以下代码表明 FMc 的代码不起作用。
该行
from name_of_file import olding_the_new,finding 指的是我在此线程中对此问题的个人回答中的代码。
* 将名称 name_of_file 命名为包含我的代码脚本的文件（位于我在此线程中的其他答案中），它将运行。
* 或者，如果您对复制粘贴我的代码感到厌烦，只需注释这行导入，下面的代码就会运行，因为我放置了一条 try-except 指令，该指令将对olding_the_new 和finding 的缺失做出正确反应
我用了两种方法来验证FMc的代码的结果：
-1/ 比较他的代码返回的跨度与 'f' 的索引和 'r' 的索引，当我们搜索短语“foobar”时，我在那里管理没有 f 和 r 除了 foobar
-2/ 与我的代码返回的第一个跨度比较，因此需要从name_of_file 进行上述导入
Nota Bene
如果将disp = None 更改为disp == True，则执行会显示有助于理解算法的中间结果。
.
import re
from name_of_file import olding_the_new,finding
def main():
# Two versions of the text: the original,
# and one without any of the "#" markers.
for text_orig in ('BLAH ##BLAH fo####o##ba###r## BL##AH',
'jkh##jh#f',
'#f#oo##ba###r##',
'a##xf#oo##ba###r##',
'ax##f#oo##ba###r##',
'ab###xyf#oo##ba###r##',
'abx###yf#oo##ba###r##',
'abxy###f#oo##ba###r##',
'iji#hkh#f#oo##ba###r##',
'mn##pps#f#oo##ba###r##',
'mn##pab###xyf#oo##ba###r##',
'lmn#pab###xyf#oo##ba###r##',
'fo##o##ba###r## aaaaaBLfoob##arAH',
'fo#o##ba####r## aaaaaBLfoob##ar#AH',
'f##oo##ba###r## aaaaaBLfoob##ar',
'f#oo##ba####r## aaaaBL#foob##arAH',
'f#oo##ba####r## aaaaBL#foob##ar#AH',
'foo##ba#####r## aaaaBL#foob##ar',
'#f#oo##ba###r## aaaBL##foob##arAH',
'#foo##ba####r## aaaBL##foob##ar#AH',
'#af#oo##ba##r## aaaBL##foob##ar',
'##afoo##ba###r## aaaaaBLfoob##arAH',
'BLAHHfo##o##ba###r aaBLfoob##ar#AH',
'BLAH#fo##o##ba###r aaBLfoob##ar',
'BLA#Hfo##o##ba###r###BLfoob##ar',
'BLA#Hfo##o##ba###r#BL##foob##ar',
):
text_clean = text_orig.replace('#', '')
# Collect data on the positions and widths
# of the markers in the original text.
rgx = re.compile(r'#+')
markers = [(m.start(), len(m.group()))
for m in rgx.finditer(text_orig)]
# Find the location of the search phrase in the cleaned text.
# At that point you'll have all the data you need to compute
# the span of the phrase in the original text.
search = 'foobar'
try:
i = text_clean.index(search)
print ('text_clean == %s\n'
"text_clean.index('%s')==%d len('%s') == %d\n"
'text_orig == %s\n'
'markers == %s'
% (text_clean,
search,i,search,len(search),
text_orig,
markers))
S,E = compute_span(i, len(search), markers)
print "span = (%d,%d) %s %s %s"\
% (S,E,
text_orig.index('f')==S,
text_orig.index('r')+1==E,
list(finding(search,text_orig,'#+')))
except ValueError:
print ('text_clean == %s\n'
"text_clean.index('%s') ***Not found***\n"
'text_orig == %s\n'
'markers == %s'
% (text_clean,
search,
text_orig,
markers))
print '--------------------------------'
.
