/a	while besti+bestsize < ahi and bestj+bestsize < bhi and \
424	n/a	not isbjunk(b[bestj+bestsize]) and \
425	n/a	a[besti+bestsize] == b[bestj+bestsize]:
426	n/a	bestsize += 1
427	n/a
428	n/a	# Now that we have a wholly interesting match (albeit possibly
429	n/a	# empty!), we may as well suck up the matching junk on each
430	n/a	# side of it too. Can't think of a good reason not to, and it
431	n/a	# saves post-processing the (possibly considerable) expense of
432	n/a	# figuring out what to do with it. In the case of an empty
433	n/a	# interesting match, this is clearly the right thing to do,
434	n/a	# because no other kind of match is possible in the regions.
435	n/a	while besti > alo and bestj > blo and \
436	n/a	isbjunk(b[bestj-1]) and \
437	n/a	a[besti-1] == b[bestj-1]:
438	n/a	besti, bestj, bestsize = besti-1, bestj-1, bestsize+1
439	n/a	while besti+bestsize < ahi and bestj+bestsize < bhi and \
440	n/a	isbjunk(b[bestj+bestsize]) and \
441	n/a	a[besti+bestsize] == b[bestj+bestsize]:
442	n/a	bestsize = bestsize + 1
443	n/a
444	n/a	return Match(besti, bestj, bestsize)
445	n/a
446	n/a	def get_matching_blocks(self):
447	n/a	"""Return list of triples describing matching subsequences.
448	n/a
449	n/a	Each triple is of the form (i, j, n), and means that
450	n/a	a[i:i+n] == b[j:j+n]. The triples are monotonically increasing in
451	n/a	i and in j. New in Python 2.5, it's also guaranteed that if
452	n/a	(i, j, n) and (i', j', n') are adjacent triples in the list, and
453	n/a	the second is not the last triple in the list, then i+n != i' or
454	n/a	j+n != j'. IOW, adjacent triples never describe adjacent equal
455	n/a	blocks.
456	n/a
457	n/a	The last triple is a dummy, (len(a), len(b), 0), and is the only
458	n/a	triple with n==0.
459	n/a
460	n/a	>>> s = SequenceMatcher(None, "abxcd", "abcd")
461	n/a	>>> list(s.get_matching_blocks())
462	n/a	[Match(a=0, b=0, size=2), Match(a=3, b=2, size=2), Match(a=5, b=4, size=0)]
463	n/a	"""
464	n/a
465	n/a	if self.matching_blocks is not None:
466	n/a	return self.matching_blocks
467	n/a	la, lb = len(self.a), len(self.b)
468	n/a
469	n/a	# This is most naturally expressed as a recursive algorithm, but
470	n/a	# at least one user bumped into extreme use cases that exceeded
471	n/a	# the recursion limit on their box. So, now we maintain a list
472	n/a	# ('queue`) of blocks we still need to look at, and append partial
473	n/a	# results to `matching_blocks` in a loop; the matches are sorted
474	n/a	# at the end.
475	n/a	queue = [(0, la, 0, lb)]
476	n/a	matching_blocks = []
477	n/a	while queue:
478	n/a	alo, ahi, blo, bhi = queue.pop()
479	n/a	i, j, k = x = self.find_longest_match(alo, ahi, blo, bhi)
480	n/a	# a[alo:i] vs b[blo:j] unknown
481	n/a	# a[i:i+k] same as b[j:j+k]
482	n/a	# a[i+k:ahi] vs b[j+k:bhi] unknown
483	n/a	if k: # if k is 0, there was no matching block
484	n/a	matching_blocks.append(x)
485	n/a	if alo < i and blo < j:
486	n/a	queue.append((alo, i, blo, j))
487	n/a	if i+k < ahi and j+k < bhi:
488	n/a	queue.append((i+k, ahi, j+k, bhi))
489	n/a	matching_blocks.sort()
490	n/a
491	n/a	# It's possible that we have adjacent equal blocks in the
492	n/a	# matching_blocks list now. Starting with 2.5, this code was added
493	n/a	# to collapse them.
494	n/a	i1 = j1 = k1 = 0
495	n/a	non_adjacent = []
496	n/a	for i2, j2, k2 in matching_blocks:
497	n/a	# Is this block adjacent to i1, j1, k1?
498	n/a	if i1 + k1 == i2 and j1 + k1 == j2:
499	n/a	# Yes, so collapse them -- this just increases the length of
500	n/a	# the first block by the length of the second, and the first
501	n/a	# block so lengthened remains the block to compare against.
502	n/a	k1 += k2
503	n/a	else:
504	n/a	# Not adjacent. Remember the first block (k1==0 means it's
505	n/a	# the dummy we started with), and make the second block the
506	n/a	# new block to compare against.
507	n/a	if k1:
508	n/a	non_adjacent.append((i1, j1, k1))
509	n/a	i1, j1, k1 = i2, j2, k2
510	n/a	if k1:
511	n/a	non_adjacent.append((i1, j1, k1))
512	n/a
513	n/a	non_adjacent.append( (la, lb, 0) )
514	n/a	self.matching_blocks = list(map(Match._make, non_adjacent))
515	n/a	return self.matching_blocks
516	n/a
517	n/a	def get_opcodes(self):
518	n/a	"""Return list of 5-tuples describing how to turn a into b.
519	n/a
520	n/a	Each tuple is of the form (tag, i1, i2, j1, j2). The first tuple
521	n/a	has i1 == j1 == 0, and remaining tuples have i1 == the i2 from the
522	n/a	tuple preceding it, and likewise for j1 == the previous j2.
523	n/a
524	n/a	The tags are strings, with these meanings:
525	n/a
526	n/a	'replace': a[i1:i2] should be replaced by b[j1:j2]
527	n/a	'delete': a[i1:i2] should be deleted.
528	n/a	Note that j1==j2 in this case.
529	n/a	'insert': b[j1:j2] should be inserted at a[i1:i1].
530	n/a	Note that i1==i2 in this case.
531	n/a	'equal': a[i1:i2] == b[j1:j2]
532	n/a
533	n/a	>>> a = "qabxcd"
534	n/a	>>> b = "abycdf"
535	n/a	>>> s = SequenceMatcher(None, a, b)
536	n/a	>>> for tag, i1, i2, j1, j2 in s.get_opcodes():
537	n/a	... print(("%7s a[%d:%d] (%s) b[%d:%d] (%s)" %
538	n/a	... (tag, i1, i2, a[i1:i2], j1, j2, b[j1:j2])))
539	n/a	delete a[0:1] (q) b[0:0] ()
540	n/a	equal a[1:3] (ab) b[0:2] (ab)
541	n/a	replace a[3:4] (x) b[2:3] (y)
542	n/a	equal a[4:6] (cd) b[3:5] (cd)
543	n/a	insert a[6:6] () b[5:6] (f)
544	n/a	"""
545	n/a
546	n/a	if self.opcodes is not None:
547	n/a	return self.opcodes
548	n/a	i = j = 0
549	n/a	self.opcodes = answer = []
550	n/a	for ai, bj, size in self.get_matching_blocks():
551	n/a	# invariant: we've pumped out correct diffs to change
552	n/a	# a[:i] into b[:j], and the next matching block is
553	n/a	# a[ai:ai+size] == b[bj:bj+size]. So we need to pump
554	n/a	# out a diff to change a[i:ai] into b[j:bj], pump out
555	n/a	# the matching block, and move (i,j) beyond the match
556	n/a	tag = ''
557	n/a	if i < ai and j < bj:
558	n/a	tag = 'replace'
559	n/a	elif i < ai:
560	n/a	tag = 'delete'
561	n/a	elif j < bj:
562	n/a	tag = 'insert'
563	n/a	if tag:
564	n/a	answer.append( (tag, i, ai, j, bj) )
565	n/a	i, j = ai+size, bj+size
566	n/a	# the list of matching blocks is terminated by a
567	n/a	# sentinel with size 0
568	n/a	if size:
569	n/a	answer.append( ('equal', ai, i, bj, j) )
570	n/a	return answer
571	n/a
572	n/a	def get_grouped_opcodes(self, n=3):
573	n/a	""" Isolate change clusters by eliminating ranges with no changes.
574	n/a
575	n/a	Return a generator of groups with up to n lines of context.
576	n/a	Each group is in the same format as returned by get_opcodes().
577	n/a
578	n/a	>>> from pprint import pprint
579	n/a	>>> a = list(map(str, range(1,40)))
580	n/a	>>> b = a[:]
581	n/a	>>> b[8:8] = ['i'] # Make an insertion
582	n/a	>>> b[20] += 'x' # Make a replacement
583	n/a	>>> b[23:28] = [] # Make a deletion
584	n/a	>>> b[30] += 'y' # Make another replacement
585	n/a	>>> pprint(list(SequenceMatcher(None,a,b).get_grouped_opcodes()))
586	n/a	[[('equal', 5, 8, 5, 8), ('insert', 8, 8, 8, 9), ('equal', 8, 11, 9, 12)],
587	n/a	[('equal', 16, 19, 17, 20),
588	n/a	('replace', 19, 20, 20, 21),
589	n/a	('equal', 20, 22, 21, 23),
590	n/a	('delete', 22, 27, 23, 23),
591	n/a	('equal', 27, 30, 23, 26)],
592	n/a	[('equal', 31, 34, 27, 30),
593	n/a	('replace', 34, 35, 30, 31),
594	n/a	('equal', 35, 38, 31, 34)]]
595	n/a	"""
596	n/a
597	n/a	codes = self.get_opcodes()
598	n/a	if not codes:
599	n/a	codes = [("equal", 0, 1, 0, 1)]
600	n/a	# Fixup leading and trailing groups if they show no changes.
601	n/a	if codes[0][0] == 'equal':
602	n/a	tag, i1, i2, j1, j2 = codes[0]
603	n/a	codes[0] = tag, max(i1, i2-n), i2, max(j1, j2-n), j2
604	n/a	if codes[-1][0] == 'equal':
605	n/a	tag, i1, i2, j1, j2 = codes[-1]
606	n/a	codes[-1] = tag, i1, min(i2, i1+n), j1, min(j2, j1+n)
607	n/a
608	n/a	nn = n + n
609	n/a	group = []
610	n/a	for tag, i1, i2, j1, j2 in codes:
611	n/a	# End the current group and start a new one whenever
612	n/a	# there is a large range with no changes.
