mirror of
https://github.com/python/cpython.git
synced 2024-11-26 11:24:40 +08:00
81b9251d59
This now does a dynamic analysis of which elements are so frequently repeated as to constitute noise. The primary benefit is an enormous speedup in find_longest_match, as the innermost loop can have factors of 100s less potential matches to worry about, in cases where the sequences have many duplicate elements. In effect, this zooms in on sequences of non-ubiquitous elements now. While I like what I've seen of the effects so far, I still consider this experimental. Please give it a try!
1119 lines
41 KiB
Python
1119 lines
41 KiB
Python
#! /usr/bin/env python
|
|
|
|
"""
|
|
Module difflib -- helpers for computing deltas between objects.
|
|
|
|
Function get_close_matches(word, possibilities, n=3, cutoff=0.6):
|
|
Use SequenceMatcher to return list of the best "good enough" matches.
|
|
|
|
Function ndiff(a, b):
|
|
Return a delta: the difference between `a` and `b` (lists of strings).
|
|
|
|
Function restore(delta, which):
|
|
Return one of the two sequences that generated an ndiff delta.
|
|
|
|
Class SequenceMatcher:
|
|
A flexible class for comparing pairs of sequences of any type.
|
|
|
|
Class Differ:
|
|
For producing human-readable deltas from sequences of lines of text.
|
|
"""
|
|
|
|
__all__ = ['get_close_matches', 'ndiff', 'restore', 'SequenceMatcher',
|
|
'Differ']
|
|
|
|
class SequenceMatcher:
|
|
|
|
"""
|
|
SequenceMatcher is a flexible class for comparing pairs of sequences of
|
|
any type, so long as the sequence elements are hashable. The basic
|
|
algorithm predates, and is a little fancier than, an algorithm
|
|
published in the late 1980's by Ratcliff and Obershelp under the
|
|
hyperbolic name "gestalt pattern matching". The basic idea is to find
|
|
the longest contiguous matching subsequence that contains no "junk"
|
|
elements (R-O doesn't address junk). The same idea is then applied
|
|
recursively to the pieces of the sequences to the left and to the right
|
|
of the matching subsequence. This does not yield minimal edit
|
|
sequences, but does tend to yield matches that "look right" to people.
|
|
|
|
SequenceMatcher tries to compute a "human-friendly diff" between two
|
|
sequences. Unlike e.g. UNIX(tm) diff, the fundamental notion is the
|
|
longest *contiguous* & junk-free matching subsequence. That's what
|
|
catches peoples' eyes. The Windows(tm) windiff has another interesting
|
|
notion, pairing up elements that appear uniquely in each sequence.
|
|
That, and the method here, appear to yield more intuitive difference
|
|
reports than does diff. This method appears to be the least vulnerable
|
|
to synching up on blocks of "junk lines", though (like blank lines in
|
|
ordinary text files, or maybe "<P>" lines in HTML files). That may be
|
|
because this is the only method of the 3 that has a *concept* of
|
|
"junk" <wink>.
|
|
|
|
Example, comparing two strings, and considering blanks to be "junk":
|
|
|
|
>>> s = SequenceMatcher(lambda x: x == " ",
|
|
... "private Thread currentThread;",
|
|
... "private volatile Thread currentThread;")
|
|
>>>
|
|
|
|
.ratio() returns a float in [0, 1], measuring the "similarity" of the
|
|
sequences. As a rule of thumb, a .ratio() value over 0.6 means the
|
|
sequences are close matches:
|
|
|
|
>>> print round(s.ratio(), 3)
|
|
0.866
|
|
>>>
|
|
|
|
If you're only interested in where the sequences match,
|
|
.get_matching_blocks() is handy:
|
|
|
|
>>> for block in s.get_matching_blocks():
|
|
... print "a[%d] and b[%d] match for %d elements" % block
|
|
a[0] and b[0] match for 8 elements
|
|
a[8] and b[17] match for 6 elements
|
|
a[14] and b[23] match for 15 elements
|
|
a[29] and b[38] match for 0 elements
|
|
|
|
Note that the last tuple returned by .get_matching_blocks() is always a
|
|
dummy, (len(a), len(b), 0), and this is the only case in which the last
|
|
tuple element (number of elements matched) is 0.
|
|
|
|
If you want to know how to change the first sequence into the second,
|
|
use .get_opcodes():
|
|
|
|
>>> for opcode in s.get_opcodes():
|
|
... print "%6s a[%d:%d] b[%d:%d]" % opcode
|
|
equal a[0:8] b[0:8]
|
|
insert a[8:8] b[8:17]
|
|
equal a[8:14] b[17:23]
|
|
equal a[14:29] b[23:38]
|
|
|
|
See the Differ class for a fancy human-friendly file differencer, which
|
|
uses SequenceMatcher both to compare sequences of lines, and to compare
|
|
sequences of characters within similar (near-matching) lines.
|
|
|
|
See also function get_close_matches() in this module, which shows how
|
|
simple code building on SequenceMatcher can be used to do useful work.
|
|
|
|
Timing: Basic R-O is cubic time worst case and quadratic time expected
|
|
case. SequenceMatcher is quadratic time for the worst case and has
|
|
expected-case behavior dependent in a complicated way on how many
|
|
elements the sequences have in common; best case time is linear.
|
|
|
|
Methods:
|
|
|
|
__init__(isjunk=None, a='', b='')
|
|
Construct a SequenceMatcher.
|
|
|
|
set_seqs(a, b)
|
|
Set the two sequences to be compared.
|
|
|
|
set_seq1(a)
|
|
Set the first sequence to be compared.
|
|
|
|
set_seq2(b)
|
|
Set the second sequence to be compared.
|
|
|
|
find_longest_match(alo, ahi, blo, bhi)
|
|
Find longest matching block in a[alo:ahi] and b[blo:bhi].
|
|
|
|
get_matching_blocks()
|
|
Return list of triples describing matching subsequences.
|
|
|
|
get_opcodes()
|
|
Return list of 5-tuples describing how to turn a into b.
