文字差异JSON [英] Textually diffing JSON
问题描述
作为发布过程的一部分，我必须比较应用程序使用的一些JSON配置数据.第一次尝试，我只是漂亮地打印了JSON并将其差异化(使用kdiff3或仅差异化).
As part of my release processes, I have to compare some JSON configuration data used by my application. As a first attempt, I just pretty-printed the JSON and diff'ed them (using kdiff3 or just diff).
但是，随着数据的增长，kdiff3混淆了输出中的不同部分，使添加看起来像是巨无穷的修改，奇数的删除等.这使得很难找出不同之处.我也尝试过其他差异工具(meld，kompare，diff，其他一些)，但是它们都存在相同的问题.
As that data has grown, however, kdiff3 confuses different parts in the output, making additions look like giant modifies, odd deletions, etc. It makes it really hard to figure out what is different. I've tried other diff tools, too (meld, kompare, diff, a few others), but they all have the same problem.
尽管尽了最大的努力，但我似乎无法以diff工具可以理解的方式来格式化JSON.
Despite my best efforts, I can't seem to format the JSON in a way that the diff tools can understand.
示例数据:
[
{
"name": "date",
"type": "date",
"nullable": true,
"state": "enabled"
},
{
"name": "owner",
"type": "string",
"nullable": false,
"state": "enabled",
}
...lots more...
]
以上可能不会导致问题(当开始有数百行时会发生问题)，但这就是所比较内容的要旨.
The above probably wouldn't cause the problem (the problem occurs when there begin to be hundreds of lines), but thats the gist of what is being compared.
那只是一个例子；完整对象是4-5个属性，有些属性中有4-5个属性.属性名称很统一，但是它们的值却很不相同.
Thats just a sample; the full objects are 4-5 attributes, and some attributes have 4-5 attributes in them. The attribute names are pretty uniform, but their values pretty varied.
通常，似乎所有的diff工具都将结束符}"与下一个关闭}"的对象混淆了.我似乎无法摆脱他们的这种习惯.
In general, it seems like all the diff tools confuse the closing "}" with the next objects closing "}". I can't seem to break them of this habit.
我尝试在各个对象之前和之后添加空格，更改缩进以及添加一些"BEGIN"和"END"字符串，但是该工具仍然感到困惑.
I've tried adding whitespace, changing indentation, and adding some "BEGIN" and "END" strings before and after the respective objects, but the tool still get confused.
推荐答案
如果您的工具中有任何选项，则耐心差异可以为您带来更好的效果.我将尝试用它找到一个工具(其他tha Git和Bazaar)并进行报告.
If any of your tool has the option, Patience Diff could work a lot better for you. I'll try to find a tool with it (other tha Git and Bazaar) and report back.
It seems that the implementation in Bazaar is usable as a standalone tool with minimal changes.
Edit2:WTH，为什么不粘贴您使我成为黑客的新酷diff脚本的源代码?在这里，我没有版权要求，只是重新整理了Bram/Canonical的代码.
WTH, why not paste the source of the new cool diff script you made me hack? Here it is, no copyright claim on my side, it's just Bram/Canonical's code re-arranged.
#!/usr/bin/env python
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
# Copyright (C) 2005 Bram Cohen, Copyright (C) 2005, 2006 Canonical Ltd
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
import os
import sys
import time
import difflib
from bisect import bisect
__all__ = ['PatienceSequenceMatcher', 'unified_diff', 'unified_diff_files']
py3k = False
try:
xrange
except NameError:
py3k = True
xrange = range
# This is a version of unified_diff which only adds a factory parameter
# so that you can override the default SequenceMatcher
# this has been submitted as a patch to python
def unified_diff(a, b, fromfile='', tofile='', fromfiledate='',
tofiledate='', n=3, lineterm='\n',
sequencematcher=None):
r"""
Compare two sequences of lines; generate the delta as a unified diff.
Unified diffs are a compact way of showing line changes and a few
lines of context. The number of context lines is set by 'n' which
defaults to three.
By default, the diff control lines (those with ---, +++, or @@) are
created with a trailing newline. This is helpful so that inputs
created from file.readlines() result in diffs that are suitable for
file.writelines() since both the inputs and outputs have trailing
newlines.
For inputs that do not have trailing newlines, set the lineterm
argument to "" so that the output will be uniformly newline free.
The unidiff format normally has a header for filenames and modification
times. Any or all of these may be specified using strings for
'fromfile', 'tofile', 'fromfiledate', and 'tofiledate'. The modification
times are normally expressed in the format returned by time.ctime().