613	n/a	if tag == 'equal' and i2-i1 > nn:
614	n/a	group.append((tag, i1, min(i2, i1+n), j1, min(j2, j1+n)))
615	n/a	yield group
616	n/a	group = []
617	n/a	i1, j1 = max(i1, i2-n), max(j1, j2-n)
618	n/a	group.append((tag, i1, i2, j1 ,j2))
619	n/a	if group and not (len(group)==1 and group[0][0] == 'equal'):
620	n/a	yield group
621	n/a
622	n/a	def ratio(self):
623	n/a	"""Return a measure of the sequences' similarity (float in [0,1]).
624	n/a
625	n/a	Where T is the total number of elements in both sequences, and
626	n/a	M is the number of matches, this is 2.0*M / T.
627	n/a	Note that this is 1 if the sequences are identical, and 0 if
628	n/a	they have nothing in common.
629	n/a
630	n/a	.ratio() is expensive to compute if you haven't already computed
631	n/a	.get_matching_blocks() or .get_opcodes(), in which case you may
632	n/a	want to try .quick_ratio() or .real_quick_ratio() first to get an
633	n/a	upper bound.
634	n/a
635	n/a	>>> s = SequenceMatcher(None, "abcd", "bcde")
636	n/a	>>> s.ratio()
637	n/a	0.75
638	n/a	>>> s.quick_ratio()
639	n/a	0.75
640	n/a	>>> s.real_quick_ratio()
641	n/a	1.0
642	n/a	"""
643	n/a
644	n/a	matches = sum(triple[-1] for triple in self.get_matching_blocks())
645	n/a	return _calculate_ratio(matches, len(self.a) + len(self.b))
646	n/a
647	n/a	def quick_ratio(self):
648	n/a	"""Return an upper bound on ratio() relatively quickly.
649	n/a
650	n/a	This isn't defined beyond that it is an upper bound on .ratio(), and
651	n/a	is faster to compute.
652	n/a	"""
653	n/a
654	n/a	# viewing a and b as multisets, set matches to the cardinality
655	n/a	# of their intersection; this counts the number of matches
656	n/a	# without regard to order, so is clearly an upper bound
657	n/a	if self.fullbcount is None:
658	n/a	self.fullbcount = fullbcount = {}
659	n/a	for elt in self.b:
660	n/a	fullbcount[elt] = fullbcount.get(elt, 0) + 1
661	n/a	fullbcount = self.fullbcount
662	n/a	# avail[x] is the number of times x appears in 'b' less the
663	n/a	# number of times we've seen it in 'a' so far ... kinda
664	n/a	avail = {}
665	n/a	availhas, matches = avail.__contains__, 0
666	n/a	for elt in self.a:
667	n/a	if availhas(elt):
668	n/a	numb = avail[elt]
669	n/a	else:
670	n/a	numb = fullbcount.get(elt, 0)
671	n/a	avail[elt] = numb - 1
672	n/a	if numb > 0:
673	n/a	matches = matches + 1
674	n/a	return _calculate_ratio(matches, len(self.a) + len(self.b))
675	n/a
676	n/a	def real_quick_ratio(self):
677	n/a	"""Return an upper bound on ratio() very quickly.
678	n/a
679	n/a	This isn't defined beyond that it is an upper bound on .ratio(), and
680	n/a	is faster to compute than either .ratio() or .quick_ratio().
681	n/a	"""
682	n/a
683	n/a	la, lb = len(self.a), len(self.b)
684	n/a	# can't have more matches than the number of elements in the
685	n/a	# shorter sequence
686	n/a	return _calculate_ratio(min(la, lb), la + lb)
687	n/a
688	n/a	def get_close_matches(word, possibilities, n=3, cutoff=0.6):
689	n/a	"""Use SequenceMatcher to return list of the best "good enough" matches.
690	n/a
691	n/a	word is a sequence for which close matches are desired (typically a
692	n/a	string).
693	n/a
694	n/a	possibilities is a list of sequences against which to match word
695	n/a	(typically a list of strings).
696	n/a
697	n/a	Optional arg n (default 3) is the maximum number of close matches to
698	n/a	return. n must be > 0.
699	n/a
700	n/a	Optional arg cutoff (default 0.6) is a float in [0, 1]. Possibilities
701	n/a	that don't score at least that similar to word are ignored.
702	n/a
703	n/a	The best (no more than n) matches among the possibilities are returned
704	n/a	in a list, sorted by similarity score, most similar first.
705	n/a
706	n/a	>>> get_close_matches("appel", ["ape", "apple", "peach", "puppy"])
707	n/a	['apple', 'ape']
708	n/a	>>> import keyword as _keyword
709	n/a	>>> get_close_matches("wheel", _keyword.kwlist)
710	n/a	['while']
711	n/a	>>> get_close_matches("Apple", _keyword.kwlist)
712	n/a	[]
713	n/a	>>> get_close_matches("accept", _keyword.kwlist)
714	n/a	['except']
715	n/a	"""
716	n/a
717	n/a	if not n > 0:
718	n/a	raise ValueError("n must be > 0: %r" % (n,))
719	n/a	if not 0.0 <= cutoff <= 1.0:
720	n/a	raise ValueError("cutoff must be in [0.0, 1.0]: %r" % (cutoff,))
721	n/a	result = []
722	n/a	s = SequenceMatcher()
723	n/a	s.set_seq2(word)
724	n/a	for x in possibilities:
725	n/a	s.set_seq1(x)
726	n/a	if s.real_quick_ratio() >= cutoff and \
727	n/a	s.quick_ratio() >= cutoff and \
728	n/a	s.ratio() >= cutoff:
729	n/a	result.append((s.ratio(), x))
730	n/a
731	n/a	# Move the best scorers to head of list
732	n/a	result = _nlargest(n, result)
733	n/a	# Strip scores for the best n matches
734	n/a	return [x for score, x in result]
735	n/a
736	n/a	def _count_leading(line, ch):
737	n/a	"""
738	n/a	Return number of `ch` characters at the start of `line`.
739	n/a
740	n/a	Example:
741	n/a
742	n/a	>>> _count_leading(' abc', ' ')
743	n/a	3
744	n/a	"""
745	n/a
746	n/a	i, n = 0, len(line)
747	n/a	while i < n and line[i] == ch:
748	n/a	i += 1
749	n/a	return i
750	n/a
751	n/a	class Differ:
752	n/a	r"""
753	n/a	Differ is a class for comparing sequences of lines of text, and
754	n/a	producing human-readable differences or deltas. Differ uses
755	n/a	SequenceMatcher both to compare sequences of lines, and to compare
756	n/a	sequences of characters within similar (near-matching) lines.
757	n/a
758	n/a	Each line of a Differ delta begins with a two-letter code:
759	n/a
760	n/a	'- ' line unique to sequence 1
761	n/a	'+ ' line unique to sequence 2
762	n/a	' ' line common to both sequences
763	n/a	'? ' line not present in either input sequence
764	n/a
765	n/a	Lines beginning with '? ' attempt to guide the eye to intraline
766	n/a	differences, and were not present in either input sequence. These lines
767	n/a	can be confusing if the sequences contain tab characters.
768	n/a
769	n/a	Note that Differ makes no claim to produce a minimal diff. To the
770	n/a	contrary, minimal diffs are often counter-intuitive, because they synch
771	n/a	up anywhere possible, sometimes accidental matches 100 pages apart.
772	n/a	Restricting synch points to contiguous matches preserves some notion of
773	n/a	locality, at the occasional cost of producing a longer diff.
774	n/a
775	n/a	Example: Comparing two texts.
776	n/a
777	n/a	First we set up the texts, sequences of individual single-line strings
778	n/a	ending with newlines (such sequences can also be obtained from the
779	n/a	`readlines()` method of file-like objects):
780	n/a
781	n/a	>>> text1 = ''' 1. Beautiful is better than ugly.
782	n/a	... 2. Explicit is better than implicit.
783	n/a	... 3. Simple is better than complex.
784	n/a	... 4. Complex is better than complicated.
785	n/a	... '''.splitlines(keepends=True)
786	n/a	>>> len(text1)
787	n/a	4
788	n/a	>>> text1[0][-1]
789	n/a	'\n'
790	n/a	>>> text2 = ''' 1. Beautiful is better than ugly.
791	n/a	... 3. Simple is better than complex.
792	n/a	... 4. Complicated is better than complex.
793	n/a	... 5. Flat is better than nested.
794	n/a	... '''.splitlines(keepends=True)
795	n/a
796	n/a	Next we instantiate a Differ object:
797	n/a
798	n/a	>>> d = Differ()
799	n/a
800	n/a	Note that when instantiating a Differ object we may pass functions to
801	n/a	filter out line and character 'junk'. See Differ.__init__ for details.
802	n/a
803	n/a	Finally, we compare the two:
804	n/a
805	n/a	>>> result = list(d.compare(text1, text2))
806	n/a
807	n/a	'result' is a list of strings, so let's pretty-print it:
808	n/a
809	n/a	>>> from pprint import pprint as _pprint
810	n/a	>>> _pprint(result)
811	n/a	[' 1. Beautiful is better than ugly.\n',
812	n/a	'- 2. Explicit is better than implicit.\n',
813	n/a	'- 3. Simple is better than complex.\n',
814	n/a	'+ 3. Simple is better than complex.\n',
815	n/a	'? ++\n',
816	n/a	'- 4. Complex is better than complicated.\n',
817	n/a	'? ^ ---- ^\n',
818	n/a	'+ 4. Complicated is better than complex.\n',
819	n/a	'? ++++ ^ ^\n',
820	n/a	'+ 5. Flat is better than nested.\n']
821	n/a
822	n/a	As a single multi-line string it looks like this:
823	n/a
824	n/a	>>> print(''.join(result), end="")
825	n/a	1. Beautiful is better than ugly.
826	n/a	- 2. Explicit is better than implicit.
827	n/a	- 3. Simple is better than complex.
828	n/a	+ 3. Simple is better than complex.
829	n/a	? ++
830	n/a	- 4. Complex is better than complicated.
831	n/a	? ^ ---- ^
832	n/a	+ 4. Complicated is better than complex.
833	n/a	? ++++ ^ ^
834	n/a	+ 5. Flat is better than nested.