|
|
|
|
ratio()
|
|
Return a measure of the sequences' similarity (float in [0,1]).
|
|
|
|
quick_ratio()
|
|
Return an upper bound on .ratio() relatively quickly.
|
|
|
|
real_quick_ratio()
|
|
Return an upper bound on ratio() very quickly.
|
|
"""
|
|
|
|
def __init__(self, isjunk=None, a='', b=''):
|
|
"""Construct a SequenceMatcher.
|
|
|
|
Optional arg isjunk is None (the default), or a one-argument
|
|
function that takes a sequence element and returns true iff the
|
|
element is junk. None is equivalent to passing "lambda x: 0", i.e.
|
|
no elements are considered to be junk. For example, pass
|
|
lambda x: x in " \\t"
|
|
if you're comparing lines as sequences of characters, and don't
|
|
want to synch up on blanks or hard tabs.
|
|
|
|
Optional arg a is the first of two sequences to be compared. By
|
|
default, an empty string. The elements of a must be hashable. See
|
|
also .set_seqs() and .set_seq1().
|
|
|
|
Optional arg b is the second of two sequences to be compared. By
|
|
default, an empty string. The elements of b must be hashable. See
|
|
also .set_seqs() and .set_seq2().
|
|
"""
|
|
|
|
# Members:
|
|
# a
|
|
# first sequence
|
|
# b
|
|
# second sequence; differences are computed as "what do
|
|
# we need to do to 'a' to change it into 'b'?"
|
|
# b2j
|
|
# for x in b, b2j[x] is a list of the indices (into b)
|
|
# at which x appears; junk elements do not appear
|
|
# fullbcount
|
|
# for x in b, fullbcount[x] == the number of times x
|
|
# appears in b; only materialized if really needed (used
|
|
# only for computing quick_ratio())
|
|
# matching_blocks
|
|
# a list of (i, j, k) triples, where a[i:i+k] == b[j:j+k];
|
|
# ascending & non-overlapping in i and in j; terminated by
|
|
# a dummy (len(a), len(b), 0) sentinel
|
|
# opcodes
|
|
# a list of (tag, i1, i2, j1, j2) tuples, where tag is
|
|
# one of
|
|
# 'replace' a[i1:i2] should be replaced by b[j1:j2]
|
|
# 'delete' a[i1:i2] should be deleted
|
|
# 'insert' b[j1:j2] should be inserted
|
|
# 'equal' a[i1:i2] == b[j1:j2]
|
|
# isjunk
|
|
# a user-supplied function taking a sequence element and
|
|
# returning true iff the element is "junk" -- this has
|
|
# subtle but helpful effects on the algorithm, which I'll
|
|
# get around to writing up someday <0.9 wink>.
|
|
# DON'T USE! Only __chain_b uses this. Use isbjunk.
|
|
# isbjunk
|
|
# for x in b, isbjunk(x) == isjunk(x) but much faster;
|
|
# it's really the has_key method of a hidden dict.
|
|
# DOES NOT WORK for x in a!
|
|
# isbpopular
|
|
# for x in b, isbpopular(x) is true iff b is reasonably long
|
|
# (at least 200 elements) and x accounts for more than 1% of
|
|
# its elements. DOES NOT WORK for x in a!
|
|
|
|
self.isjunk = isjunk
|
|
self.a = self.b = None
|
|
self.set_seqs(a, b)
|
|
|
|
def set_seqs(self, a, b):
|
|
"""Set the two sequences to be compared.
|
|
|
|
>>> s = SequenceMatcher()
|
|
>>> s.set_seqs("abcd", "bcde")
|
|
>>> s.ratio()
|
|
0.75
|
|
"""
|
|
|
|
self.set_seq1(a)
|
|
self.set_seq2(b)
|
|
|
|
def set_seq1(self, a):
|
|
"""Set the first sequence to be compared.
|
|
|
|
The second sequence to be compared is not changed.
|
|
|
|
>>> s = SequenceMatcher(None, "abcd", "bcde")
|
|
>>> s.ratio()
|
|
0.75
|
|
>>> s.set_seq1("bcde")
|
|
>>> s.ratio()
|
|
1.0
|
|
>>>
|
|
|
|
SequenceMatcher computes and caches detailed information about the
|
|
second sequence, so if you want to compare one sequence S against
|
|
many sequences, use .set_seq2(S) once and call .set_seq1(x)
|
|
repeatedly for each of the other sequences.
|
|
|
|
See also set_seqs() and set_seq2().
|
|
"""
|
|
|
|
if a is self.a:
|
|
return
|
|
self.a = a
|
|
self.matching_blocks = self.opcodes = None
|
|
|
|
def set_seq2(self, b):
|
|
"""Set the second sequence to be compared.
|
|
|
|
The first sequence to be compared is not changed.
|
|
|
|
>>> s = SequenceMatcher(None, "abcd", "bcde")
|
|
>>> s.ratio()
|
|
0.75
|
|
>>> s.set_seq2("abcd")
|
|
>>> s.ratio()
|
|
1.0
|
|
>>>
|
|
|
|
SequenceMatcher computes and caches detailed information about the
|
|
second sequence, so if you want to compare one sequence S against
|
|
many sequences, use .set_seq2(S) once and call .set_seq1(x)
|
|
repeatedly for each of the other sequences.
|
|
|
|
See also set_seqs() and set_seq1().