def compute_span(start, width, markers):
# start and width are in expurgated text
# markers are in original text
disp = None # if disp==True => displaying of intermediary results
span_start = start
if disp:
print ('\nAt beginning in compute_span():\n'
' span_start==start==%d width==%d'
% (start,width))
for s, w in markers: # s and w are in original text
if disp:
print ('\ns,w==%d,%d'
' s+w-1(%d)<start(%d) %s'
' s(%d)==start(%d) %s'
% (s,w,s+w-1,start,s+w-1<start,s,start,s==start))
if s + w - 1 < start:
#mwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmmwmwmwmwmwm
# the following if-else section is justified to be used
# only after correction of the above line to this one:
# if s+w-1 <= start or s==start:
#mwmwmwmwmwmwmwmwmwmwmwmwmwmwmwm
if s + w - 1 <= start and disp:
print ' 1a) s + w - 1 (%d) <= start (%d) marker at left'\
% (s+w-1, start)
elif disp:
print ' 1b) s(%d) == start(%d)' % (s,start)
#mwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmwmmwmwmwmwmwm
# Situation: marker fully to left of our text.
# Adjust our start points rightward.
start += w
span_start += w
if disp:
print ' span_start == %d start, width == %d, %d' % (span_start, start, width)
elif start + width - 1 < s:
if disp:
print (' 2) start + width - 1 (%d) < s (%d) marker at right\n'
' break' % (start+width-1, s))
# Situation: marker fully to the right of our text.
break
else:
# Situation: marker interrupts our text.
# Advance the start point for the remaining text
# rightward, and reduce the remaining width.
if disp:
print " 3) In 'else': s - start == %d marker interrupts" % (s - start)
start += w
width = width - (s - start)
if disp:
print ' span_start == %d start, width == %d, %d' % (span_start, start, width)
return (span_start, start + width)
.
main()
结果
>>>
text_clean == BLAH BLAH foobar BLAH
text_clean.index('foobar')==10 len('foobar') == 6
text_orig == BLAH ##BLAH fo####o##ba###r## BL##AH
markers == [(5, 2), (14, 4), (19, 2), (23, 3), (27, 2), (32, 2)]
span = (12,26) True False [(12, 27)]
--------------------------------
text_clean == jkhjhf
text_clean.index('foobar') ***Not found***
text_orig == jkh##jh#f
markers == [(3, 2), (7, 1)]
--------------------------------
text_clean == foobar
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == #f#oo##ba###r##
markers == [(0, 1), (2, 1), (5, 2), (9, 3), (13, 2)]
span = (0,11) False False [(1, 13)]
--------------------------------
text_clean == axfoobar
text_clean.index('foobar')==2 len('foobar') == 6
text_orig == a##xf#oo##ba###r##
markers == [(1, 2), (5, 1), (8, 2), (12, 3), (16, 2)]
span = (2,16) False True [(4, 16)]
--------------------------------
text_clean == axfoobar
text_clean.index('foobar')==2 len('foobar') == 6
text_orig == ax##f#oo##ba###r##
markers == [(2, 2), (5, 1), (8, 2), (12, 3), (16, 2)]
span = (2,15) False False [(4, 16)]
--------------------------------
text_clean == abxyfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == ab###xyf#oo##ba###r##
markers == [(2, 3), (8, 1), (11, 2), (15, 3), (19, 2)]
span = (4,19) False True [(7, 19)]
--------------------------------
text_clean == abxyfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == abx###yf#oo##ba###r##
markers == [(3, 3), (8, 1), (11, 2), (15, 3), (19, 2)]
span = (4,18) False False [(7, 19)]
--------------------------------
text_clean == abxyfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == abxy###f#oo##ba###r##
markers == [(4, 3), (8, 1), (11, 2), (15, 3), (19, 2)]
span = (4,19) False True [(7, 19)]
--------------------------------
text_clean == ijihkhfoobar
text_clean.index('foobar')==6 len('foobar') == 6
text_orig == iji#hkh#f#oo##ba###r##
markers == [(3, 1), (7, 1), (9, 1), (12, 2), (16, 3), (20, 2)]
span = (7,18) False False [(8, 20)]
--------------------------------
text_clean == mnppsfoobar
text_clean.index('foobar')==5 len('foobar') == 6
text_orig == mn##pps#f#oo##ba###r##