835	n/a
836	n/a	Methods:
837	n/a
838	n/a	__init__(linejunk=None, charjunk=None)
839	n/a	Construct a text differencer, with optional filters.
840	n/a
841	n/a	compare(a, b)
842	n/a	Compare two sequences of lines; generate the resulting delta.
843	n/a	"""
844	n/a
845	n/a	def __init__(self, linejunk=None, charjunk=None):
846	n/a	"""
847	n/a	Construct a text differencer, with optional filters.
848	n/a
849	n/a	The two optional keyword parameters are for filter functions:
850	n/a
851	n/a	- `linejunk`: A function that should accept a single string argument,
852	n/a	and return true iff the string is junk. The module-level function
853	n/a	`IS_LINE_JUNK` may be used to filter out lines without visible
854	n/a	characters, except for at most one splat ('#'). It is recommended
855	n/a	to leave linejunk None; the underlying SequenceMatcher class has
856	n/a	an adaptive notion of "noise" lines that's better than any static
857	n/a	definition the author has ever been able to craft.
858	n/a
859	n/a	- `charjunk`: A function that should accept a string of length 1. The
860	n/a	module-level function `IS_CHARACTER_JUNK` may be used to filter out
861	n/a	whitespace characters (a blank or tab; note: bad idea to include
862	n/a	newline in this!). Use of IS_CHARACTER_JUNK is recommended.
863	n/a	"""
864	n/a
865	n/a	self.linejunk = linejunk
866	n/a	self.charjunk = charjunk
867	n/a
868	n/a	def compare(self, a, b):
869	n/a	r"""
870	n/a	Compare two sequences of lines; generate the resulting delta.
871	n/a
872	n/a	Each sequence must contain individual single-line strings ending with
873	n/a	newlines. Such sequences can be obtained from the `readlines()` method
874	n/a	of file-like objects. The delta generated also consists of newline-
875	n/a	terminated strings, ready to be printed as-is via the writeline()
876	n/a	method of a file-like object.
877	n/a
878	n/a	Example:
879	n/a
880	n/a	>>> print(''.join(Differ().compare('one\ntwo\nthree\n'.splitlines(True),
881	n/a	... 'ore\ntree\nemu\n'.splitlines(True))),
882	n/a	... end="")
883	n/a	- one
884	n/a	? ^
885	n/a	+ ore
886	n/a	? ^
887	n/a	- two
888	n/a	- three
889	n/a	? -
890	n/a	+ tree
891	n/a	+ emu
892	n/a	"""
893	n/a
894	n/a	cruncher = SequenceMatcher(self.linejunk, a, b)
895	n/a	for tag, alo, ahi, blo, bhi in cruncher.get_opcodes():
896	n/a	if tag == 'replace':
897	n/a	g = self._fancy_replace(a, alo, ahi, b, blo, bhi)
898	n/a	elif tag == 'delete':
899	n/a	g = self._dump('-', a, alo, ahi)
900	n/a	elif tag == 'insert':
901	n/a	g = self._dump('+', b, blo, bhi)
902	n/a	elif tag == 'equal':
903	n/a	g = self._dump(' ', a, alo, ahi)
904	n/a	else:
905	n/a	raise ValueError('unknown tag %r' % (tag,))
906	n/a
907	n/a	yield from g
908	n/a
909	n/a	def _dump(self, tag, x, lo, hi):
910	n/a	"""Generate comparison results for a same-tagged range."""
911	n/a	for i in range(lo, hi):
912	n/a	yield '%s %s' % (tag, x[i])
913	n/a
914	n/a	def _plain_replace(self, a, alo, ahi, b, blo, bhi):
915	n/a	assert alo < ahi and blo < bhi
916	n/a	# dump the shorter block first -- reduces the burden on short-term
917	n/a	# memory if the blocks are of very different sizes
918	n/a	if bhi - blo < ahi - alo:
919	n/a	first = self._dump('+', b, blo, bhi)
920	n/a	second = self._dump('-', a, alo, ahi)
921	n/a	else:
922	n/a	first = self._dump('-', a, alo, ahi)
923	n/a	second = self._dump('+', b, blo, bhi)
924	n/a
925	n/a	for g in first, second:
926	n/a	yield from g
927	n/a
928	n/a	def _fancy_replace(self, a, alo, ahi, b, blo, bhi):
929	n/a	r"""
930	n/a	When replacing one block of lines with another, search the blocks
931	n/a	for similar lines; the best-matching pair (if any) is used as a
932	n/a	synch point, and intraline difference marking is done on the
933	n/a	similar pair. Lots of work, but often worth it.
934	n/a
935	n/a	Example:
936	n/a
937	n/a	>>> d = Differ()
938	n/a	>>> results = d._fancy_replace(['abcDefghiJkl\n'], 0, 1,
939	n/a	... ['abcdefGhijkl\n'], 0, 1)
940	n/a	>>> print(''.join(results), end="")
941	n/a	- abcDefghiJkl
942	n/a	? ^ ^ ^
943	n/a	+ abcdefGhijkl
944	n/a	? ^ ^ ^
945	n/a	"""
946	n/a
947	n/a	# don't synch up unless the lines have a similarity score of at
948	n/a	# least cutoff; best_ratio tracks the best score seen so far
949	n/a	best_ratio, cutoff = 0.74, 0.75
950	n/a	cruncher = SequenceMatcher(self.charjunk)
951	n/a	eqi, eqj = None, None # 1st indices of equal lines (if any)
952	n/a
953	n/a	# search for the pair that matches best without being identical
954	n/a	# (identical lines must be junk lines, & we don't want to synch up
955	n/a	# on junk -- unless we have to)
956	n/a	for j in range(blo, bhi):
957	n/a	bj = b[j]
958	n/a	cruncher.set_seq2(bj)
959	n/a	for i in range(alo, ahi):
960	n/a	ai = a[i]
961	n/a	if ai == bj:
962	n/a	if eqi is None:
963	n/a	eqi, eqj = i, j
964	n/a	continue
965	n/a	cruncher.set_seq1(ai)
966	n/a	# computing similarity is expensive, so use the quick
967	n/a	# upper bounds first -- have seen this speed up messy
968	n/a	# compares by a factor of 3.
969	n/a	# note that ratio() is only expensive to compute the first
970	n/a	# time it's called on a sequence pair; the expensive part
971	n/a	# of the computation is cached by cruncher
972	n/a	if cruncher.real_quick_ratio() > best_ratio and \
973	n/a	cruncher.quick_ratio() > best_ratio and \
974	n/a	cruncher.ratio() > best_ratio:
975	n/a	best_ratio, best_i, best_j = cruncher.ratio(), i, j
976	n/a	if best_ratio < cutoff:
977	n/a	# no non-identical "pretty close" pair
978	n/a	if eqi is None:
979	n/a	# no identical pair either -- treat it as a straight replace
980	n/a	yield from self._plain_replace(a, alo, ahi, b, blo, bhi)
981	n/a	return
982	n/a	# no close pair, but an identical pair -- synch up on that
983	n/a	best_i, best_j, best_ratio = eqi, eqj, 1.0
984	n/a	else:
985	n/a	# there's a close pair, so forget the identical pair (if any)
986	n/a	eqi = None
987	n/a
988	n/a	# a[best_i] very similar to b[best_j]; eqi is None iff they're not
989	n/a	# identical
990	n/a
991	n/a	# pump out diffs from before the synch point
992	n/a	yield from self._fancy_helper(a, alo, best_i, b, blo, best_j)
993	n/a
994	n/a	# do intraline marking on the synch pair
995	n/a	aelt, belt = a[best_i], b[best_j]
996	n/a	if eqi is None:
997	n/a	# pump out a '-', '?', '+', '?' quad for the synched lines
998	n/a	atags = btags = ""
999	n/a	cruncher.set_seqs(aelt, belt)
1000	n/a	for tag, ai1, ai2, bj1, bj2 in cruncher.get_opcodes():
1001	n/a	la, lb = ai2 - ai1, bj2 - bj1
1002	n/a	if tag == 'replace':
1003	n/a	atags += '^' * la
1004	n/a	btags += '^' * lb
1005	n/a	elif tag == 'delete':
1006	n/a	atags += '-' * la
1007	n/a	elif tag == 'insert':
1008	n/a	btags += '+' * lb
1009	n/a	elif tag == 'equal':
1010	n/a	atags += ' ' * la
1011	n/a	btags += ' ' * lb
1012	n/a	else:
1013	n/a	raise ValueError('unknown tag %r' % (tag,))
1014	n/a	yield from self._qformat(aelt, belt, atags, btags)
1015	n/a	else:
1016	n/a	# the synch pair is identical
1017	n/a	yield ' ' + aelt
1018	n/a
1019	n/a	# pump out diffs from after the synch point
1020	n/a	yield from self._fancy_helper(a, best_i+1, ahi, b, best_j+1, bhi)
1021	n/a
1022	n/a	def _fancy_helper(self, a, alo, ahi, b, blo, bhi):
1023	n/a	g = []
1024	n/a	if alo < ahi:
1025	n/a	if blo < bhi:
1026	n/a	g = self._fancy_replace(a, alo, ahi, b, blo, bhi)
1027	n/a	else:
1028	n/a	g = self._dump('-', a, alo, ahi)
1029	n/a	elif blo < bhi:
1030	n/a	g = self._dump('+', b, blo, bhi)
1031	n/a
1032	n/a	yield from g
1033	n/a
1034	n/a	def _qformat(self, aline, bline, atags, btags):
1035	n/a	r"""
1036	n/a	Format "?" output and deal with leading tabs.
1037	n/a
1038	n/a	Example:
1039	n/a
1040	n/a	>>> d = Differ()
1041	n/a	>>> results = d._qformat('\tabcDefghiJkl\n', '\tabcdefGhijkl\n',
1042	n/a	... ' ^ ^ ^ ', ' ^ ^ ^ ')
1043	n/a	>>> for line in results: print(repr(line))
1044	n/a	...
1045	n/a	'- \tabcDefghiJkl\n'
1046	n/a	'? \t ^ ^ ^\n'
1047	n/a	'+ \tabcdefGhijkl\n'
1048	n/a	'? \t ^ ^ ^\n'
1049	n/a	"""
1050	n/a
1051	n/a	# Can hurt, but will probably help most of the time.
