20
votes

I need to use regex to strip punctuation at the start and end of a word. It seems like regex would be the best option for this. I don't want punctuation removed from words like 'you're', which is why I'm not using .replace().

4
For those coming here looking for a way to distinguish between Unicode alphanumeric characters and everything else, while using Python 3.x, you can just use \w and \W in your regular expression. This just helped me code the Control-Shift-Left/Right functionality in a Tkinter text widget (to skip past all the stuff like punctuation before a word). I found your post before I found that solution. So, I figure it may help someone in a similar boat. - Brōtsyorfuzthrāx

4 Answers

55
votes

You don't need regular expression to do this task. Use str.strip with string.punctuation:

>>> import string
>>> string.punctuation
'!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~'
>>> '!Hello.'.strip(string.punctuation)
'Hello'

>>> ' '.join(word.strip(string.punctuation) for word in "Hello, world. I'm a boy, you're a girl.".split())
"Hello world I'm a boy you're a girl"
2
votes

I think this function will be helpful and concise in removing punctuation:

import re
def remove_punct(text):
    new_words = []
    for word in text:
        w = re.sub(r'[^\w\s]','',word) #remove everything except words and space
        w = re.sub(r'_','',w) #how to remove underscore as well
        new_words.append(w)
    return new_words
0
votes

If you persist in using Regex, I recommend this solution:

import re
import string
p = re.compile("[" + re.escape(string.punctuation) + "]")
print(p.sub("", "\"hello world!\", he's told me."))
### hello world hes told me

Note also that you can pass your own punctuation marks:

my_punct = ['!', '"', '#', '$', '%', '&', "'", '(', ')', '*', '+', ',', '.',
           '/', ':', ';', '<', '=', '>', '?', '@', '[', '\\', ']', '^', '_', 
           '`', '{', '|', '}', '~', '»', '«', '“', '”']

punct_pattern = re.compile("[" + re.escape("".join(my_punct)) + "]")
re.sub(punct_pattern, "", "I've been vaccinated against *covid-19*!") # the "-" symbol should remain
### Ive been vaccinated against covid-19
-1
votes

You can remove punctuation from a text file or a particular string file using regular expression as follows -

new_data=[]
with open('/home/rahul/align.txt','r') as f:
    f1 = f.read()
    f2 = f1.split()



    all_words = f2 
    punctuations = '''!()-[]{};:'"\,<>./?@#$%^&*_~''' 
    # You can add and remove punctuations as per your choice 
    #removing stop words in hungarian text and  english text and 
    #display the unpunctuated string
    # To remove from a string, replace new_data with new_str 
    # new_str = "My name$#@ is . rahul -~"

    for word in all_words: 
        if word not in punctuations:
           new_data.append(word)

    print (new_data)

P.S. - Do the identation properly as per required. Hope this helps!!