|
|
"""
|
|
|
|
if b is self.b:
|
|
return
|
|
self.b = b
|
|
self.matching_blocks = self.opcodes = None
|
|
self.fullbcount = None
|
|
self.__chain_b()
|
|
|
|
# For each element x in b, set b2j[x] to a list of the indices in
|
|
# b where x appears; the indices are in increasing order; note that
|
|
# the number of times x appears in b is len(b2j[x]) ...
|
|
# when self.isjunk is defined, junk elements don't show up in this
|
|
# map at all, which stops the central find_longest_match method
|
|
# from starting any matching block at a junk element ...
|
|
# also creates the fast isbjunk function ...
|
|
# b2j also does not contain entries for "popular" elements, meaning
|
|
# elements that account for more than 1% of the total elements, and
|
|
# when the sequence is reasonably large (>= 200 elements); this can
|
|
# be viewed as an adaptive notion of semi-junk, and yields an enormous
|
|
# speedup when, e.g., comparing program files with hundreds of
|
|
# instances of "return NULL;" ...
|
|
# note that this is only called when b changes; so for cross-product
|
|
# kinds of matches, it's best to call set_seq2 once, then set_seq1
|
|
# repeatedly
|
|
|
|
def __chain_b(self):
|
|
# Because isjunk is a user-defined (not C) function, and we test
|
|
# for junk a LOT, it's important to minimize the number of calls.
|
|
# Before the tricks described here, __chain_b was by far the most
|
|
# time-consuming routine in the whole module! If anyone sees
|
|
# Jim Roskind, thank him again for profile.py -- I never would
|
|
# have guessed that.
|
|
# The first trick is to build b2j ignoring the possibility
|
|
# of junk. I.e., we don't call isjunk at all yet. Throwing
|
|
# out the junk later is much cheaper than building b2j "right"
|
|
# from the start.
|
|
b = self.b
|
|
n = len(b)
|
|
self.b2j = b2j = {}
|
|
populardict = {}
|
|
for i, elt in enumerate(b):
|
|
if elt in b2j:
|
|
indices = b2j[elt]
|
|
if n >= 200 and len(indices) * 100 > n:
|
|
populardict[elt] = 1
|
|
del indices[:]
|
|
else:
|
|
indices.append(i)
|
|
else:
|
|
b2j[elt] = [i]
|
|
|
|
# Purge leftover indices for popular elements.
|
|
for elt in populardict:
|
|
del b2j[elt]
|
|
|
|
# Now b2j.keys() contains elements uniquely, and especially when
|
|
# the sequence is a string, that's usually a good deal smaller
|
|
# than len(string). The difference is the number of isjunk calls
|
|
# saved.
|
|
isjunk = self.isjunk
|
|
junkdict = {}
|
|
if isjunk:
|
|
for d in populardict, b2j:
|
|
for elt in d.keys():
|
|
if isjunk(elt):
|
|
junkdict[elt] = 1
|
|
del d[elt]
|
|
|
|
# Now for x in b, isjunk(x) == junkdict.has_key(x), but the
|
|
# latter is much faster. Note too that while there may be a
|
|
# lot of junk in the sequence, the number of *unique* junk
|
|
# elements is probably small. So the memory burden of keeping
|
|
# this dict alive is likely trivial compared to the size of b2j.
|
|
self.isbjunk = junkdict.has_key
|
|
self.isbpopular = populardict.has_key
|
|
|
|
def find_longest_match(self, alo, ahi, blo, bhi):
|
|
"""Find longest matching block in a[alo:ahi] and b[blo:bhi].
|
|
|
|
If isjunk is not defined:
|
|
|
|
Return (i,j,k) such that a[i:i+k] is equal to b[j:j+k], where
|
|
alo <= i <= i+k <= ahi
|
|
blo <= j <= j+k <= bhi
|
|
and for all (i',j',k') meeting those conditions,
|
|
k >= k'
|
|
i <= i'
|
|
and if i == i', j <= j'
|
|
|
|
In other words, of all maximal matching blocks, return one that
|
|
starts earliest in a, and of all those maximal matching blocks that
|
|
start earliest in a, return the one that starts earliest in b.
|
|
|
|
>>> s = SequenceMatcher(None, " abcd", "abcd abcd")
|
|
>>> s.find_longest_match(0, 5, 0, 9)
|
|
(0, 4, 5)
|
|
|
|
If isjunk is defined, first the longest matching block is
|
|
determined as above, but with the additional restriction that no
|
|
junk element appears in the block. Then that block is extended as
|
|
far as possible by matching (only) junk elements on both sides. So
|
|
the resulting block never matches on junk except as identical junk
|
|
happens to be adjacent to an "interesting" match.
|
|
|
|
Here's the same example as before, but considering blanks to be
|
|
junk. That prevents " abcd" from matching the " abcd" at the tail
|
|
end of the second sequence directly. Instead only the "abcd" can
|
|
match, and matches the leftmost "abcd" in the second sequence:
|
|
|
|
>>> s = SequenceMatcher(lambda x: x==" ", " abcd", "abcd abcd")
|
|
>>> s.find_longest_match(0, 5, 0, 9)
|
|
(1, 0, 4)
|
|
|
|
If no blocks match, return (alo, blo, 0).
|
|
|
|
>>> s = SequenceMatcher(None, "ab", "c")
|
|
>>> s.find_longest_match(0, 2, 0, 1)
|
|
(0, 0, 0)
|
|
"""
|
|
|
|
# CAUTION: stripping common prefix or suffix would be incorrect.
|
|
# E.g.,
|
|
# ab
|
|
# acab
|
|
# Longest matching block is "ab", but if common prefix is
|
|
# stripped, it's "a" (tied with "b"). UNIX(tm) diff does so
|
|
# strip, so ends up claiming that ab is changed to acab by
|
|
# inserting "ca" in the middle. That's minimal but unintuitive:
|
|
# "it's obvious" that someone inserted "ac" at the front.
|
|
# Windiff ends up at the same place as diff, but by pairing up
|
|
# the unique 'b's and then matching the first two 'a's.
|
|
|
|
a, b, b2j, isbjunk = self.a, self.b, self.b2j, self.isbjunk
|
|
besti, bestj, bestsize = alo, blo, 0
|
|
# find longest junk-free match
|
|
# during an iteration of the loop, j2len[j] = length of longest
|
|
# junk-free match ending with a[i-1] and b[j]
|
|
j2len = {}
|
|
nothing = []
|
|
for i in xrange(alo, ahi):
|
|
# look at all instances of a[i] in b; note that because
|
|
# b2j has no junk keys, the loop is skipped if a[i] is junk
|
|
j2lenget = j2len.get
|
|
newj2len = {}
|
|
for j in b2j.get(a[i], nothing):
|
|
# a[i] matches b[j]
|
|
if j < blo:
|
|
continue
|
|
if j >= bhi:
|
|
break
|
|
k = newj2len[j] = j2lenget(j-1, 0) + 1
|
|
if k > bestsize:
