[英]Search and replace specific strings with floating point representations in python
Problem: I'm trying to replace mutiple specific sequences in a string with another mutiple specific sequences of floating point representations in Python.问题:我试图用 Python 中的另一个浮点表示的多个特定序列替换字符串中的多个特定序列。
I have an array of strings in a JSON-file, which I load to a python script via the json-module.我在 JSON 文件中有一个字符串数组,我通过 json 模块将其加载到 python 脚本。 The array of strings:字符串数组:
{
"LinesToReplace": [
"_ __ ___ ____ _____ ______ _______ ",
"_._ __._ ___._ ____._ _____._ ______._ ",
"_._ _.__ _.___ _.____ _._____ _.______ ",
"_._ __.__ ___.___ ____.____ _____._____ ",
"_. __. ___. ____. _____. ______. "
]
}
I load the JSON-file via the json-module:我通过 json 模块加载 JSON 文件:
with open("myFile.json") as jsonFile:
data = json.load(jsonFile)
I'm trying to replace the sequences of _
with specific substrings of floating point representations.我正在尝试用浮点表示的特定子字符串替换_
的序列。
Specification:规格:
_
or a sequence of multiple _
.要在字符串中查找的字符必须是单个_
或多个_
的序列。_
is unknown. _
序列的长度未知。_
or a sequence of multiple _
is followed by a .
如果单个_
或多个_
的序列后跟一个.
, which is again followed by a single _
or a sequence of multiple _
, the .
, 后面又跟着一个_
或多个_
的序列, .
is part of the _
-sequence.是_
序列的一部分。.
的.
is used to specify decimals用于指定小数.
如果.
isn't followed by a single _
or a sequence of multiple _
, the .
后面没有单个_
或多个_
的序列, .
is not part of the _
-sequence.不是_
序列的一部分。_
and .
_
和 的序列.
is to be replaced by floating point representations, ie, %f1.0
.将替换为浮点表示,即%f1.0
。_
- and .
表示取决于_
- 和.
-sequences. -序列。Examples:例子:
__
is to be replaced by %f2.0
. __
将被替换为%f2.0
。_.___
is to be replaced by %f1.3
. _.___
将被替换为%f1.3
。____.__
is to be replaced by %f4.2
. ____.__
将被替换为%f4.2
。___.
is to be replaced by %3.0
.将被替换为%3.0
。For the above JSON-file, the result should be:对于上面的 JSON 文件,结果应该是:
{
"ReplacedLines": [
"%f1.0 %f2.0 %f3.0 %f4.0 %f5.0 %f6.0 %f7.0 ",
"%f1.1 %f2.1 %f3.1 %f4.1 %f5.1 %f6.1 ",
"%f1.1 %f1.2 %f1.3 %f1.4 %f1.5 %f1.6 ",
"%f1.1 %f2.2 %f3.3 %f4.4 %f5.5 ",
"%f1.0. %f.0. %f3.0. %f4.0. %f5.0. %f6.0. "
]
}
Some code, which tries to replace single _
with %f1.0
(that doesn't work...):一些代码,它试图用%f1.0
替换单个_
(这不起作用......):
with open("myFile.json") as jsonFile:
data = json.load(jsonFile)
strToFind = "_"
for line in data["LinesToReplace"]:
for idl, l in enumerate(line):
if (line[idl] == strToFind and line[idl+1] != ".") and (line[idl+1] != strToFind and line[idl-1] != strToFind):
l = l[:idl] + "%f1.0" + l[idl+1:] # replace string
Any ideas on how to do this?关于如何做到这一点的任何想法? I have also though about using regular expressions.我也考虑过使用正则表达式。
EDIT编辑
The algorithm should be able to check if the character is a "_", ie to be able to format this:该算法应该能够检查字符是否为“_”,即能够格式化:
{
"LinesToReplace": [
"Ex1:_ Ex2:_. Ex3:._ Ex4:_._ Ex5:_._. ",
"Ex6:._._ Ex7:._._. Ex8:__._ Ex9: _.__ ",
"Ex10: _ Ex11: _. Ex12: ._ Ex13: _._ ",
"Ex5:._._..Ex6:.._._.Ex7:.._._._._._._._."
]
}
Solution:解决方案:
{
"LinesToReplace": [
"Ex1:%f1.0 Ex2:%f1.0. Ex3:.%f1.0 Ex4:%f1.1 Ex5:%f1.1. ",
"Ex6:.%f1.1 Ex7:.%f1.1. Ex8:%f2.1 Ex9: %f1.2 ",
"Ex10: %f1.0 Ex11: %f1.0. Ex12: .%f1.0 Ex13: %f1.1 ",
"Ex5:.%f1.1..Ex6:..%f1.1.Ex7:..%f1.1.%f1.1.%f1.1.%f1.0."