1052	n/a	common = min(_count_leading(aline, "\t"),
1053	n/a	_count_leading(bline, "\t"))
1054	n/a	common = min(common, _count_leading(atags[:common], " "))
1055	n/a	common = min(common, _count_leading(btags[:common], " "))
1056	n/a	atags = atags[common:].rstrip()
1057	n/a	btags = btags[common:].rstrip()
1058	n/a
1059	n/a	yield "- " + aline
1060	n/a	if atags:
1061	n/a	yield "? %s%s\n" % ("\t" * common, atags)
1062	n/a
1063	n/a	yield "+ " + bline
1064	n/a	if btags:
1065	n/a	yield "? %s%s\n" % ("\t" * common, btags)
1066	n/a
1067	n/a	# With respect to junk, an earlier version of ndiff simply refused to
1068	n/a	# start a match with a junk element. The result was cases like this:
1069	n/a	# before: private Thread currentThread;
1070	n/a	# after: private volatile Thread currentThread;
1071	n/a	# If you consider whitespace to be junk, the longest contiguous match
1072	n/a	# not starting with junk is "e Thread currentThread". So ndiff reported
1073	n/a	# that "e volatil" was inserted between the 't' and the 'e' in "private".
1074	n/a	# While an accurate view, to people that's absurd. The current version
1075	n/a	# looks for matching blocks that are entirely junk-free, then extends the
1076	n/a	# longest one of those as far as possible but only with matching junk.
1077	n/a	# So now "currentThread" is matched, then extended to suck up the
1078	n/a	# preceding blank; then "private" is matched, and extended to suck up the
1079	n/a	# following blank; then "Thread" is matched; and finally ndiff reports
1080	n/a	# that "volatile " was inserted before "Thread". The only quibble
1081	n/a	# remaining is that perhaps it was really the case that " volatile"
1082	n/a	# was inserted after "private". I can live with that <wink>.
1083	n/a
1084	n/a	import re
1085	n/a
1086	n/a	def IS_LINE_JUNK(line, pat=re.compile(r"\s#?\s$").match):
1087	n/a	r"""
1088	n/a	Return 1 for ignorable line: iff `line` is blank or contains a single '#'.
1089	n/a
1090	n/a	Examples:
1091	n/a
1092	n/a	>>> IS_LINE_JUNK('\n')
1093	n/a	True
1094	n/a	>>> IS_LINE_JUNK(' # \n')
1095	n/a	True
1096	n/a	>>> IS_LINE_JUNK('hello\n')
1097	n/a	False
1098	n/a	"""
1099	n/a
1100	n/a	return pat(line) is not None
1101	n/a
1102	n/a	def IS_CHARACTER_JUNK(ch, ws=" \t"):
1103	n/a	r"""
1104	n/a	Return 1 for ignorable character: iff `ch` is a space or tab.
1105	n/a
1106	n/a	Examples:
1107	n/a
1108	n/a	>>> IS_CHARACTER_JUNK(' ')
1109	n/a	True
1110	n/a	>>> IS_CHARACTER_JUNK('\t')
1111	n/a	True
1112	n/a	>>> IS_CHARACTER_JUNK('\n')
1113	n/a	False
1114	n/a	>>> IS_CHARACTER_JUNK('x')
1115	n/a	False
1116	n/a	"""
1117	n/a
1118	n/a	return ch in ws
1119	n/a
1120	n/a
1121	n/a	########################################################################
1122	n/a	### Unified Diff
1123	n/a	########################################################################
1124	n/a
1125	n/a	def _format_range_unified(start, stop):
1126	n/a	'Convert range to the "ed" format'
1127	n/a	# Per the diff spec at http://www.unix.org/single_unix_specification/
1128	n/a	beginning = start + 1 # lines start numbering with one
1129	n/a	length = stop - start
1130	n/a	if length == 1:
1131	n/a	return '{}'.format(beginning)
1132	n/a	if not length:
1133	n/a	beginning -= 1 # empty ranges begin at line just before the range
1134	n/a	return '{},{}'.format(beginning, length)
1135	n/a
1136	n/a	def unified_diff(a, b, fromfile='', tofile='', fromfiledate='',
1137	n/a	tofiledate='', n=3, lineterm='\n'):
1138	n/a	r"""
1139	n/a	Compare two sequences of lines; generate the delta as a unified diff.
1140	n/a
1141	n/a	Unified diffs are a compact way of showing line changes and a few
1142	n/a	lines of context. The number of context lines is set by 'n' which
1143	n/a	defaults to three.
1144	n/a
1145	n/a	By default, the diff control lines (those with ---, +++, or @@) are
1146	n/a	created with a trailing newline. This is helpful so that inputs
1147	n/a	created from file.readlines() result in diffs that are suitable for
1148	n/a	file.writelines() since both the inputs and outputs have trailing
1149	n/a	newlines.
1150	n/a
1151	n/a	For inputs that do not have trailing newlines, set the lineterm
1152	n/a	argument to "" so that the output will be uniformly newline free.
1153	n/a
1154	n/a	The unidiff format normally has a header for filenames and modification
1155	n/a	times. Any or all of these may be specified using strings for
1156	n/a	'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'.
1157	n/a	The modification times are normally expressed in the ISO 8601 format.
1158	n/a
1159	n/a	Example:
1160	n/a
1161	n/a	>>> for line in unified_diff('one two three four'.split(),
1162	n/a	... 'zero one tree four'.split(), 'Original', 'Current',
1163	n/a	... '2005-01-26 23:30:50', '2010-04-02 10:20:52',
1164	n/a	... lineterm=''):
1165	n/a	... print(line) # doctest: +NORMALIZE_WHITESPACE
1166	n/a	--- Original 2005-01-26 23:30:50
1167	n/a	+++ Current 2010-04-02 10:20:52
1168	n/a	@@ -1,4 +1,4 @@
1169	n/a	+zero
1170	n/a	one
1171	n/a	-two
1172	n/a	-three
1173	n/a	+tree
1174	n/a	four
1175	n/a	"""
1176	n/a
1177	n/a	_check_types(a, b, fromfile, tofile, fromfiledate, tofiledate, lineterm)
1178	n/a	started = False
1179	n/a	for group in SequenceMatcher(None,a,b).get_grouped_opcodes(n):
1180	n/a	if not started:
1181	n/a	started = True
1182	n/a	fromdate = '\t{}'.format(fromfiledate) if fromfiledate else ''
1183	n/a	todate = '\t{}'.format(tofiledate) if tofiledate else ''
1184	n/a	yield '--- {}{}{}'.format(fromfile, fromdate, lineterm)
1185	n/a	yield '+++ {}{}{}'.format(tofile, todate, lineterm)
1186	n/a
1187	n/a	first, last = group[0], group[-1]
1188	n/a	file1_range = _format_range_unified(first[1], last[2])
1189	n/a	file2_range = _format_range_unified(first[3], last[4])
1190	n/a	yield '@@ -{} +{} @@{}'.format(file1_range, file2_range, lineterm)
1191	n/a
1192	n/a	for tag, i1, i2, j1, j2 in group:
1193	n/a	if tag == 'equal':
1194	n/a	for line in a[i1:i2]:
1195	n/a	yield ' ' + line
1196	n/a	continue
1197	n/a	if tag in {'replace', 'delete'}:
1198	n/a	for line in a[i1:i2]:
1199	n/a	yield '-' + line
1200	n/a	if tag in {'replace', 'insert'}:
1201	n/a	for line in b[j1:j2]:
1202	n/a	yield '+' + line
1203	n/a
1204	n/a
1205	n/a	########################################################################
1206	n/a	### Context Diff
1207	n/a	########################################################################
1208	n/a
1209	n/a	def _format_range_context(start, stop):
1210	n/a	'Convert range to the "ed" format'
1211	n/a	# Per the diff spec at http://www.unix.org/single_unix_specification/
1212	n/a	beginning = start + 1 # lines start numbering with one
1213	n/a	length = stop - start
1214	n/a	if not length:
1215	n/a	beginning -= 1 # empty ranges begin at line just before the range
1216	n/a	if length <= 1:
1217	n/a	return '{}'.format(beginning)
1218	n/a	return '{},{}'.format(beginning, beginning + length - 1)
1219	n/a
1220	n/a	# See http://www.unix.org/single_unix_specification/
1221	n/a	def context_diff(a, b, fromfile='', tofile='',
1222	n/a	fromfiledate='', tofiledate='', n=3, lineterm='\n'):
1223	n/a	r"""
1224	n/a	Compare two sequences of lines; generate the delta as a context diff.
1225	n/a
1226	n/a	Context diffs are a compact way of showing line changes and a few
1227	n/a	lines of context. The number of context lines is set by 'n' which
1228	n/a	defaults to three.
1229	n/a
1230	n/a	By default, the diff control lines (those with *** or ---) are
1231	n/a	created with a trailing newline. This is helpful so that inputs
1232	n/a	created from file.readlines() result in diffs that are suitable for
1233	n/a	file.writelines() since both the inputs and outputs have trailing
1234	n/a	newlines.
1235	n/a
1236	n/a	For inputs that do not have trailing newlines, set the lineterm
1237	n/a	argument to "" so that the output will be uniformly newline free.
1238	n/a
1239	n/a	The context diff format normally has a header for filenames and
1240	n/a	modification times. Any or all of these may be specified using
1241	n/a	strings for 'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'.
1242	n/a	The modification times are normally expressed in the ISO 8601 format.
1243	n/a	If not specified, the strings default to blanks.