|
|
besti, bestj, bestsize = i-k+1, j-k+1, k
|
|
j2len = newj2len
|
|
|
|
# Extend the best by non-junk elements on each end. In particular,
|
|
# "popular" non-junk elements aren't in b2j, which greatly speeds
|
|
# the inner loop above, but also means "the best" match so far
|
|
# doesn't contain any junk *or* popular non-junk elements.
|
|
while besti > alo and bestj > blo and \
|
|
not isbjunk(b[bestj-1]) and \
|
|
a[besti-1] == b[bestj-1]:
|
|
besti, bestj, bestsize = besti-1, bestj-1, bestsize+1
|
|
while besti+bestsize < ahi and bestj+bestsize < bhi and \
|
|
not isbjunk(b[bestj+bestsize]) and \
|
|
a[besti+bestsize] == b[bestj+bestsize]:
|
|
bestsize += 1
|
|
|
|
# Now that we have a wholly interesting match (albeit possibly
|
|
# empty!), we may as well suck up the matching junk on each
|
|
# side of it too. Can't think of a good reason not to, and it
|
|
# saves post-processing the (possibly considerable) expense of
|
|
# figuring out what to do with it. In the case of an empty
|
|
# interesting match, this is clearly the right thing to do,
|
|
# because no other kind of match is possible in the regions.
|
|
while besti > alo and bestj > blo and \
|
|
isbjunk(b[bestj-1]) and \
|
|
a[besti-1] == b[bestj-1]:
|
|
besti, bestj, bestsize = besti-1, bestj-1, bestsize+1
|
|
while besti+bestsize < ahi and bestj+bestsize < bhi and \
|
|
isbjunk(b[bestj+bestsize]) and \
|
|
a[besti+bestsize] == b[bestj+bestsize]:
|
|
bestsize = bestsize + 1
|
|
|
|
return besti, bestj, bestsize
|
|
|
|
def get_matching_blocks(self):
|
|
"""Return list of triples describing matching subsequences.
|
|
|
|
Each triple is of the form (i, j, n), and means that
|
|
a[i:i+n] == b[j:j+n]. The triples are monotonically increasing in
|
|
i and in j.
|
|
|
|
The last triple is a dummy, (len(a), len(b), 0), and is the only
|
|
triple with n==0.
|
|
|
|
>>> s = SequenceMatcher(None, "abxcd", "abcd")
|
|
>>> s.get_matching_blocks()
|
|
[(0, 0, 2), (3, 2, 2), (5, 4, 0)]
|
|
"""
|
|
|
|
if self.matching_blocks is not None:
|
|
return self.matching_blocks
|
|
self.matching_blocks = []
|
|
la, lb = len(self.a), len(self.b)
|
|
self.__helper(0, la, 0, lb, self.matching_blocks)
|
|
self.matching_blocks.append( (la, lb, 0) )
|
|
return self.matching_blocks
|
|
|
|
# builds list of matching blocks covering a[alo:ahi] and
|
|
# b[blo:bhi], appending them in increasing order to answer
|
|
|
|
def __helper(self, alo, ahi, blo, bhi, answer):
|
|
i, j, k = x = self.find_longest_match(alo, ahi, blo, bhi)
|
|
# a[alo:i] vs b[blo:j] unknown
|
|
# a[i:i+k] same as b[j:j+k]
|
|
# a[i+k:ahi] vs b[j+k:bhi] unknown
|
|
if k:
|
|
if alo < i and blo < j:
|
|
self.__helper(alo, i, blo, j, answer)
|
|
answer.append(x)
|
|
if i+k < ahi and j+k < bhi:
|
|
self.__helper(i+k, ahi, j+k, bhi, answer)
|
|
|
|
def get_opcodes(self):
|
|
"""Return list of 5-tuples describing how to turn a into b.
|
|
|
|
Each tuple is of the form (tag, i1, i2, j1, j2). The first tuple
|
|
has i1 == j1 == 0, and remaining tuples have i1 == the i2 from the
|
|
tuple preceding it, and likewise for j1 == the previous j2.
|
|
|
|
The tags are strings, with these meanings:
|
|
|
|
'replace': a[i1:i2] should be replaced by b[j1:j2]
|
|
'delete': a[i1:i2] should be deleted.
|
|
Note that j1==j2 in this case.
|
|
'insert': b[j1:j2] should be inserted at a[i1:i1].
|
|
Note that i1==i2 in this case.
|
|
'equal': a[i1:i2] == b[j1:j2]
|
|
|
|
>>> a = "qabxcd"
|
|
>>> b = "abycdf"
|
|
>>> s = SequenceMatcher(None, a, b)
|
|
>>> for tag, i1, i2, j1, j2 in s.get_opcodes():
|
|
... print ("%7s a[%d:%d] (%s) b[%d:%d] (%s)" %
|
|
... (tag, i1, i2, a[i1:i2], j1, j2, b[j1:j2]))
|
|
delete a[0:1] (q) b[0:0] ()
|
|
equal a[1:3] (ab) b[0:2] (ab)
|
|
replace a[3:4] (x) b[2:3] (y)
|
|
equal a[4:6] (cd) b[3:5] (cd)
|
|
insert a[6:6] () b[5:6] (f)
|
|
"""
|
|
|
|
if self.opcodes is not None:
|
|
return self.opcodes
|
|
i = j = 0
|
|
self.opcodes = answer = []
|
|
for ai, bj, size in self.get_matching_blocks():
|
|
# invariant: we've pumped out correct diffs to change
|
|
# a[:i] into b[:j], and the next matching block is
|
|
# a[ai:ai+size] == b[bj:bj+size]. So we need to pump
|
|
# out a diff to change a[i:ai] into b[j:bj], pump out
|
|
# the matching block, and move (i,j) beyond the match
|
|
tag = ''
|
|
if i < ai and j < bj:
|
|
tag = 'replace'
|
|
elif i < ai:
|
|
tag = 'delete'
|
|
elif j < bj:
|
|
tag = 'insert'
|
|
if tag:
|
|
answer.append( (tag, i, ai, j, bj) )
|
|
i, j = ai+size, bj+size
|
|
# the list of matching blocks is terminated by a
|
|
# sentinel with size 0
|
|
if size:
|
|
answer.append( ('equal', ai, i, bj, j) )
|
|
return answer
|
|
|
|
def ratio(self):
|
|
"""Return a measure of the sequences' similarity (float in [0,1]).