]
}
I have tried the following algorithm based on the above criteria, but I can't figure out how to implement it:我已经根据上述标准尝试了以下算法,但我无法弄清楚如何实现它:
def replaceFunc3(lines: list[str]) -> list[str]:
result = []
charToFind = '_'
charMatrix = []
# Find indicies of all "_" in lines
for line in lines:
charIndices = [idx for idx, c in enumerate(line) if c == charToFind]
charMatrix.append(charIndices)
for (line, char) in zip(lines, charMatrix):
if not char: # No "_" in current line, append the whole line
result.append(line)
else:
pass
# result.append(Something)
# TODO: Insert "%fx.x on all the placeholders"
return result
Neat problem.整洁的问题。 Personally, here is how I would do it:就个人而言,我会这样做:
from pprint import pprint
d = {
"LinesToReplace": [
"_ __ ___ ____ _____ ______ _______ ",
"_._ __._ ___._ ____._ _____._ ______._ ",
"_._ _.__ _.___ _.____ _._____ _.______ ",
"_._ __.__ ___.___ ____.____ _____._____ ",
"_. __. ___. ____. _____. ______. "
]
}
def get_replaced_lines(lines: list[str]) -> list[str]:
result = []
for line in lines:
trimmed_line = line.rstrip()
trailing_spaces = len(line) - len(trimmed_line)
underscores = trimmed_line.split()
repl_line = []
for s in underscores:
n = len(s)
if '.' in s:
if s.endswith('.'):
repl_line.append(f'%f{n - 1}.0.')
else:
idx = s.index('.')
repl_line.append(f'%f{idx}.{n - idx - 1}')
else:
repl_line.append(f'%f{n}.0')
result.append(' '.join(repl_line) + ' ' * trailing_spaces)
return result
if __name__ == '__main__':
pprint(get_replaced_lines(d['LinesToReplace']))
Output: Output:
['%f1.0 %f2.0 %f3.0 %f4.0 %f5.0 %f6.0 %f7.0 ',
'%f1.1 %f2.1 %f3.1 %f4.1 %f5.1 %f6.1 ',
'%f1.1 %f1.2 %f1.3 %f1.4 %f1.5 %f1.6 ',
'%f1.1 %f2.2 %f3.3 %f4.4 %f5.5 ',
'%f1.0. %f2.0. %f3.0. %f4.0. %f5.0. %f6.0. ']
If curious, I've also timed it at the alternate regex approach, and found this to be 40% faster overall.如果好奇的话,我还用备用正则表达式方法对它进行了计时,发现它总体上快了 40% 。 I only like this test because it proves that in general, regex is a little slower than just doing it by hand.我只喜欢这个测试,因为它证明一般来说,正则表达式比手工做要慢一点。 Though the regex approach is nice because it is certainly shorter:-)虽然正则表达式方法很好,因为它肯定更短:-)
Here is my test code:这是我的测试代码:
import re
from timeit import timeit
d = {
"LinesToReplace": [
"_ __ ___ ____ _____ ______ _______ ",
"_._ __._ ___._ ____._ _____._ ______._ ",
"_._ _.__ _.___ _.____ _._____ _.______ ",
"_._ __.__ ___.___ ____.____ _____._____ ",
"_. __. ___. ____. _____. ______. "
]
}
def get_replaced_lines(lines: list[str]) -> list[str]:
result = []
dot = '.'
space = ' '
for line in lines:
trimmed_line = line.rstrip()
trailing_spaces = len(line) - len(trimmed_line)
underscores = trimmed_line.split()
repl_line = []
for s in underscores:
n = len(s)
if dot in s:
if s[n - 1] == dot: # if last character is a '.'
repl_line.append(f'%f{n - 1}.0.')
else:
idx = s.index(dot)
repl_line.append(f'%f{idx}.{n - idx - 1}')
else:
repl_line.append(f'%f{n}.0')
result.append(space.join(repl_line) + space * trailing_spaces)
return result
def get_replaced_lines_regex(lines_to_replace):
return [re.sub(
'(_+)([.]_+)?',
lambda m: f'%f{len(m.group(1))}.{len(m.group(2) or ".")-1}',
line,
) for line in lines_to_replace]
if __name__ == '__main__':
n = 100_000
time_1 = timeit("get_replaced_lines(d['LinesToReplace'])", number=n, globals=globals())
time_2 = timeit("get_replaced_lines_regex(d['LinesToReplace'])", number=n, globals=globals())
print(f'get_replaced_lines: {time_1:.3f}')
print(f'get_replaced_lines_regex: {time_2:.3f}')
print(f'The first (non-regex) approach is faster by {(1 - time_1 / time_2) * 100:.2f}%')
assert get_replaced_lines(d['LinesToReplace']) == get_replaced_lines_regex(d['LinesToReplace'])
Results on my M1 Mac:在我的 M1 Mac 上的结果:
get_replaced_lines: 0.813
get_replaced_lines_regex: 1.359
The first (non-regex) approach is faster by 40.14%
You can use regular expression's re.sub
together with a replacement function that performs the logic on the capture groups:您可以将正则表达式的re.sub
与执行捕获组逻辑的替换 function 一起使用:
import re
def replace(line):
return re.sub(
'(_+)([.]_+)?',
lambda m: f'%f{len(m.group(1))}.{len(m.group(2) or ".")-1}',
line,
)
lines = [replace(line) for line in lines_to_replace]
Explanation of regex:正则解释:
(_+)
matches one or more underscores; (_+)
匹配一个或多个下划线; the ()
part makes them available as a capture group (the first such group, ie m.group(1)
). ()
部分使它们可用作捕获组(第一个这样的组,即m.group(1)
)。([.]_+)?
optionally matches a dot followed by one or more trailing underscores (made optional by the trailing ?
);可选地匹配一个点后跟一个或多个尾随下划线(由尾随?
变为可选); the dot is part of a character class ( []
) because otherwise it would have the special meaning "any character" .点是字符 class ( []
) 的一部分,否则它将具有特殊含义“任何字符” 。 The ()
make this part available as the second capture group ( m.group(2)
). ()
使此部分可用作第二个捕获组 ( m.group(2)
)。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.