1244	n/a
1245	n/a	Example:
1246	n/a
1247	n/a	>>> print(''.join(context_diff('one\ntwo\nthree\nfour\n'.splitlines(True),
1248	n/a	... 'zero\none\ntree\nfour\n'.splitlines(True), 'Original', 'Current')),
1249	n/a	... end="")
1250	n/a	*** Original
1251	n/a	--- Current
1252	n/a	***************
1253	n/a	* 1,4 **
1254	n/a	one
1255	n/a	! two
1256	n/a	! three
1257	n/a	four
1258	n/a	--- 1,4 ----
1259	n/a	+ zero
1260	n/a	one
1261	n/a	! tree
1262	n/a	four
1263	n/a	"""
1264	n/a
1265	n/a	_check_types(a, b, fromfile, tofile, fromfiledate, tofiledate, lineterm)
1266	n/a	prefix = dict(insert='+ ', delete='- ', replace='! ', equal=' ')
1267	n/a	started = False
1268	n/a	for group in SequenceMatcher(None,a,b).get_grouped_opcodes(n):
1269	n/a	if not started:
1270	n/a	started = True
1271	n/a	fromdate = '\t{}'.format(fromfiledate) if fromfiledate else ''
1272	n/a	todate = '\t{}'.format(tofiledate) if tofiledate else ''
1273	n/a	yield '*** {}{}{}'.format(fromfile, fromdate, lineterm)
1274	n/a	yield '--- {}{}{}'.format(tofile, todate, lineterm)
1275	n/a
1276	n/a	first, last = group[0], group[-1]
1277	n/a	yield '***************' + lineterm
1278	n/a
1279	n/a	file1_range = _format_range_context(first[1], last[2])
1280	n/a	yield '* {} **{}'.format(file1_range, lineterm)
1281	n/a
1282	n/a	if any(tag in {'replace', 'delete'} for tag, _, _, _, _ in group):
1283	n/a	for tag, i1, i2, _, _ in group:
1284	n/a	if tag != 'insert':
1285	n/a	for line in a[i1:i2]:
1286	n/a	yield prefix[tag] + line
1287	n/a
1288	n/a	file2_range = _format_range_context(first[3], last[4])
1289	n/a	yield '--- {} ----{}'.format(file2_range, lineterm)
1290	n/a
1291	n/a	if any(tag in {'replace', 'insert'} for tag, _, _, _, _ in group):
1292	n/a	for tag, _, _, j1, j2 in group:
1293	n/a	if tag != 'delete':
1294	n/a	for line in b[j1:j2]:
1295	n/a	yield prefix[tag] + line
1296	n/a
1297	n/a	def _check_types(a, b, *args):
1298	n/a	# Checking types is weird, but the alternative is garbled output when
1299	n/a	# someone passes mixed bytes and str to {unified,context}_diff(). E.g.
1300	n/a	# without this check, passing filenames as bytes results in output like
1301	n/a	# --- b'oldfile.txt'
1302	n/a	# +++ b'newfile.txt'
1303	n/a	# because of how str.format() incorporates bytes objects.
1304	n/a	if a and not isinstance(a[0], str):
1305	n/a	raise TypeError('lines to compare must be str, not %s (%r)' %
1306	n/a	(type(a[0]).__name__, a[0]))
1307	n/a	if b and not isinstance(b[0], str):
1308	n/a	raise TypeError('lines to compare must be str, not %s (%r)' %
1309	n/a	(type(b[0]).__name__, b[0]))
1310	n/a	for arg in args:
1311	n/a	if not isinstance(arg, str):
1312	n/a	raise TypeError('all arguments must be str, not: %r' % (arg,))
1313	n/a
1314	n/a	def diff_bytes(dfunc, a, b, fromfile=b'', tofile=b'',
1315	n/a	fromfiledate=b'', tofiledate=b'', n=3, lineterm=b'\n'):
1316	n/a	r"""
1317	n/a	Compare `a` and `b`, two sequences of lines represented as bytes rather
1318	n/a	than str. This is a wrapper for `dfunc`, which is typically either
1319	n/a	unified_diff() or context_diff(). Inputs are losslessly converted to
1320	n/a	strings so that `dfunc` only has to worry about strings, and encoded
1321	n/a	back to bytes on return. This is necessary to compare files with
1322	n/a	unknown or inconsistent encoding. All other inputs (except `n`) must be
1323	n/a	bytes rather than str.
1324	n/a	"""
1325	n/a	def decode(s):
1326	n/a	try:
1327	n/a	return s.decode('ascii', 'surrogateescape')
1328	n/a	except AttributeError as err:
1329	n/a	msg = ('all arguments must be bytes, not %s (%r)' %
1330	n/a	(type(s).__name__, s))
1331	n/a	raise TypeError(msg) from err
1332	n/a	a = list(map(decode, a))
1333	n/a	b = list(map(decode, b))
1334	n/a	fromfile = decode(fromfile)
1335	n/a	tofile = decode(tofile)
1336	n/a	fromfiledate = decode(fromfiledate)
1337	n/a	tofiledate = decode(tofiledate)
1338	n/a	lineterm = decode(lineterm)
1339	n/a
1340	n/a	lines = dfunc(a, b, fromfile, tofile, fromfiledate, tofiledate, n, lineterm)
1341	n/a	for line in lines:
1342	n/a	yield line.encode('ascii', 'surrogateescape')
1343	n/a
1344	n/a	def ndiff(a, b, linejunk=None, charjunk=IS_CHARACTER_JUNK):
1345	n/a	r"""
1346	n/a	Compare `a` and `b` (lists of strings); return a `Differ`-style delta.
1347	n/a
1348	n/a	Optional keyword parameters `linejunk` and `charjunk` are for filter
1349	n/a	functions, or can be None:
1350	n/a
1351	n/a	- linejunk: A function that should accept a single string argument and
1352	n/a	return true iff the string is junk. The default is None, and is
1353	n/a	recommended; the underlying SequenceMatcher class has an adaptive
1354	n/a	notion of "noise" lines.
1355	n/a
1356	n/a	- charjunk: A function that accepts a character (string of length
1357	n/a	1), and returns true iff the character is junk. The default is
1358	n/a	the module-level function IS_CHARACTER_JUNK, which filters out
1359	n/a	whitespace characters (a blank or tab; note: it's a bad idea to
1360	n/a	include newline in this!).
1361	n/a
1362	n/a	Tools/scripts/ndiff.py is a command-line front-end to this function.
1363	n/a
1364	n/a	Example:
1365	n/a
1366	n/a	>>> diff = ndiff('one\ntwo\nthree\n'.splitlines(keepends=True),
1367	n/a	... 'ore\ntree\nemu\n'.splitlines(keepends=True))
1368	n/a	>>> print(''.join(diff), end="")
1369	n/a	- one
1370	n/a	? ^
1371	n/a	+ ore
1372	n/a	? ^
1373	n/a	- two
1374	n/a	- three
1375	n/a	? -
1376	n/a	+ tree
1377	n/a	+ emu
1378	n/a	"""
1379	n/a	return Differ(linejunk, charjunk).compare(a, b)
1380	n/a
1381	n/a	def _mdiff(fromlines, tolines, context=None, linejunk=None,
1382	n/a	charjunk=IS_CHARACTER_JUNK):
1383	n/a	r"""Returns generator yielding marked up from/to side by side differences.
1384	n/a
1385	n/a	Arguments:
1386	n/a	fromlines -- list of text lines to compared to tolines
1387	n/a	tolines -- list of text lines to be compared to fromlines
1388	n/a	context -- number of context lines to display on each side of difference,
1389	n/a	if None, all from/to text lines will be generated.
1390	n/a	linejunk -- passed on to ndiff (see ndiff documentation)
1391	n/a	charjunk -- passed on to ndiff (see ndiff documentation)
1392	n/a
1393	n/a	This function returns an iterator which returns a tuple:
1394	n/a	(from line tuple, to line tuple, boolean flag)
1395	n/a
1396	n/a	from/to line tuple -- (line num, line text)
1397	n/a	line num -- integer or None (to indicate a context separation)
1398	n/a	line text -- original line text with following markers inserted:
1399	n/a	'\0+' -- marks start of added text
1400	n/a	'\0-' -- marks start of deleted text
1401	n/a	'\0^' -- marks start of changed text
1402	n/a	'\1' -- marks end of added/deleted/changed text
1403	n/a
1404	n/a	boolean flag -- None indicates context separation, True indicates
1405	n/a	either "from" or "to" line contains a change, otherwise False.
1406	n/a
1407	n/a	This function/iterator was originally developed to generate side by side
1408	n/a	file difference for making HTML pages (see HtmlDiff class for example
1409	n/a	usage).
1410	n/a
1411	n/a	Note, this function utilizes the ndiff function to generate the side by
1412	n/a	side difference markup. Optional ndiff arguments may be passed to this
1413	n/a	function and they in turn will be passed to ndiff.
1414	n/a	"""
1415	n/a	import re
1416	n/a
1417	n/a	# regular expression for finding intraline change indices
1418	n/a	change_re = re.compile(r'(\++\|\-+\|\^+)')
1419	n/a
1420	n/a	# create the difference iterator to generate the differences
1421	n/a	diff_lines_iterator = ndiff(fromlines,tolines,linejunk,charjunk)
1422	n/a
1423	n/a	def _make_line(lines, format_key, side, num_lines=[0,0]):
1424	n/a	"""Returns line of text with user's change markup and line formatting.
1425	n/a
1426	n/a	lines -- list of lines from the ndiff generator to produce a line of
1427	n/a	text from. When producing the line of text to return, the
1428	n/a	lines used are removed from this list.
1429	n/a	format_key -- '+' return first line in list with "add" markup around
1430	n/a	the entire line.
1431	n/a	'-' return first line in list with "delete" markup around
1432	n/a	the entire line.
1433	n/a	'?' return first line in list with add/delete/change
1434	n/a	intraline markup (indices obtained from second line)
1435	n/a	None return first line in list with no markup
1436	n/a	side -- indice into the num_lines list (0=from,1=to)
1437	n/a	num_lines -- from/to current line number. This is NOT intended to be a
1438	n/a	passed parameter. It is present as a keyword argument to
1439	n/a	maintain memory of the current line numbers between calls
1440	n/a	of this function.
1441	n/a
1442	n/a	Note, this function is purposefully not defined at the module scope so
1443	n/a	that data it needs from its parent function (within whose context it
1444	n/a	is defined) does not need to be of module scope.
1445	n/a	"""
1446	n/a	num_lines[side] += 1
1447	n/a	# Handle case where no user markup is to be added, just return line of
1448	n/a	# text with user's line format to allow for usage of the line number.
1449	n/a	if format_key is None:
1450	n/a	return (num_lines[side],lines.pop(0)[2:])
1451	n/a	# Handle case of intraline changes
1452	n/a	if format_key == '?':
1453	n/a	text, markers = lines.pop(0), lines.pop(0)
1454	n/a	# find intraline changes (store change type and indices in tuples)
1455	n/a	sub_info = []
1456	n/a	def record_sub_info(match_object,sub_info=sub_info):
1457	n/a	sub_info.append([match_object.group(1)[0],match_object.span()])
1458	n/a	return match_object.group(1)
1459	n/a	change_re.sub(record_sub_info,markers)
1460	n/a	# process each tuple inserting our special marks that won't be
1461	n/a	# noticed by an xml/html escaper.