|
|
|
|
Where T is the total number of elements in both sequences, and
|
|
M is the number of matches, this is 2,0*M / T.
|
|
Note that this is 1 if the sequences are identical, and 0 if
|
|
they have nothing in common.
|
|
|
|
.ratio() is expensive to compute if you haven't already computed
|
|
.get_matching_blocks() or .get_opcodes(), in which case you may
|
|
want to try .quick_ratio() or .real_quick_ratio() first to get an
|
|
upper bound.
|
|
|
|
>>> s = SequenceMatcher(None, "abcd", "bcde")
|
|
>>> s.ratio()
|
|
0.75
|
|
>>> s.quick_ratio()
|
|
0.75
|
|
>>> s.real_quick_ratio()
|
|
1.0
|
|
"""
|
|
|
|
matches = reduce(lambda sum, triple: sum + triple[-1],
|
|
self.get_matching_blocks(), 0)
|
|
return 2.0 * matches / (len(self.a) + len(self.b))
|
|
|
|
def quick_ratio(self):
|
|
"""Return an upper bound on ratio() relatively quickly.
|
|
|
|
This isn't defined beyond that it is an upper bound on .ratio(), and
|
|
is faster to compute.
|
|
"""
|
|
|
|
# viewing a and b as multisets, set matches to the cardinality
|
|
# of their intersection; this counts the number of matches
|
|
# without regard to order, so is clearly an upper bound
|
|
if self.fullbcount is None:
|
|
self.fullbcount = fullbcount = {}
|
|
for elt in self.b:
|
|
fullbcount[elt] = fullbcount.get(elt, 0) + 1
|
|
fullbcount = self.fullbcount
|
|
# avail[x] is the number of times x appears in 'b' less the
|
|
# number of times we've seen it in 'a' so far ... kinda
|
|
avail = {}
|
|
availhas, matches = avail.has_key, 0
|
|
for elt in self.a:
|
|
if availhas(elt):
|
|
numb = avail[elt]
|
|
else:
|
|
numb = fullbcount.get(elt, 0)
|
|
avail[elt] = numb - 1
|
|
if numb > 0:
|
|
matches = matches + 1
|
|
return 2.0 * matches / (len(self.a) + len(self.b))
|
|
|
|
def real_quick_ratio(self):
|
|
"""Return an upper bound on ratio() very quickly.
|
|
|
|
This isn't defined beyond that it is an upper bound on .ratio(), and
|
|
is faster to compute than either .ratio() or .quick_ratio().
|
|
"""
|
|
|
|
la, lb = len(self.a), len(self.b)
|
|
# can't have more matches than the number of elements in the
|
|
# shorter sequence
|
|
return 2.0 * min(la, lb) / (la + lb)
|
|
|
|
def get_close_matches(word, possibilities, n=3, cutoff=0.6):
|
|
"""Use SequenceMatcher to return list of the best "good enough" matches.
|
|
|
|
word is a sequence for which close matches are desired (typically a
|
|
string).
|
|
|
|
possibilities is a list of sequences against which to match word
|
|
(typically a list of strings).
|
|
|
|
Optional arg n (default 3) is the maximum number of close matches to
|
|
return. n must be > 0.
|
|
|
|
Optional arg cutoff (default 0.6) is a float in [0, 1]. Possibilities
|
|
that don't score at least that similar to word are ignored.
|
|
|
|
The best (no more than n) matches among the possibilities are returned
|
|
in a list, sorted by similarity score, most similar first.
|
|
|
|
>>> get_close_matches("appel", ["ape", "apple", "peach", "puppy"])
|
|
['apple', 'ape']
|
|
>>> import keyword as _keyword
|
|
>>> get_close_matches("wheel", _keyword.kwlist)
|
|
['while']
|
|
>>> get_close_matches("apple", _keyword.kwlist)
|
|
[]
|
|
>>> get_close_matches("accept", _keyword.kwlist)
|
|
['except']
|
|
"""
|
|
|
|
if not n > 0:
|
|
raise ValueError("n must be > 0: " + `n`)
|
|
if not 0.0 <= cutoff <= 1.0:
|
|
raise ValueError("cutoff must be in [0.0, 1.0]: " + `cutoff`)
|
|
result = []
|
|
s = SequenceMatcher()
|
|
s.set_seq2(word)
|
|
for x in possibilities:
|
|
s.set_seq1(x)
|
|
if s.real_quick_ratio() >= cutoff and \
|
|
s.quick_ratio() >= cutoff and \
|
|
s.ratio() >= cutoff:
|
|
result.append((s.ratio(), x))
|
|
# Sort by score.
|
|
result.sort()
|
|
# Retain only the best n.
|
|
result = result[-n:]
|
|
# Move best-scorer to head of list.
|
|
result.reverse()
|
|
# Strip scores.
|
|
return [x for score, x in result]
|
|
|
|
|
|
def _count_leading(line, ch):
|
|
"""
|
|
Return number of `ch` characters at the start of `line`.
|
|
|
|
Example:
|
|
|
|
>>> _count_leading(' abc', ' ')
|
|
3
|
|
"""
|
|
|
|
i, n = 0, len(line)
|
|
while i < n and line[i] == ch:
|
|
i += 1
|
|
return i
|
|
|
|
class Differ:
|
|
r"""
|
|
Differ is a class for comparing sequences of lines of text, and
|
|
producing human-readable differences or deltas. Differ uses
|
|
SequenceMatcher both to compare sequences of lines, and to compare
|
|
sequences of characters within similar (near-matching) lines.
