如何用Python去除字符串边缘的动词缩写与标点等噪声

Got it, let's figure out how to solve this problem in Python. The core requirement here is to strip only the noise from the edges of a string (or word sequence) while leaving any internal noise completely untouched—regular expressions are perfect for this job, since we can target specific patterns at the start and end of the string without affecting the middle.

Approach #

We need to handle two key parts:

1. Leading noise: Remove all non-alphanumeric characters (like $, ¿, !, etc.) from the start of the string.
2. Trailing noise: Remove both verb abbreviations (like 's, 're, 'm) and non-alphanumeric punctuation (like ., ?, ,) from the end of the string.

We'll use targeted regex substitutions to zero in on these edge cases specifically.

Python Implementation #

import re

def clean_edge_noise(input_str):
    # Step 1: Remove leading non-alphanumeric characters (punctuation, symbols, etc.)
    cleaned = re.sub(r'^[^a-zA-Z0-9]+', '', input_str)
    
    # Step 2: Remove trailing verb abbreviations + non-alphanumeric characters
    # Covers abbreviations like 's, 're, 'm, 'd, 'll, 've (including smart quotes ’)
    cleaned = re.sub(r'([\'’](s|re|m|d|ll|ve))?[^a-zA-Z0-9]+$', '', cleaned)
    
    # Handle edge case: String ends only with an abbreviation (no trailing punctuation)
    cleaned = re.sub(r'([\'’](s|re|m|d|ll|ve))$', '', cleaned)
    
    return cleaned

How It Works #

Let's break down the regex patterns to understand what's happening:

• Leading noise pattern: ^[^a-zA-Z0-9]+

  • ^ anchors the match to the very start of the string.
  • [^a-zA-Z0-9] matches any character that's not a letter or number (our definition of leading noise).
  • + ensures we remove all consecutive leading noise characters in one go.

• Trailing noise pattern: ([\'’](s|re|m|d|ll|ve))?[^a-zA-Z0-9]+$

  • ([\'’](s|re|m|d|ll|ve))? matches optional verb abbreviations (supports both straight ' and smart ’ quotes). The ? makes this part optional so we can handle strings that end with just punctuation.
  • [^a-zA-Z0-9]+ matches any trailing non-alphanumeric punctuation.
  • $ anchors the match to the very end of the string.

• Final abbreviation check: ([\'’](s|re|m|d|ll|ve))$ handles cases where a string ends only with an abbreviation (like "They're" or "I'm") with no trailing punctuation.

Test the Function #

Let's verify with your example cases (plus a few extra edge cases to cover all bases):

test_cases = [
    ("Apple.", "Apple"),
    ("Trump's", "Trump"),  # Corrected your likely typo from "Donald Trump's" → "Trump"
    ("They're", "They"),
    ("I'm", "I"),
    ("¿Hablas espanol?", "Hablas espanol"),
    ("$12", "12"),
    ("H4ck3r", "H4ck3r"),
    ("What's up", "What's up"),
    ("!@#World's?;", "World"),
    ("''Hello''", "Hello"),
    ("Don't stop", "Don't stop")
]

for input_str, expected in test_cases:
    result = clean_edge_noise(input_str)
    status = "✅ Pass" if result == expected else "❌ Fail"
    print(f"Input: {repr(input_str)} → Output: {repr(result)} | {status}")

Running this will confirm the function behaves exactly as required—internal noise like the 't in "Don't stop" stays intact, while only edge noise gets stripped.

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