1462	n/a	for key,(begin,end) in reversed(sub_info):
1463	n/a	text = text[0:begin]+'\0'+key+text[begin:end]+'\1'+text[end:]
1464	n/a	text = text[2:]
1465	n/a	# Handle case of add/delete entire line
1466	n/a	else:
1467	n/a	text = lines.pop(0)[2:]
1468	n/a	# if line of text is just a newline, insert a space so there is
1469	n/a	# something for the user to highlight and see.
1470	n/a	if not text:
1471	n/a	text = ' '
1472	n/a	# insert marks that won't be noticed by an xml/html escaper.
1473	n/a	text = '\0' + format_key + text + '\1'
1474	n/a	# Return line of text, first allow user's line formatter to do its
1475	n/a	# thing (such as adding the line number) then replace the special
1476	n/a	# marks with what the user's change markup.
1477	n/a	return (num_lines[side],text)
1478	n/a
1479	n/a	def _line_iterator():
1480	n/a	"""Yields from/to lines of text with a change indication.
1481	n/a
1482	n/a	This function is an iterator. It itself pulls lines from a
1483	n/a	differencing iterator, processes them and yields them. When it can
1484	n/a	it yields both a "from" and a "to" line, otherwise it will yield one
1485	n/a	or the other. In addition to yielding the lines of from/to text, a
1486	n/a	boolean flag is yielded to indicate if the text line(s) have
1487	n/a	differences in them.
1488	n/a
1489	n/a	Note, this function is purposefully not defined at the module scope so
1490	n/a	that data it needs from its parent function (within whose context it
1491	n/a	is defined) does not need to be of module scope.
1492	n/a	"""
1493	n/a	lines = []
1494	n/a	num_blanks_pending, num_blanks_to_yield = 0, 0
1495	n/a	while True:
1496	n/a	# Load up next 4 lines so we can look ahead, create strings which
1497	n/a	# are a concatenation of the first character of each of the 4 lines
1498	n/a	# so we can do some very readable comparisons.
1499	n/a	while len(lines) < 4:
1500	n/a	lines.append(next(diff_lines_iterator, 'X'))
1501	n/a	s = ''.join([line[0] for line in lines])
1502	n/a	if s.startswith('X'):
1503	n/a	# When no more lines, pump out any remaining blank lines so the
1504	n/a	# corresponding add/delete lines get a matching blank line so
1505	n/a	# all line pairs get yielded at the next level.
1506	n/a	num_blanks_to_yield = num_blanks_pending
1507	n/a	elif s.startswith('-?+?'):
1508	n/a	# simple intraline change
1509	n/a	yield _make_line(lines,'?',0), _make_line(lines,'?',1), True
1510	n/a	continue
1511	n/a	elif s.startswith('--++'):
1512	n/a	# in delete block, add block coming: we do NOT want to get
1513	n/a	# caught up on blank lines yet, just process the delete line
1514	n/a	num_blanks_pending -= 1
1515	n/a	yield _make_line(lines,'-',0), None, True
1516	n/a	continue
1517	n/a	elif s.startswith(('--?+', '--+', '- ')):
1518	n/a	# in delete block and see an intraline change or unchanged line
1519	n/a	# coming: yield the delete line and then blanks
1520	n/a	from_line,to_line = _make_line(lines,'-',0), None
1521	n/a	num_blanks_to_yield,num_blanks_pending = num_blanks_pending-1,0
1522	n/a	elif s.startswith('-+?'):
1523	n/a	# intraline change
1524	n/a	yield _make_line(lines,None,0), _make_line(lines,'?',1), True
1525	n/a	continue
1526	n/a	elif s.startswith('-?+'):
1527	n/a	# intraline change
1528	n/a	yield _make_line(lines,'?',0), _make_line(lines,None,1), True
1529	n/a	continue
1530	n/a	elif s.startswith('-'):
1531	n/a	# delete FROM line
1532	n/a	num_blanks_pending -= 1
1533	n/a	yield _make_line(lines,'-',0), None, True
1534	n/a	continue
1535	n/a	elif s.startswith('+--'):
1536	n/a	# in add block, delete block coming: we do NOT want to get
1537	n/a	# caught up on blank lines yet, just process the add line
1538	n/a	num_blanks_pending += 1
1539	n/a	yield None, _make_line(lines,'+',1), True
1540	n/a	continue
1541	n/a	elif s.startswith(('+ ', '+-')):
1542	n/a	# will be leaving an add block: yield blanks then add line
1543	n/a	from_line, to_line = None, _make_line(lines,'+',1)
1544	n/a	num_blanks_to_yield,num_blanks_pending = num_blanks_pending+1,0
1545	n/a	elif s.startswith('+'):
1546	n/a	# inside an add block, yield the add line
1547	n/a	num_blanks_pending += 1
1548	n/a	yield None, _make_line(lines,'+',1), True
1549	n/a	continue
1550	n/a	elif s.startswith(' '):
1551	n/a	# unchanged text, yield it to both sides
1552	n/a	yield _make_line(lines[:],None,0),_make_line(lines,None,1),False
1553	n/a	continue
1554	n/a	# Catch up on the blank lines so when we yield the next from/to
1555	n/a	# pair, they are lined up.
1556	n/a	while(num_blanks_to_yield < 0):
1557	n/a	num_blanks_to_yield += 1
1558	n/a	yield None,('','\n'),True
1559	n/a	while(num_blanks_to_yield > 0):
1560	n/a	num_blanks_to_yield -= 1
1561	n/a	yield ('','\n'),None,True
1562	n/a	if s.startswith('X'):
1563	n/a	return
1564	n/a	else:
1565	n/a	yield from_line,to_line,True
1566	n/a
1567	n/a	def _line_pair_iterator():
1568	n/a	"""Yields from/to lines of text with a change indication.
1569	n/a
1570	n/a	This function is an iterator. It itself pulls lines from the line
1571	n/a	iterator. Its difference from that iterator is that this function
1572	n/a	always yields a pair of from/to text lines (with the change
1573	n/a	indication). If necessary it will collect single from/to lines
1574	n/a	until it has a matching pair from/to pair to yield.
1575	n/a
1576	n/a	Note, this function is purposefully not defined at the module scope so
1577	n/a	that data it needs from its parent function (within whose context it
1578	n/a	is defined) does not need to be of module scope.
1579	n/a	"""
1580	n/a	line_iterator = _line_iterator()
1581	n/a	fromlines,tolines=[],[]
1582	n/a	while True:
1583	n/a	# Collecting lines of text until we have a from/to pair
1584	n/a	while (len(fromlines)==0 or len(tolines)==0):
1585	n/a	try:
1586	n/a	from_line, to_line, found_diff = next(line_iterator)
1587	n/a	except StopIteration:
1588	n/a	return
1589	n/a	if from_line is not None:
1590	n/a	fromlines.append((from_line,found_diff))
1591	n/a	if to_line is not None:
1592	n/a	tolines.append((to_line,found_diff))
1593	n/a	# Once we have a pair, remove them from the collection and yield it
1594	n/a	from_line, fromDiff = fromlines.pop(0)
1595	n/a	to_line, to_diff = tolines.pop(0)
1596	n/a	yield (from_line,to_line,fromDiff or to_diff)
1597	n/a
1598	n/a	# Handle case where user does not want context differencing, just yield
1599	n/a	# them up without doing anything else with them.
1600	n/a	line_pair_iterator = _line_pair_iterator()
1601	n/a	if context is None:
1602	n/a	yield from line_pair_iterator
1603	n/a	# Handle case where user wants context differencing. We must do some
1604	n/a	# storage of lines until we know for sure that they are to be yielded.
1605	n/a	else:
1606	n/a	context += 1
1607	n/a	lines_to_write = 0
1608	n/a	while True:
1609	n/a	# Store lines up until we find a difference, note use of a
1610	n/a	# circular queue because we only need to keep around what
1611	n/a	# we need for context.
1612	n/a	index, contextLines = 0, [None]*(context)
1613	n/a	found_diff = False
1614	n/a	while(found_diff is False):
1615	n/a	try:
1616	n/a	from_line, to_line, found_diff = next(line_pair_iterator)
1617	n/a	except StopIteration:
1618	n/a	return
1619	n/a	i = index % context
1620	n/a	contextLines[i] = (from_line, to_line, found_diff)
1621	n/a	index += 1
1622	n/a	# Yield lines that we have collected so far, but first yield
1623	n/a	# the user's separator.