markers == [(2, 2), (7, 1), (9, 1), (12, 2), (16, 3), (20, 2)]
span = (7,18) False False [(8, 20)]
--------------------------------
text_clean == mnpabxyfoobar
text_clean.index('foobar')==7 len('foobar') == 6
text_orig == mn##pab###xyf#oo##ba###r##
markers == [(2, 2), (7, 3), (13, 1), (16, 2), (20, 3), (24, 2)]
span = (9,24) False True [(12, 24)]
--------------------------------
text_clean == lmnpabxyfoobar
text_clean.index('foobar')==8 len('foobar') == 6
text_orig == lmn#pab###xyf#oo##ba###r##
markers == [(3, 1), (7, 3), (13, 1), (16, 2), (20, 3), (24, 2)]
span = (9,24) False True [(12, 24)]
--------------------------------
text_clean == foobar aaaaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == fo##o##ba###r## aaaaaBLfoob##arAH
markers == [(2, 2), (5, 2), (9, 3), (13, 2), (27, 2)]
span = (0,9) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == fo#o##ba####r## aaaaaBLfoob##ar#AH
markers == [(2, 1), (4, 2), (8, 4), (13, 2), (27, 2), (31, 1)]
span = (0,7) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaaaBLfoobar
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == f##oo##ba###r## aaaaaBLfoob##ar
markers == [(1, 2), (5, 2), (9, 3), (13, 2), (27, 2)]
span = (0,11) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == f#oo##ba####r## aaaaBL#foob##arAH
markers == [(1, 1), (4, 2), (8, 4), (13, 2), (22, 1), (27, 2)]
span = (0,8) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == f#oo##ba####r## aaaaBL#foob##ar#AH
markers == [(1, 1), (4, 2), (8, 4), (13, 2), (22, 1), (27, 2), (31, 1)]
span = (0,8) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaaBLfoobar
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == foo##ba#####r## aaaaBL#foob##ar
markers == [(3, 2), (7, 5), (13, 2), (22, 1), (27, 2)]
span = (0,7) True False [(0, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == #f#oo##ba###r## aaaBL##foob##arAH
markers == [(0, 1), (2, 1), (5, 2), (9, 3), (13, 2), (21, 2), (27, 2)]
span = (0,11) False False [(1, 13), (23, 31)]
--------------------------------
text_clean == foobar aaaBLfoobarAH
text_clean.index('foobar')==0 len('foobar') == 6
text_orig == #foo##ba####r## aaaBL##foob##ar#AH
markers == [(0, 1), (4, 2), (8, 4), (13, 2), (21, 2), (27, 2), (31, 1)]
span = (0,12) False False [(1, 13), (23, 31)]
--------------------------------
text_clean == afoobar aaaBLfoobar
text_clean.index('foobar')==1 len('foobar') == 6
text_orig == #af#oo##ba##r## aaaBL##foob##ar
markers == [(0, 1), (3, 1), (6, 2), (10, 2), (13, 2), (21, 2), (27, 2)]
span = (2,10) True False [(2, 13), (23, 31)]
--------------------------------
text_clean == afoobar aaaaaBLfoobarAH
text_clean.index('foobar')==1 len('foobar') == 6
text_orig == ##afoo##ba###r## aaaaaBLfoob##arAH
markers == [(0, 2), (6, 2), (10, 3), (14, 2), (28, 2)]
span = (1,14) False True [(3, 14), (24, 32)]
--------------------------------
text_clean == BLAHHfoobar aaBLfoobarAH
text_clean.index('foobar')==5 len('foobar') == 6
text_orig == BLAHHfo##o##ba###r aaBLfoob##ar#AH
markers == [(7, 2), (10, 2), (14, 3), (27, 2), (31, 1)]
span = (5,14) True False [(5, 18), (23, 31)]
--------------------------------
text_clean == BLAHfoobar aaBLfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == BLAH#fo##o##ba###r aaBLfoob##ar
markers == [(4, 1), (7, 2), (10, 2), (14, 3), (27, 2)]
span = (4,16) False False [(5, 18), (23, 31)]
--------------------------------
text_clean == BLAHfoobarBLfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == BLA#Hfo##o##ba###r###BLfoob##ar
markers == [(3, 1), (7, 2), (10, 2), (14, 3), (18, 3), (27, 2)]
span = (5,14) True False [(5, 18), (23, 31)]
--------------------------------
text_clean == BLAHfoobarBLfoobar
text_clean.index('foobar')==4 len('foobar') == 6
text_orig == BLA#Hfo##o##ba###r#BL##foob##ar
markers == [(3, 1), (7, 2), (10, 2), (14, 3), (18, 1), (21, 2), (27, 2)]
span = (5,14) True False [(5, 18), (23, 31)]
--------------------------------
>>>
.