|
|
|
|
Each line of a Differ delta begins with a two-letter code:
|
|
|
|
'- ' line unique to sequence 1
|
|
'+ ' line unique to sequence 2
|
|
' ' line common to both sequences
|
|
'? ' line not present in either input sequence
|
|
|
|
Lines beginning with '? ' attempt to guide the eye to intraline
|
|
differences, and were not present in either input sequence. These lines
|
|
can be confusing if the sequences contain tab characters.
|
|
|
|
Note that Differ makes no claim to produce a *minimal* diff. To the
|
|
contrary, minimal diffs are often counter-intuitive, because they synch
|
|
up anywhere possible, sometimes accidental matches 100 pages apart.
|
|
Restricting synch points to contiguous matches preserves some notion of
|
|
locality, at the occasional cost of producing a longer diff.
|
|
|
|
Example: Comparing two texts.
|
|
|
|
First we set up the texts, sequences of individual single-line strings
|
|
ending with newlines (such sequences can also be obtained from the
|
|
`readlines()` method of file-like objects):
|
|
|
|
>>> text1 = ''' 1. Beautiful is better than ugly.
|
|
... 2. Explicit is better than implicit.
|
|
... 3. Simple is better than complex.
|
|
... 4. Complex is better than complicated.
|
|
... '''.splitlines(1)
|
|
>>> len(text1)
|
|
4
|
|
>>> text1[0][-1]
|
|
'\n'
|
|
>>> text2 = ''' 1. Beautiful is better than ugly.
|
|
... 3. Simple is better than complex.
|
|
... 4. Complicated is better than complex.
|
|
... 5. Flat is better than nested.
|
|
... '''.splitlines(1)
|
|
|
|
Next we instantiate a Differ object:
|
|
|
|
>>> d = Differ()
|
|
|
|
Note that when instantiating a Differ object we may pass functions to
|
|
filter out line and character 'junk'. See Differ.__init__ for details.
|
|
|
|
Finally, we compare the two:
|
|
|
|
>>> result = list(d.compare(text1, text2))
|
|
|
|
'result' is a list of strings, so let's pretty-print it:
|
|
|
|
>>> from pprint import pprint as _pprint
|
|
>>> _pprint(result)
|
|
[' 1. Beautiful is better than ugly.\n',
|
|
'- 2. Explicit is better than implicit.\n',
|
|
'- 3. Simple is better than complex.\n',
|
|
'+ 3. Simple is better than complex.\n',
|
|
'? ++\n',
|
|
'- 4. Complex is better than complicated.\n',
|
|
'? ^ ---- ^\n',
|
|
'+ 4. Complicated is better than complex.\n',
|
|
'? ++++ ^ ^\n',
|
|
'+ 5. Flat is better than nested.\n']
|
|
|
|
As a single multi-line string it looks like this:
|
|
|
|
>>> print ''.join(result),
|
|
1. Beautiful is better than ugly.
|
|
- 2. Explicit is better than implicit.
|
|
- 3. Simple is better than complex.
|
|
+ 3. Simple is better than complex.
|
|
? ++
|
|
- 4. Complex is better than complicated.
|
|
? ^ ---- ^
|
|
+ 4. Complicated is better than complex.
|
|
? ++++ ^ ^
|
|
+ 5. Flat is better than nested.
|
|
|
|
Methods:
|
|
|
|
__init__(linejunk=None, charjunk=None)
|
|
Construct a text differencer, with optional filters.
|
|
|
|
compare(a, b)
|
|
Compare two sequences of lines; generate the resulting delta.
|
|
"""
|
|
|
|
def __init__(self, linejunk=None, charjunk=None):
|
|
"""
|
|
Construct a text differencer, with optional filters.
|
|
|
|
The two optional keyword parameters are for filter functions:
|
|
|
|
- `linejunk`: A function that should accept a single string argument,
|
|
and return true iff the string is junk. The module-level function
|
|
`IS_LINE_JUNK` may be used to filter out lines without visible
|
|
characters, except for at most one splat ('#'). It is recommended
|
|
to leave linejunk None; as of Python 2.3, the underlying
|
|
SequenceMatcher class has grown an adaptive notion of "noise" lines
|
|
that's better than any static definition the author has ever been
|
|
able to craft.
|
|
|
|
- `charjunk`: A function that should accept a string of length 1. The
|
|
module-level function `IS_CHARACTER_JUNK` may be used to filter out
|
|
whitespace characters (a blank or tab; **note**: bad idea to include
|
|
newline in this!). Use of IS_CHARACTER_JUNK is recommended.
|
|
"""
|
|
|
|
self.linejunk = linejunk
|
|
self.charjunk = charjunk
|
|
|
|
def compare(self, a, b):
|
|
r"""
|
|
Compare two sequences of lines; generate the resulting delta.
|
|
|
|
Each sequence must contain individual single-line strings ending with
|
|
newlines. Such sequences can be obtained from the `readlines()` method
|
|
of file-like objects. The delta generated also consists of newline-
|
|
terminated strings, ready to be printed as-is via the writeline()
|
|
method of a file-like object.
|
|
|
|
Example:
|
|
|
|
>>> print ''.join(Differ().compare('one\ntwo\nthree\n'.splitlines(1),
|
|
... 'ore\ntree\nemu\n'.splitlines(1))),
|
|
- one
|
|
? ^
|
|
+ ore
|
|
? ^
|
|
- two
|
|
- three
|
|
? -
|
|
+ tree
|
|
+ emu
|
|
"""
|
|
|
|
cruncher = SequenceMatcher(self.linejunk, a, b)
|
|
for tag, alo, ahi, blo, bhi in cruncher.get_opcodes():
|
|
if tag == 'replace':
|
|
g = self._fancy_replace(a, alo, ahi, b, blo, bhi)
|
|
elif tag == 'delete':
|
|
g = self._dump('-', a, alo, ahi)
|
|
elif tag == 'insert':
|
|
g = self._dump('+', b, blo, bhi)
|
|
elif tag == 'equal':
|
|
g = self._dump(' ', a, alo, ahi)
|
|
else:
|
|
raise ValueError, 'unknown tag ' + `tag`
|
|
|
|
for line in g:
|
|
yield line
|
|
|
|
def _dump(self, tag, x, lo, hi):
|
|
"""Generate comparison results for a same-tagged range."""