1624	n/a	if index > context:
1625	n/a	yield None, None, None
1626	n/a	lines_to_write = context
1627	n/a	else:
1628	n/a	lines_to_write = index
1629	n/a	index = 0
1630	n/a	while(lines_to_write):
1631	n/a	i = index % context
1632	n/a	index += 1
1633	n/a	yield contextLines[i]
1634	n/a	lines_to_write -= 1
1635	n/a	# Now yield the context lines after the change
1636	n/a	lines_to_write = context-1
1637	n/a	while(lines_to_write):
1638	n/a	from_line, to_line, found_diff = next(line_pair_iterator)
1639	n/a	# If another change within the context, extend the context
1640	n/a	if found_diff:
1641	n/a	lines_to_write = context-1
1642	n/a	else:
1643	n/a	lines_to_write -= 1
1644	n/a	yield from_line, to_line, found_diff
1645	n/a
1646	n/a
1647	n/a	_file_template = """
1648	n/a	<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
1649	n/a	"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
1650	n/a
1651	n/a	<html>
1652	n/a
1653	n/a	<head>
1654	n/a	<meta http-equiv="Content-Type"
1655	n/a	content="text/html; charset=%(charset)s" />
1656	n/a	<title></title>
1657	n/a	<style type="text/css">%(styles)s
1658	n/a	</style>
1659	n/a	</head>
1660	n/a
1661	n/a	<body>
1662	n/a	%(table)s%(legend)s
1663	n/a	</body>
1664	n/a
1665	n/a	</html>"""
1666	n/a
1667	n/a	_styles = """
1668	n/a	table.diff {font-family:Courier; border:medium;}
1669	n/a	.diff_header {background-color:#e0e0e0}
1670	n/a	td.diff_header {text-align:right}
1671	n/a	.diff_next {background-color:#c0c0c0}
1672	n/a	.diff_add {background-color:#aaffaa}
1673	n/a	.diff_chg {background-color:#ffff77}
1674	n/a	.diff_sub {background-color:#ffaaaa}"""
1675	n/a
1676	n/a	_table_template = """
1677	n/a	<table class="diff" id="difflib_chg_%(prefix)s_top"
1678	n/a	cellspacing="0" cellpadding="0" rules="groups" >
1679	n/a	<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>
1680	n/a	<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>
1681	n/a	%(header_row)s
1682	n/a	<tbody>
1683	n/a	%(data_rows)s </tbody>
1684	n/a	</table>"""
1685	n/a
1686	n/a	_legend = """
1687	n/a	<table class="diff" summary="Legends">
1688	n/a	<tr> <th colspan="2"> Legends </th> </tr>
1689	n/a	<tr> <td> <table border="" summary="Colors">
1690	n/a	<tr><th> Colors </th> </tr>
1691	n/a	<tr><td class="diff_add"> Added </td></tr>
1692	n/a	<tr><td class="diff_chg">Changed</td> </tr>
1693	n/a	<tr><td class="diff_sub">Deleted</td> </tr>
1694	n/a	</table></td>
1695	n/a	<td> <table border="" summary="Links">
1696	n/a	<tr><th colspan="2"> Links </th> </tr>
1697	n/a	<tr><td>(f)irst change</td> </tr>
1698	n/a	<tr><td>(n)ext change</td> </tr>
1699	n/a	<tr><td>(t)op</td> </tr>
1700	n/a	</table></td> </tr>
1701	n/a	</table>"""
1702	n/a
1703	n/a	class HtmlDiff(object):
1704	n/a	"""For producing HTML side by side comparison with change highlights.
1705	n/a
1706	n/a	This class can be used to create an HTML table (or a complete HTML file
1707	n/a	containing the table) showing a side by side, line by line comparison
1708	n/a	of text with inter-line and intra-line change highlights. The table can
1709	n/a	be generated in either full or contextual difference mode.
1710	n/a
1711	n/a	The following methods are provided for HTML generation:
1712	n/a
1713	n/a	make_table -- generates HTML for a single side by side table
1714	n/a	make_file -- generates complete HTML file with a single side by side table
1715	n/a
1716	n/a	See tools/scripts/diff.py for an example usage of this class.
1717	n/a	"""
1718	n/a
1719	n/a	_file_template = _file_template
1720	n/a	_styles = _styles
1721	n/a	_table_template = _table_template
1722	n/a	_legend = _legend
1723	n/a	_default_prefix = 0
1724	n/a
1725	n/a	def __init__(self,tabsize=8,wrapcolumn=None,linejunk=None,
1726	n/a	charjunk=IS_CHARACTER_JUNK):
1727	n/a	"""HtmlDiff instance initializer
1728	n/a
1729	n/a	Arguments:
1730	n/a	tabsize -- tab stop spacing, defaults to 8.
1731	n/a	wrapcolumn -- column number where lines are broken and wrapped,
1732	n/a	defaults to None where lines are not wrapped.
1733	n/a	linejunk,charjunk -- keyword arguments passed into ndiff() (used by
1734	n/a	HtmlDiff() to generate the side by side HTML differences). See
1735	n/a	ndiff() documentation for argument default values and descriptions.
1736	n/a	"""
1737	n/a	self._tabsize = tabsize
1738	n/a	self._wrapcolumn = wrapcolumn
1739	n/a	self._linejunk = linejunk
1740	n/a	self._charjunk = charjunk
1741	n/a
1742	n/a	def make_file(self, fromlines, tolines, fromdesc='', todesc='',
1743	n/a	context=False, numlines=5, *, charset='utf-8'):
1744	n/a	"""Returns HTML file of side by side comparison with change highlights
1745	n/a
1746	n/a	Arguments:
1747	n/a	fromlines -- list of "from" lines
1748	n/a	tolines -- list of "to" lines
1749	n/a	fromdesc -- "from" file column header string
1750	n/a	todesc -- "to" file column header string
1751	n/a	context -- set to True for contextual differences (defaults to False
1752	n/a	which shows full differences).
1753	n/a	numlines -- number of context lines. When context is set True,
1754	n/a	controls number of lines displayed before and after the change.
1755	n/a	When context is False, controls the number of lines to place
1756	n/a	the "next" link anchors before the next change (so click of
1757	n/a	"next" link jumps to just before the change).
1758	n/a	charset -- charset of the HTML document
1759	n/a	"""
1760	n/a
1761	n/a	return (self._file_template % dict(
1762	n/a	styles=self._styles,
1763	n/a	legend=self._legend,
1764	n/a	table=self.make_table(fromlines, tolines, fromdesc, todesc,
1765	n/a	context=context, numlines=numlines),
1766	n/a	charset=charset
1767	n/a	)).encode(charset, 'xmlcharrefreplace').decode(charset)
1768	n/a
1769	n/a	def _tab_newline_replace(self,fromlines,tolines):
1770	n/a	"""Returns from/to line lists with tabs expanded and newlines removed.
1771	n/a
1772	n/a	Instead of tab characters being replaced by the number of spaces
1773	n/a	needed to fill in to the next tab stop, this function will fill
1774	n/a	the space with tab characters. This is done so that the difference
1775	n/a	algorithms can identify changes in a file when tabs are replaced by
1776	n/a	spaces and vice versa. At the end of the HTML generation, the tab
1777	n/a	characters will be replaced with a nonbreakable space.
1778	n/a	"""
1779	n/a	def expand_tabs(line):
1780	n/a	# hide real spaces
1781	n/a	line = line.replace(' ','\0')
1782	n/a	# expand tabs into spaces
1783	n/a	line = line.expandtabs(self._tabsize)
1784	n/a	# replace spaces from expanded tabs back into tab characters
1785	n/a	# (we'll replace them with markup after we do differencing)
1786	n/a	line = line.replace(' ','\t')
1787	n/a	return line.replace('\0',' ').rstrip('\n')
1788	n/a	fromlines = [expand_tabs(line) for line in fromlines]
1789	n/a	tolines = [expand_tabs(line) for line in tolines]
1790	n/a	return fromlines,tolines
1791	n/a
1792	n/a	def _split_line(self,data_list,line_num,text):
1793	n/a	"""Builds list of text lines by splitting text lines at wrap point
1794	n/a
1795	n/a	This function will determine if the input text line needs to be
1796	n/a	wrapped (split) into separate lines. If so, the first wrap point
1797	n/a	will be determined and the first line appended to the output
1798	n/a	text line list. This function is used recursively to handle
1799	n/a	the second part of the split line to further split it.
1800	n/a	"""
1801	n/a	# if blank line or context separator, just add it to the output list
1802	n/a	if not line_num:
1803	n/a	data_list.append((line_num,text))
1804	n/a	return
1805	n/a
1806	n/a	# if line text doesn't need wrapping, just add it to the output list
1807	n/a	size = len(text)
1808	n/a	max = self._wrapcolumn
1809	n/a	if (size <= max) or ((size -(text.count('\0')*3)) <= max):
1810	n/a	data_list.append((line_num,text))
1811	n/a	return
1812	n/a
1813	n/a	# scan text looking for the wrap point, keeping track if the wrap
1814	n/a	# point is inside markers
1815	n/a	i = 0
1816	n/a	n = 0
1817	n/a	mark = ''
1818	n/a	while n < max and i < size:
1819	n/a	if text[i] == '\0':
1820	n/a	i += 1
1821	n/a	mark = text[i]
1822	n/a	i += 1
1823	n/a	elif text[i] == '\1':
1824	n/a	i += 1
1825	n/a	mark = ''
1826	n/a	else:
1827	n/a	i += 1
1828	n/a	n += 1
1829	n/a
1830	n/a	# wrap point is inside text, break it up into separate lines
1831	n/a	line1 = text[:i]
1832	n/a	line2 = text[i:]
1833	n/a
1834	n/a	# if wrap point is inside markers, place end marker at end of first
1835	n/a	# line and start marker at beginning of second line because each
1836	n/a	# line will have its own table tag markup around it.
1837	n/a	if mark:
1838	n/a	line1 = line1 + '\1'
1839	n/a	line2 = '\0' + mark + line2
1840	n/a
1841	n/a	# tack on first line onto the output list
1842	n/a	data_list.append((line_num,line1))
1843	n/a
1844	n/a	# use this routine again to wrap the remaining text
1845	n/a	self._split_line(data_list,'>',line2)
1846	n/a
1847	n/a	def _line_wrapper(self,diffs):
1848	n/a	"""Returns iterator that splits (wraps) mdiff text lines"""
1849	n/a
1850	n/a	# pull from/to data and flags from mdiff iterator
1851	n/a	for fromdata,todata,flag in diffs:
1852	n/a	# check for context separators and pass them through
1853	n/a	if flag is None:
1854	n/a	yield fromdata,todata,flag
1855	n/a	continue
1856	n/a	(fromline,fromtext),(toline,totext) = fromdata,todata
1857	n/a	# for each from/to line split it at the wrap column to form
1858	n/a	# list of text lines.
1859	n/a	fromlist,tolist = [],[]
1860	n/a	self._split_line(fromlist,fromline,fromtext)
1861	n/a	self._split_line(tolist,toline,totext)
1862	n/a	# yield from/to line in pairs inserting blank lines as
1863	n/a	# necessary when one side has more wrapped lines
1864	n/a	while fromlist or tolist:
1865	n/a	if fromlist:
1866	n/a	fromdata = fromlist.pop(0)
1867	n/a	else:
1868	n/a	fromdata = ('',' ')
1869	n/a	if tolist:
1870	n/a	todata = tolist.pop(0)
1871	n/a	else:
1872	n/a	todata = ('',' ')
1873	n/a	yield fromdata,todata,flag
1874	n/a
1875	n/a	def _collect_lines(self,diffs):
1876	n/a	"""Collects mdiff output into separate lists
1877	n/a
1878	n/a	Before storing the mdiff from/to data into a list, it is converted
1879	n/a	into a single line of text with HTML markup.
