Replace character with another only if repeated and not part of a word

Question

In Python3, the following code works to replace a string (two or more) of *'s with x's.

import re
re.sub(r'\*(?=\*)|(?<=\*)\*', 'x', 'Replace this *** but not this *')
# 'Replace this xxx but not this *'

But what if I also want to exempt a string of *'s that are part of a "word" like the following? (I.e. the string is attached to one or more [a-zA-Z] characters.)

text = "Don't replace foo*** or **bar, either."
# unmodified text expected

How do I do this? I can probably match the exempted cases as well and use a replacement function to deal with them, but is there a better way?

Answer 1

regex = r"\s\*{2,}[\s\n]"

This matches 2 or more * chars, surrounded by whitespace (or ended with a newline)

Call it maybe like this?

regex = r"\s\*{2,}[\s\n]"


def replacer(match):
    return 'x' * len(match.group())

re.sub(regex, replacer, your_string_here)

Answer 2

This answer is inspired Danielle M.'s. This pattern below seems to give me what I want. The rest is the same as hers.

regex = r'(?<![a-zA-Z])\*{2,}(?![a-zA-Z])'