|
|
for i in xrange(lo, hi):
|
|
yield '%s %s' % (tag, x[i])
|
|
|
|
def _plain_replace(self, a, alo, ahi, b, blo, bhi):
|
|
assert alo < ahi and blo < bhi
|
|
# dump the shorter block first -- reduces the burden on short-term
|
|
# memory if the blocks are of very different sizes
|
|
if bhi - blo < ahi - alo:
|
|
first = self._dump('+', b, blo, bhi)
|
|
second = self._dump('-', a, alo, ahi)
|
|
else:
|
|
first = self._dump('-', a, alo, ahi)
|
|
second = self._dump('+', b, blo, bhi)
|
|
|
|
for g in first, second:
|
|
for line in g:
|
|
yield line
|
|
|
|
def _fancy_replace(self, a, alo, ahi, b, blo, bhi):
|
|
r"""
|
|
When replacing one block of lines with another, search the blocks
|
|
for *similar* lines; the best-matching pair (if any) is used as a
|
|
synch point, and intraline difference marking is done on the
|
|
similar pair. Lots of work, but often worth it.
|
|
|
|
Example:
|
|
|
|
>>> d = Differ()
|
|
>>> d._fancy_replace(['abcDefghiJkl\n'], 0, 1, ['abcdefGhijkl\n'], 0, 1)
|
|
>>> print ''.join(d.results),
|
|
- abcDefghiJkl
|
|
? ^ ^ ^
|
|
+ abcdefGhijkl
|
|
? ^ ^ ^
|
|
"""
|
|
|
|
# don't synch up unless the lines have a similarity score of at
|
|
# least cutoff; best_ratio tracks the best score seen so far
|
|
best_ratio, cutoff = 0.74, 0.75
|
|
cruncher = SequenceMatcher(self.charjunk)
|
|
eqi, eqj = None, None # 1st indices of equal lines (if any)
|
|
|
|
# search for the pair that matches best without being identical
|
|
# (identical lines must be junk lines, & we don't want to synch up
|
|
# on junk -- unless we have to)
|
|
for j in xrange(blo, bhi):
|
|
bj = b[j]
|
|
cruncher.set_seq2(bj)
|
|
for i in xrange(alo, ahi):
|
|
ai = a[i]
|
|
if ai == bj:
|
|
if eqi is None:
|
|
eqi, eqj = i, j
|
|
continue
|
|
cruncher.set_seq1(ai)
|
|
# computing similarity is expensive, so use the quick
|
|
# upper bounds first -- have seen this speed up messy
|
|
# compares by a factor of 3.
|
|
# note that ratio() is only expensive to compute the first
|
|
# time it's called on a sequence pair; the expensive part
|
|
# of the computation is cached by cruncher
|
|
if cruncher.real_quick_ratio() > best_ratio and \
|
|
cruncher.quick_ratio() > best_ratio and \
|
|
cruncher.ratio() > best_ratio:
|
|
best_ratio, best_i, best_j = cruncher.ratio(), i, j
|
|
if best_ratio < cutoff:
|
|
# no non-identical "pretty close" pair
|
|
if eqi is None:
|
|
# no identical pair either -- treat it as a straight replace
|
|
for line in self._plain_replace(a, alo, ahi, b, blo, bhi):
|
|
yield line
|
|
return
|
|
# no close pair, but an identical pair -- synch up on that
|
|
best_i, best_j, best_ratio = eqi, eqj, 1.0
|
|
else:
|
|
# there's a close pair, so forget the identical pair (if any)
|
|
eqi = None
|
|
|
|
# a[best_i] very similar to b[best_j]; eqi is None iff they're not
|
|
# identical
|
|
|
|
# pump out diffs from before the synch point
|
|
for line in self._fancy_helper(a, alo, best_i, b, blo, best_j):
|
|
yield line
|
|
|
|
# do intraline marking on the synch pair
|
|
aelt, belt = a[best_i], b[best_j]
|
|
if eqi is None:
|
|
# pump out a '-', '?', '+', '?' quad for the synched lines
|
|
atags = btags = ""
|
|
cruncher.set_seqs(aelt, belt)
|
|
for tag, ai1, ai2, bj1, bj2 in cruncher.get_opcodes():
|
|
la, lb = ai2 - ai1, bj2 - bj1
|
|
if tag == 'replace':
|
|
atags += '^' * la
|
|
btags += '^' * lb
|
|
elif tag == 'delete':
|
|
atags += '-' * la
|
|
elif tag == 'insert':
|
|
btags += '+' * lb
|
|
elif tag == 'equal':
|
|
atags += ' ' * la
|
|
btags += ' ' * lb
|
|
else:
|
|
raise ValueError, 'unknown tag ' + `tag`
|
|
for line in self._qformat(aelt, belt, atags, btags):
|
|
yield line
|
|
else:
|
|
# the synch pair is identical
|
|
yield ' ' + aelt
|
|
|
|
# pump out diffs from after the synch point
|
|
for line in self._fancy_helper(a, best_i+1, ahi, b, best_j+1, bhi):
|
|
yield line
|
|
|
|
def _fancy_helper(self, a, alo, ahi, b, blo, bhi):
|
|
g = []
|
|
if alo < ahi:
|
|
if blo < bhi:
|
|
g = self._fancy_replace(a, alo, ahi, b, blo, bhi)
|
|
else:
|
|
g = self._dump('-', a, alo, ahi)
|
|
elif blo < bhi:
|
|
g = self._dump('+', b, blo, bhi)
|
|
|
|
for line in g:
|
|
yield line
|
|
|
|
def _qformat(self, aline, bline, atags, btags):
|
|
r"""
|
|
Format "?" output and deal with leading tabs.