1880	n/a	"""
1881	n/a
1882	n/a	fromlist,tolist,flaglist = [],[],[]
1883	n/a	# pull from/to data and flags from mdiff style iterator
1884	n/a	for fromdata,todata,flag in diffs:
1885	n/a	try:
1886	n/a	# store HTML markup of the lines into the lists
1887	n/a	fromlist.append(self._format_line(0,flag,*fromdata))
1888	n/a	tolist.append(self._format_line(1,flag,*todata))
1889	n/a	except TypeError:
1890	n/a	# exceptions occur for lines where context separators go
1891	n/a	fromlist.append(None)
1892	n/a	tolist.append(None)
1893	n/a	flaglist.append(flag)
1894	n/a	return fromlist,tolist,flaglist
1895	n/a
1896	n/a	def _format_line(self,side,flag,linenum,text):
1897	n/a	"""Returns HTML markup of "from" / "to" text lines
1898	n/a
1899	n/a	side -- 0 or 1 indicating "from" or "to" text
1900	n/a	flag -- indicates if difference on line
1901	n/a	linenum -- line number (used for line number column)
1902	n/a	text -- line text to be marked up
1903	n/a	"""
1904	n/a	try:
1905	n/a	linenum = '%d' % linenum
1906	n/a	id = ' id="%s%s"' % (self._prefix[side],linenum)
1907	n/a	except TypeError:
1908	n/a	# handle blank lines where linenum is '>' or ''
1909	n/a	id = ''
1910	n/a	# replace those things that would get confused with HTML symbols
1911	n/a	text=text.replace("&","&").replace(">",">").replace("<","<")
1912	n/a
1913	n/a	# make space non-breakable so they don't get compressed or line wrapped
1914	n/a	text = text.replace(' ',' ').rstrip()
1915	n/a
1916	n/a	return '<td class="diff_header"%s>%s</td><td nowrap="nowrap">%s</td>' \
1917	n/a	% (id,linenum,text)
1918	n/a
1919	n/a	def _make_prefix(self):
1920	n/a	"""Create unique anchor prefixes"""
1921	n/a
1922	n/a	# Generate a unique anchor prefix so multiple tables
1923	n/a	# can exist on the same HTML page without conflicts.
1924	n/a	fromprefix = "from%d_" % HtmlDiff._default_prefix
1925	n/a	toprefix = "to%d_" % HtmlDiff._default_prefix
1926	n/a	HtmlDiff._default_prefix += 1
1927	n/a	# store prefixes so line format method has access
1928	n/a	self._prefix = [fromprefix,toprefix]
1929	n/a
1930	n/a	def _convert_flags(self,fromlist,tolist,flaglist,context,numlines):
1931	n/a	"""Makes list of "next" links"""
1932	n/a
1933	n/a	# all anchor names will be generated using the unique "to" prefix
1934	n/a	toprefix = self._prefix[1]
1935	n/a
1936	n/a	# process change flags, generating middle column of next anchors/links
1937	n/a	next_id = ['']*len(flaglist)
1938	n/a	next_href = ['']*len(flaglist)
1939	n/a	num_chg, in_change = 0, False
1940	n/a	last = 0
1941	n/a	for i,flag in enumerate(flaglist):
1942	n/a	if flag:
1943	n/a	if not in_change:
1944	n/a	in_change = True
1945	n/a	last = i
1946	n/a	# at the beginning of a change, drop an anchor a few lines
1947	n/a	# (the context lines) before the change for the previous
1948	n/a	# link
1949	n/a	i = max([0,i-numlines])
1950	n/a	next_id[i] = ' id="difflib_chg_%s_%d"' % (toprefix,num_chg)
1951	n/a	# at the beginning of a change, drop a link to the next
1952	n/a	# change
1953	n/a	num_chg += 1
1954	n/a	next_href[last] = '<a href="#difflib_chg_%s_%d">n</a>' % (
1955	n/a	toprefix,num_chg)
1956	n/a	else:
1957	n/a	in_change = False
1958	n/a	# check for cases where there is no content to avoid exceptions
1959	n/a	if not flaglist:
1960	n/a	flaglist = [False]
1961	n/a	next_id = ['']
1962	n/a	next_href = ['']
1963	n/a	last = 0
1964	n/a	if context:
1965	n/a	fromlist = ['<td></td><td> No Differences Found </td>']
1966	n/a	tolist = fromlist
1967	n/a	else:
1968	n/a	fromlist = tolist = ['<td></td><td> Empty File </td>']
1969	n/a	# if not a change on first line, drop a link
1970	n/a	if not flaglist[0]:
1971	n/a	next_href[0] = '<a href="#difflib_chg_%s_0">f</a>' % toprefix
1972	n/a	# redo the last link to link to the top
1973	n/a	next_href[last] = '<a href="#difflib_chg_%s_top">t</a>' % (toprefix)
1974	n/a
1975	n/a	return fromlist,tolist,flaglist,next_href,next_id
1976	n/a
1977	n/a	def make_table(self,fromlines,tolines,fromdesc='',todesc='',context=False,
1978	n/a	numlines=5):
1979	n/a	"""Returns HTML table of side by side comparison with change highlights
1980	n/a
1981	n/a	Arguments:
1982	n/a	fromlines -- list of "from" lines
1983	n/a	tolines -- list of "to" lines
1984	n/a	fromdesc -- "from" file column header string
1985	n/a	todesc -- "to" file column header string
1986	n/a	context -- set to True for contextual differences (defaults to False
1987	n/a	which shows full differences).
1988	n/a	numlines -- number of context lines. When context is set True,
1989	n/a	controls number of lines displayed before and after the change.
1990	n/a	When context is False, controls the number of lines to place
1991	n/a	the "next" link anchors before the next change (so click of
1992	n/a	"next" link jumps to just before the change).
1993	n/a	"""
1994	n/a
1995	n/a	# make unique anchor prefixes so that multiple tables may exist
1996	n/a	# on the same page without conflict.
1997	n/a	self._make_prefix()
1998	n/a
1999	n/a	# change tabs to spaces before it gets more difficult after we insert
2000	n/a	# markup
2001	n/a	fromlines,tolines = self._tab_newline_replace(fromlines,tolines)
2002	n/a
2003	n/a	# create diffs iterator which generates side by side from/to data
2004	n/a	if context:
2005	n/a	context_lines = numlines
2006	n/a	else:
2007	n/a	context_lines = None
2008	n/a	diffs = _mdiff(fromlines,tolines,context_lines,linejunk=self._linejunk,
2009	n/a	charjunk=self._charjunk)
2010	n/a
2011	n/a	# set up iterator to wrap lines that exceed desired width
2012	n/a	if self._wrapcolumn:
2013	n/a	diffs = self._line_wrapper(diffs)
2014	n/a
2015	n/a	# collect up from/to lines and flags into lists (also format the lines)
2016	n/a	fromlist,tolist,flaglist = self._collect_lines(diffs)
2017	n/a
2018	n/a	# process change flags, generating middle column of next anchors/links
2019	n/a	fromlist,tolist,flaglist,next_href,next_id = self._convert_flags(
2020	n/a	fromlist,tolist,flaglist,context,numlines)
2021	n/a
2022	n/a	s = []
2023	n/a	fmt = ' <tr><td class="diff_next"%s>%s</td>%s' + \
2024	n/a	'<td class="diff_next">%s</td>%s</tr>\n'
2025	n/a	for i in range(len(flaglist)):
2026	n/a	if flaglist[i] is None:
2027	n/a	# mdiff yields None on separator lines skip the bogus ones
2028	n/a	# generated for the first line
2029	n/a	if i > 0:
2030	n/a	s.append(' </tbody> \n <tbody>\n')
2031	n/a	else:
2032	n/a	s.append( fmt % (next_id[i],next_href[i],fromlist[i],
2033	n/a	next_href[i],tolist[i]))
2034	n/a	if fromdesc or todesc:
2035	n/a	header_row = '<thead><tr>%s%s%s%s</tr></thead>' % (
2036	n/a	'<th class="diff_next"><br /></th>',
2037	n/a	'<th colspan="2" class="diff_header">%s</th>' % fromdesc,
2038	n/a	'<th class="diff_next"><br /></th>',
2039	n/a	'<th colspan="2" class="diff_header">%s</th>' % todesc)
2040	n/a	else:
2041	n/a	header_row = ''
2042	n/a
2043	n/a	table = self._table_template % dict(
2044	n/a	data_rows=''.join(s),
2045	n/a	header_row=header_row,
2046	n/a	prefix=self._prefix[1])
2047	n/a
2048	n/a	return table.replace('\0+','<span class="diff_add">'). \
2049	n/a	replace('\0-','<span class="diff_sub">'). \
2050	n/a	replace('\0^','<span class="diff_chg">'). \
2051	n/a	replace('\1','</span>'). \
2052	n/a	replace('\t',' ')
2053	n/a
2054	n/a	del re
2055	n/a
2056	n/a	def restore(delta, which):
2057	n/a	r"""
2058	n/a	Generate one of the two sequences that generated a delta.
2059	n/a
2060	n/a	Given a `delta` produced by `Differ.compare()` or `ndiff()`, extract
2061	n/a	lines originating from file 1 or 2 (parameter `which`), stripping off line
2062	n/a	prefixes.
2063	n/a
2064	n/a	Examples:
2065	n/a
2066	n/a	>>> diff = ndiff('one\ntwo\nthree\n'.splitlines(keepends=True),
2067	n/a	... 'ore\ntree\nemu\n'.splitlines(keepends=True))
2068	n/a	>>> diff = list(diff)
2069	n/a	>>> print(''.join(restore(diff, 1)), end="")
2070	n/a	one
2071	n/a	two
2072	n/a	three
2073	n/a	>>> print(''.join(restore(diff, 2)), end="")
2074	n/a	ore
2075	n/a	tree
2076	n/a	emu
2077	n/a	"""
2078	n/a	try:
2079	n/a	tag = {1: "- ", 2: "+ "}[int(which)]
2080	n/a	except KeyError:
2081	n/a	raise ValueError('unknown delta choice (must be 1 or 2): %r'
2082	n/a	% which)
2083	n/a	prefixes = (" ", tag)
2084	n/a	for line in delta:
2085	n/a	if line[:2] in prefixes:
2086	n/a	yield line[2:]
2087	n/a
2088	n/a	def _test():
2089	n/a	import doctest, difflib
2090	n/a	return doctest.testmod(difflib)
2091	n/a
2092	n/a	if __name__ == "__main__":
2093	n/a	_test()