Example:
>>> for line in unified_diff('one two three four'.split(),
... 'zero one tree four'.split(), 'Original', 'Current',
... 'Sat Jan 26 23:30:50 1991', 'Fri Jun 06 10:20:52 2003',
... lineterm=''):
... print line
--- Original Sat Jan 26 23:30:50 1991
+++ Current Fri Jun 06 10:20:52 2003
@@ -1,4 +1,4 @@
+zero
one
-two
-three
+tree
four
"""
if sequencematcher is None:
import difflib
sequencematcher = difflib.SequenceMatcher
if fromfiledate:
fromfiledate = '\t' + str(fromfiledate)
if tofiledate:
tofiledate = '\t' + str(tofiledate)
started = False
for group in sequencematcher(None,a,b).get_grouped_opcodes(n):
if not started:
yield '--- %s%s%s' % (fromfile, fromfiledate, lineterm)
yield '+++ %s%s%s' % (tofile, tofiledate, lineterm)
started = True
i1, i2, j1, j2 = group[0][3], group[-1][4], group[0][5], group[-1][6]
yield "@@ -%d,%d +%d,%d @@%s" % (i1+1, i2-i1, j1+1, j2-j1, lineterm)
for tag, i1, i2, j1, j2 in group:
if tag == 'equal':
for line in a[i1:i2]:
yield ' ' + line
continue
if tag == 'replace' or tag == 'delete':
for line in a[i1:i2]:
yield '-' + line
if tag == 'replace' or tag == 'insert':
for line in b[j1:j2]:
yield '+' + line
def unified_diff_files(a, b, sequencematcher=None):
"""Generate the diff for two files.
"""
mode = 'rb'
if py3k: mode = 'r'
# Should this actually be an error?
if a == b:
return []
if a == '-':
file_a = sys.stdin
time_a = time.time()
else:
file_a = open(a, mode)
time_a = os.stat(a).st_mtime
if b == '-':
file_b = sys.stdin
time_b = time.time()
else:
file_b = open(b, mode)
time_b = os.stat(b).st_mtime
# TODO: Include fromfiledate and tofiledate
return unified_diff(file_a.readlines(), file_b.readlines(),
fromfile=a, tofile=b,
sequencematcher=sequencematcher)
def unique_lcs_py(a, b):
"""Find the longest common subset for unique lines.
:param a: An indexable object (such as string or list of strings)
:param b: Another indexable object (such as string or list of strings)
:return: A list of tuples, one for each line which is matched.
[(line_in_a, line_in_b), ...]
This only matches lines which are unique on both sides.
This helps prevent common lines from over influencing match
results.
The longest common subset uses the Patience Sorting algorithm:
http://en.wikipedia.org/wiki/Patience_sorting
"""
# set index[line in a] = position of line in a unless
# a is a duplicate, in which case it's set to None
index = {}
for i in xrange(len(a)):
line = a[i]
if line in index:
index[line] = None
else:
index[line]= i
# make btoa[i] = position of line i in a, unless
# that line doesn't occur exactly once in both,
# in which case it's set to None
btoa = [None] * len(b)
index2 = {}
for pos, line in enumerate(b):
next = index.get(line)
if next is not None:
if line in index2:
# unset the previous mapping, which we now know to
# be invalid because the line isn't unique
btoa[index2[line]] = None
del index[line]
else:
index2[line] = pos
btoa[pos] = next
# this is the Patience sorting algorithm
# see http://en.wikipedia.org/wiki/Patience_sorting
backpointers = [None] * len(b)
stacks = []
lasts = []
k = 0
for bpos, apos in enumerate(btoa):
if apos is None:
continue
# as an optimization, check if the next line comes at the end,
# because it usually does
if stacks and stacks[-1] < apos:
k = len(stacks)
# as an optimization, check if the next line comes right after
# the previous line, because usually it does
elif stacks and stacks[k] < apos and (k == len(stacks) - 1 or
stacks[k+1] > apos):
k += 1
else:
k = bisect(stacks, apos)
if k > 0:
backpointers[bpos] = lasts[k-1]
if k < len(stacks):
stacks[k] = apos
lasts[k] = bpos
else:
stacks.append(apos)
lasts.append(bpos)
if len(lasts) == 0:
return []
result = []
k = lasts[-1]
while k is not None:
result.append((btoa[k], k))
k = backpointers[k]
result.reverse()
return result
def recurse_matches_py(a, b, alo, blo, ahi, bhi, answer, maxrecursion):
"""Find all of the matching text in the lines of a and b.
:param a: A sequence
:param b: Another sequence
:param alo: The start location of a to check, typically 0
:param ahi: The start location of b to check, typically 0
:param ahi: The maximum length of a to check, typically len(a)
:param bhi: The maximum length of b to check, typically len(b)
:param answer: The return array. Will be filled with tuples
indicating [(line_in_a, line_in_b)]
:param maxrecursion: The maximum depth to recurse.
Must be a positive integer.