|
|
|
|
Example:
|
|
|
|
>>> d = Differ()
|
|
>>> d._qformat('\tabcDefghiJkl\n', '\t\tabcdefGhijkl\n',
|
|
... ' ^ ^ ^ ', '+ ^ ^ ^ ')
|
|
>>> for line in d.results: print repr(line)
|
|
...
|
|
'- \tabcDefghiJkl\n'
|
|
'? \t ^ ^ ^\n'
|
|
'+ \t\tabcdefGhijkl\n'
|
|
'? \t ^ ^ ^\n'
|
|
"""
|
|
|
|
# Can hurt, but will probably help most of the time.
|
|
common = min(_count_leading(aline, "\t"),
|
|
_count_leading(bline, "\t"))
|
|
common = min(common, _count_leading(atags[:common], " "))
|
|
atags = atags[common:].rstrip()
|
|
btags = btags[common:].rstrip()
|
|
|
|
yield "- " + aline
|
|
if atags:
|
|
yield "? %s%s\n" % ("\t" * common, atags)
|
|
|
|
yield "+ " + bline
|
|
if btags:
|
|
yield "? %s%s\n" % ("\t" * common, btags)
|
|
|
|
# With respect to junk, an earlier version of ndiff simply refused to
|
|
# *start* a match with a junk element. The result was cases like this:
|
|
# before: private Thread currentThread;
|
|
# after: private volatile Thread currentThread;
|
|
# If you consider whitespace to be junk, the longest contiguous match
|
|
# not starting with junk is "e Thread currentThread". So ndiff reported
|
|
# that "e volatil" was inserted between the 't' and the 'e' in "private".
|
|
# While an accurate view, to people that's absurd. The current version
|
|
# looks for matching blocks that are entirely junk-free, then extends the
|
|
# longest one of those as far as possible but only with matching junk.
|
|
# So now "currentThread" is matched, then extended to suck up the
|
|
# preceding blank; then "private" is matched, and extended to suck up the
|
|
# following blank; then "Thread" is matched; and finally ndiff reports
|
|
# that "volatile " was inserted before "Thread". The only quibble
|
|
# remaining is that perhaps it was really the case that " volatile"
|
|
# was inserted after "private". I can live with that <wink>.
|
|
|
|
import re
|
|
|
|
def IS_LINE_JUNK(line, pat=re.compile(r"\s*#?\s*$").match):
|
|
r"""
|
|
Return 1 for ignorable line: iff `line` is blank or contains a single '#'.
|
|
|
|
Examples:
|
|
|
|
>>> IS_LINE_JUNK('\n')
|
|
True
|
|
>>> IS_LINE_JUNK(' # \n')
|
|
True
|
|
>>> IS_LINE_JUNK('hello\n')
|
|
False
|
|
"""
|
|
|
|
return pat(line) is not None
|
|
|
|
def IS_CHARACTER_JUNK(ch, ws=" \t"):
|
|
r"""
|
|
Return 1 for ignorable character: iff `ch` is a space or tab.
|
|
|
|
Examples:
|
|
|
|
>>> IS_CHARACTER_JUNK(' ')
|
|
True
|
|
>>> IS_CHARACTER_JUNK('\t')
|
|
True
|
|
>>> IS_CHARACTER_JUNK('\n')
|
|
False
|
|
>>> IS_CHARACTER_JUNK('x')
|
|
False
|
|
"""
|
|
|
|
return ch in ws
|
|
|
|
del re
|
|
|
|
def ndiff(a, b, linejunk=None, charjunk=IS_CHARACTER_JUNK):
|
|
r"""
|
|
Compare `a` and `b` (lists of strings); return a `Differ`-style delta.
|
|
|
|
Optional keyword parameters `linejunk` and `charjunk` are for filter
|
|
functions (or None):
|
|
|
|
- linejunk: A function that should accept a single string argument, and
|
|
return true iff the string is junk. The default is None, and is
|
|
recommended; as of Python 2.3, an adaptive notion of "noise" lines is
|
|
used that does a good job on its own.
|
|
|
|
- charjunk: A function that should accept a string of length 1. The
|
|
default is module-level function IS_CHARACTER_JUNK, which filters out
|
|
whitespace characters (a blank or tab; note: bad idea to include newline
|
|
in this!).
|
|
|
|
Tools/scripts/ndiff.py is a command-line front-end to this function.
|
|
|
|
Example:
|
|
|
|
>>> diff = ndiff('one\ntwo\nthree\n'.splitlines(1),
|
|
... 'ore\ntree\nemu\n'.splitlines(1))
|
|
>>> print ''.join(diff),
|
|
- one
|
|
? ^
|
|
+ ore
|
|
? ^
|
|
- two
|
|
- three
|
|
? -
|
|
+ tree
|
|
+ emu
|
|
"""
|
|
return Differ(linejunk, charjunk).compare(a, b)
|
|
|
|
def restore(delta, which):
|
|
r"""
|
|
Generate one of the two sequences that generated a delta.
|
|
|
|
Given a `delta` produced by `Differ.compare()` or `ndiff()`, extract
|
|
lines originating from file 1 or 2 (parameter `which`), stripping off line
|
|
prefixes.
|
|
|
|
Examples:
|
|
|
|
>>> diff = ndiff('one\ntwo\nthree\n'.splitlines(1),
|
|
... 'ore\ntree\nemu\n'.splitlines(1))
|
|
>>> diff = list(diff)
|
|
>>> print ''.join(restore(diff, 1)),
|
|
one
|
|
two
|
|
three
|
|
>>> print ''.join(restore(diff, 2)),
|
|
ore
|
|
tree
|
|
emu
|
|
"""
|
|
try:
|
|
tag = {1: "- ", 2: "+ "}[int(which)]
|
|
except KeyError:
|
|
raise ValueError, ('unknown delta choice (must be 1 or 2): %r'
|
|
% which)
|
|
prefixes = (" ", tag)
|
|
for line in delta:
|
|
if line[:2] in prefixes:
|
|
yield line[2:]
|
|
|
|
def _test():
|
|
import doctest, difflib
|
|
return doctest.testmod(difflib)
|
|
|
|
if __name__ == "__main__":
|
|
_test()
|