如何实现文本按多标点分句并保留原标点？代码优化问询

Got it, let's fix this up for you! Your current code only splits on . and ends up replacing all sentence endings with periods, which isn't what you want. Plus it can't handle cases where a sentence ends without a space before the next one (like today.Are in your example).

Here's a better approach using Python's regex module to correctly identify and preserve all three punctuation marks, while cleaning up spacing issues:

import re

# Assume your input text is stored in this variable (replace with your file reading logic)
text = "I have a friend. I have many friends! I am very happy today.Are you happy? I wish you were."

# Use regex to extract all sentences ending with ., ?, or !
# The pattern matches: optional leading spaces + sentence content + ending punctuation
sentences = re.findall(r'\s*[^.!?]+[.!?]', text)

# Clean up each sentence by removing extra leading/trailing whitespace
processed_sentences = [sent.strip() for sent in sentences]

# Option 1: Write sentences joined with spaces (matches your example output)
with open("file.txt", 'w') as writefile:
    writefile.write(' '.join(processed_sentences) + '\n')

# Option 2: Write each sentence on a new line (like your original code intended)
# with open("file.txt", 'w') as writefile:
#     for sent in processed_sentences:
#         writefile.write(sent + '\n')

How This Works: #

• Regex Pattern Breakdown: r'\s*[^.!?]+[.!?]'
  • \s*: Matches any number of leading spaces (handles inconsistent spacing after punctuation)
  • [^.!?]+: Matches one or more characters that aren't ., ?, or ! (the main content of the sentence)
  • [.!?]: Matches the ending punctuation mark we want to preserve
• re.findall(): Extracts all full sentences (with their original punctuation) from the text
• strip(): Removes extra spaces from the start/end of each sentence so we don't get messy spacing in the output

Why Your Original Code Failed: #

• text.split('.') only splits on periods, so it ignores ! and ? entirely
• It discards the original punctuation and adds a period to every split segment, which breaks sentences that should end with ! or ?
• It can't fix cases where punctuation is followed directly by the next sentence (like today.Are)

Testing this with your example input will produce exactly the output you want:

Input: "I have a friend. I have many friends! I am very happy today.Are you happy? I wish you were."
Output: "I have a friend. I have many friends! I am very happy today. Are you happy? I wish you were."

内容的提问来源于stack exchange，提问作者Subha Maharjan