当前位置：网站首页>Improve the readability of your regular expressions a hundred times

Improve the readability of your regular expressions a hundred times

2022-07-29 04:48:00 【The way of Python data】

author ：kingname

source ： Unheard of Code

Regular expressions , Powerful is powerful , But it's written like an emoticon . Write your own expression , Come and see in a month , I don't remember what it means . Like this one down here ：

pattern = r"((?:\(\s*)?[A-Z]*H\d+[a-z]*(?:\s*\+\s*[A-Z]*H\d+[a-z]*)*(?:\s*[\):+])?)(.*?)(?=(?:\(\s*)?[A-Z]*H\d+[a-z]*(?:\s*\+\s*[A-Z]*H\d+[a-z]*)*(?:\s*[\):+])?(?![^\w\s])|$)"

Is there any way to improve the readability of regular expressions ？ We know , One way to improve code readability is to write comments , So can regular expressions write comments ？

For example, for the following sentence ：

msg = ' My name is Qingnan , My password is :123kingname456, Please keep it confidential .'

I want to extract the password 123kingname456, Then my regular expression may be like this ：

pattern = ':(.*?),'

Can I write it like this ：

pattern = '''
:  #  Start sign 
(.*?)  # Any character starting from the next character of the start flag 
,  # Stop when you encounter an English comma 
'''

It's much clearer to write in this way , The function of each part is clear .

But obviously, nothing can be extracted directly , As shown in the figure below ：

But I'm shopping today Python Regular expression documents ^[1] When , Found a good thing ：

Use it , You can make your regular expressions have comments , As shown in the figure below ：

re.VERBOSE It can also be abbreviated as re.X, As shown in the figure below ：

Complex regular expressions at the beginning of this article , After using comments , It will become more readable ：

pattern = r"""
(                       # code (capture)
    # BEGIN multicode

    (?: \( \s* )?       # maybe open paren and maybe space

    # code
    [A-Z]*H  # prefix
    \d+      # digits
    [a-z]*   # suffix

    (?:                 # maybe followed by other codes,
        \s* \+ \s*      # ... plus-separated

        # code
        [A-Z]*H  # prefix
        \d+      # digits
        [a-z]*   # suffix
    )*

    (?: \s* [\):+] )?   # maybe space and maybe close paren or colon or plus

    # END multicode
)

( .*? )                 # message (capture): everything ...

(?=                     # ... up to (but excluding) ...
    # ... the next code

    # BEGIN multicode

    (?: \( \s* )?       # maybe open paren and maybe space

    # code
    [A-Z]*H  # prefix
    \d+      # digits
    [a-z]*   # suffix

    (?:                 # maybe followed by other codes,
        \s* \+ \s*      # ... plus-separated

        # code
        [A-Z]*H  # prefix
        \d+      # digits
        [a-z]*   # suffix
    )*

    (?: \s* [\):+] )?   # maybe space and maybe close paren or colon or plus

    # END multicode

        # (but not when followed by punctuation)
        (?! [^\w\s] )

    # ... or the end
    | $
)
"""