在客户评论中用逗号替换一些点的正则表达式

Question

我需要写一个正则表达式，在一些病人对药物的评论中用'.'代替','。他们应该在提到副作用后用逗号，但是他们中的一些人使用点。例如：

text = "the drug side-effects are: night mare. nausea. night sweat. bad dream. dizziness. severe headache.  I suffered. she suffered. she told I should change it."

我编写了一个正则表达式代码来检测一个单词(如头痛)或两个单词(例如，恶梦)周围的两个点：

发现一个被两个点包围的词：

text=  re.sub (r'(\.)(\s*\w+\s*\.)',r',\2 ', text )

探测到两个被两个点包围的单词：

text =  re.sub (r'(\.)(\s*\w+\s\w+\s*\.)',r',\2 ', text11 )

这是输出：

the drug side-effects are: night mare, nausea,  night sweat.  bad dream, dizziness,  severe headache.   I suffered, she suffered.  she told I should change it.

但应该是：

the drug side-effects are: night mare, nausea,  night sweat,  bad dream, dizziness,  severe headache.   I suffered. she suffered.  she told I should change it.

我的代码没有在dot之后替换night sweat to ','。我加上，if a sentence starts with a subject pronoun (such as I and she) I do not want to change dot to comma after it, even if it has two words (such as, I suffered)。我不知道如何将此条件添加到代码中。

有什么建议吗？谢谢！

Answer 1

您可以使用以下模式：

\.(\s*(?!(?:i|she)\b)\w+(?:\s+\w+)?\s*)(?=[^\w\s]|$)

这匹配一个点，然后捕获一个或两个单词，其中第一个不是你提到的代词(你将需要扩大名单最有可能)。后面必须是既不是单词字符也不是空格的字符(例如，.、!、:、,)或字符串的结尾。

然后，您必须用,\1替换它。

巨蟒

import re
text = "the drug side-effects are: night mare. nausea. night sweat. bad dream. dizziness. severe headache.  I suffered. she suffered. she told I should change it."
text = re.sub(r'\.(\s*(?!(?:i|she)\b)\w+(?:\s+\w+)?\s*)(?=[^\w\s]|$)', r',\1', text, flags=re.I)
print(text)

输出

the drug side-effects are: night mare, nausea, night sweat, bad dream, dizziness, severe headache.  I suffered. she suffered. she told I should change it.

这可能不是绝对安全的，您可能需要扩展某些边缘情况下的模式。