:return: None, the return value is in the parameter answer, which
should be a list
"""
if maxrecursion < 0:
print('max recursion depth reached')
# this will never happen normally, this check is to prevent DOS attacks
return
oldlength = len(answer)
if alo == ahi or blo == bhi:
return
last_a_pos = alo-1
last_b_pos = blo-1
for apos, bpos in unique_lcs_py(a[alo:ahi], b[blo:bhi]):
# recurse between lines which are unique in each file and match
apos += alo
bpos += blo
# Most of the time, you will have a sequence of similar entries
if last_a_pos+1 != apos or last_b_pos+1 != bpos:
recurse_matches_py(a, b, last_a_pos+1, last_b_pos+1,
apos, bpos, answer, maxrecursion - 1)
last_a_pos = apos
last_b_pos = bpos
answer.append((apos, bpos))
if len(answer) > oldlength:
# find matches between the last match and the end
recurse_matches_py(a, b, last_a_pos+1, last_b_pos+1,
ahi, bhi, answer, maxrecursion - 1)
elif a[alo] == b[blo]:
# find matching lines at the very beginning
while alo < ahi and blo < bhi and a[alo] == b[blo]:
answer.append((alo, blo))
alo += 1
blo += 1
recurse_matches_py(a, b, alo, blo,
ahi, bhi, answer, maxrecursion - 1)
elif a[ahi - 1] == b[bhi - 1]:
# find matching lines at the very end
nahi = ahi - 1
nbhi = bhi - 1
while nahi > alo and nbhi > blo and a[nahi - 1] == b[nbhi - 1]:
nahi -= 1
nbhi -= 1
recurse_matches_py(a, b, last_a_pos+1, last_b_pos+1,
nahi, nbhi, answer, maxrecursion - 1)
for i in xrange(ahi - nahi):
answer.append((nahi + i, nbhi + i))
def _collapse_sequences(matches):
"""Find sequences of lines.
Given a sequence of [(line_in_a, line_in_b),]
find regions where they both increment at the same time
"""
answer = []
start_a = start_b = None
length = 0
for i_a, i_b in matches:
if (start_a is not None
and (i_a == start_a + length)
and (i_b == start_b + length)):
length += 1
else:
if start_a is not None:
answer.append((start_a, start_b, length))
start_a = i_a
start_b = i_b
length = 1
if length != 0:
answer.append((start_a, start_b, length))
return answer
def _check_consistency(answer):
# For consistency sake, make sure all matches are only increasing
next_a = -1
next_b = -1
for (a, b, match_len) in answer:
if a < next_a:
raise ValueError('Non increasing matches for a')
if b < next_b:
raise ValueError('Non increasing matches for b')
next_a = a + match_len
next_b = b + match_len
class PatienceSequenceMatcher_py(difflib.SequenceMatcher):
"""Compare a pair of sequences using longest common subset."""
_do_check_consistency = True
def __init__(self, isjunk=None, a='', b=''):
if isjunk is not None:
raise NotImplementedError('Currently we do not support'
' isjunk for sequence matching')
difflib.SequenceMatcher.__init__(self, isjunk, a, b)
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 = PatienceSequenceMatcher(None, "abxcd", "abcd")
>>> s.get_matching_blocks()
[(0, 0, 2), (3, 2, 2), (5, 4, 0)]
"""
# jam 20060525 This is the python 2.4.1 difflib get_matching_blocks
# implementation which uses __helper. 2.4.3 got rid of helper for
# doing it inline with a queue.
# We should consider doing the same for recurse_matches
if self.matching_blocks is not None:
return self.matching_blocks
matches = []
recurse_matches_py(self.a, self.b, 0, 0,
len(self.a), len(self.b), matches, 10)
# Matches now has individual line pairs of
# line A matches line B, at the given offsets
self.matching_blocks = _collapse_sequences(matches)
self.matching_blocks.append( (len(self.a), len(self.b), 0) )
if PatienceSequenceMatcher_py._do_check_consistency:
if __debug__:
_check_consistency(self.matching_blocks)
return self.matching_blocks
unique_lcs = unique_lcs_py
recurse_matches = recurse_matches_py
PatienceSequenceMatcher = PatienceSequenceMatcher_py
def main(args):
import optparse
p = optparse.OptionParser(usage='%prog [options] file_a file_b'
'\nFiles can be "-" to read from stdin')
p.add_option('--patience', dest='matcher', action='store_const', const='patience',
default='patience', help='Use the patience difference algorithm')
p.add_option('--difflib', dest='matcher', action='store_const', const='difflib',
default='patience', help='Use python\'s difflib algorithm')
algorithms = {'patience':PatienceSequenceMatcher, 'difflib':difflib.SequenceMatcher}
(opts, args) = p.parse_args(args)
matcher = algorithms[opts.matcher]
if len(args) != 2:
print('You must supply 2 filenames to diff')
return -1
for line in unified_diff_files(args[0], args[1], sequencematcher=matcher):
sys.stdout.write(line)
if __name__ == '__main__':
sys.exit(main(sys.argv[1:]))
我还制作了最小独立版本 "http://neil.fraser.name/writing/diff/" rel ="noreferrer">尼尔·弗雷泽的
Edit 3: I've also made a minimally standalone version of Neil Fraser's Diff Match and Patch, I'd be very interested in a comparison of results for your use case. Again, I claim no copyrights.
我刚刚发现 DataDiff ，这可能是另一种尝试的工具.
Edit 4: I just found DataDiff, which might be another tool to try.
DataDiff是要提供的库 易于理解的python数据差异 结构.它可以处理序列 类型(列表，元组等)，集合和 字典.
DataDiff is a library to provide human-readable diffs of python data structures. It can handle sequence types (lists, tuples, etc), sets, and dictionaries.
字典和序列将是 如果适用，则递归比较.
Dictionaries and sequences will be diffed recursively, when applicable.
这篇关于文字差异JSON的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
