将字边界添加到 python 正则表达式捕获组

Question

I am trying to preprocess text before parsing them to StanfordCoreNLP server. Some of my text looks like this.

" Conversion of code written in C# to Visual Basic .NET (VB.NET)."

The ".NET" confuses the server because it appears as a period and makes the single sentence into two. I wanted to replace '.' that appears in front of a word with 'DOT' so that sentence remains the same. Note that I don't want to change anything in 'VB.NET' because the StanfordCoreNLP recognizes that as one word (Proper noun).

This is what I tried so far.

print(re.sub(r"\.(\S+)", r"DOT\g<0>", text))

The result looks like this.

Conversion of code written in C# to Visual Basic DOT.NET (VBDOT.NET).

I tried adding word boundaries to the pattern r"\b\.(\S+)\b" . It didn't work.

Any help would be appreciated.

Answer 1

You can use

re.sub(r"\B\.\b", "DOT", text)

See the regex demo .

The \B\.\b regex matches a dot that is either at the start of string or immediately preceded with a non-word char, and that is followed with a word char.

See the Python demo :

import re
text = "Conversion of code written in C# to Visual Basic .NET (VB.NET)."
print( re.sub(r"\B\.\b", "DOT", text) )
# => Conversion of code written in C# to Visual Basic DOTNET (VB.NET).