---------------------------------------------- H3>
FMc的代码很微妙，我花了好长时间才明白它的原理，然后才能够改正。
我会让任何人理解算法的任务。我只说使 FMc 的代码正常工作所需的更正：
.
第一次更正：
if s + w - 1 < start:
# must be changed to
if s + w - 1 <= start or (s==start):
编辑
在我最初的回答中，
我写了... or (s<=start)。
那是我的错误，其实我是有意写的
.. or (s==start)
关于此编辑的注意事项：
这个错误在用我在这里描述的两个更正来更正FMc的初始代码的代码中没有任何后果（第一个，因为目前他已经更改了两次）。
事实上，如果你用这两个更正来更正代码，你将获得正确的结果，所有 25 个例子都是为 text_orig 以及 ... or (s <= start) 和 ... or (s==start)。
所以我认为当第一个条件s+w-1 <= start为False时，s < start为True的情况永远不会发生，大概是基于w总是大于0的事实以及由于配置的其他一些原因标记和非标记序列.....
所以我试图找到这种印象的示范......但我失败了。
此外，我达到了一种我什至不再了解 FMc 的算法的状态（他进行任何编辑之前的第一个算法）！！
尽管如此，我还是让这个答案保持原样，并在这个答案的末尾发布解释，试图解释为什么需要这些更正。
但我警告：FMc 的第一个算法非常古怪且难以理解，因为它会比较属于两个不同字符串的索引，一个是带有标记 #### 的 text_orig，另一个是已清理的所有这些标记....现在我不再相信这可能有意义....
.
第二次更正::
start += w
width = width - (s - start)
# must be changed to
width -= (s-start) # this line MUST BE before the following one
start = s + w # because start += (s-start) + w
-------
我很惊讶有 2 个人支持 FMc 的答案，尽管它给出了错误的代码。这意味着他们在没有测试给定代码的情况下对答案进行了投票。
----------------------------------------
.
编辑
为什么一定要把条件if s + w - 1 < start:改成这个：
if s + w - 1 <= start or (s==start):？
因为s + w - 1 < start 应该是 False 并且 s 等于 start 可能会发生。
在这种情况下，执行转到else 部分并执行（在更正的代码中）：
width -= (s - start)
start = s + w
因此，width 不会改变，而当我们看到相关序列时它显然应该改变。
这种情况可能发生在检查第一个标记时，如下所示：
'#f#oo##ba###r##' : s,w==0,1 , 0==s==start==0
'ax##f#oo##ba###r##' : s,w==2,2 , 2==s==start==2
'abxy###f#oo##ba###r##' : s,w==4,3 , 4==s==start==4
'#f#oo##ba###r## aaaBL##foob##arAH' : s,w==0,1 , 0==s==start==0
'BLAH#fo##o##ba###r aaBLfoob##ar' : s,w==4,1 4==s==start==4
对于第二个标记的检查发生以下情况：
'iji#hkh#f#oo##ba###r##' : s,w==7,1 , 7==s==start==7
'mn##pps#f#oo##ba###r##' : s,w==7,1 , 7==s==start==7
通过设置disp = True执行我的代码可以更好地理解。
当s + w - 1 <= start 被验证时，s 可能等于start 的事实并不麻烦，因为执行不会转到else 部分，它会转到第一个部分，其中只有将w 添加到s 和start。
但是当s + w - 1 <= start 为假而s 等于start 时，执行将转到else 部分，其中指令width -= (s-start) 的执行不会将任何内容更改为width 的值这很麻烦。
因此，必须添加条件or (s==start) 以防止出现此else 目标，并且需要将其放在or 之后以防止出现此目标，即使s+w-1 <= start 为False，如某些示例所示，这可能会发生。
.
关于指令s+w-1 < start必须更改为s+w-1 <= start（带=），
这是因为w==1 的大小写只对应 1 个字符 # ,
至于案例
mn##pps#f#oo##ba###r##（第二个标记）
和BLAH#fo##o##ba###r（第一个标记）。