Create a script to clean email signatures

Here is a Python script to clean email signatures from a block of email text. It identifies and removes common email signature patterns using heuristics such as delimiters (--, Thanks, Regards, etc.), excessive contact info, and short lines at the end of the email body:

python
CopyEdit
import re

def clean_email_signature(email_body: str) -> str:
    # Normalize line endings
    email_body = email_body.replace('rn', 'n').strip()

    # Common signature delimiters
    signature_delimiters = [
        r'^--s*$',                           # "--" on its own line
        r'^__+$',                             # "____" line
        r'^thanks[.,s]*$',                  # "Thanks", "Thanks."
        r'^regards[.,s]*$',                 # "Regards", "Regards,"
        r'^best[s]+(regards|wishes)[.,s]*$',  # "Best regards"
        r'^cheers[.,s]*$',                  # "Cheers"
        r'^sincerely[.,s]*$',               # "Sincerely"
    ]

    # Compile all delimiter regexes
    signature_patterns = [re.compile(p, re.IGNORECASE) for p in signature_delimiters]

    # Split the email into lines for analysis
    lines = email_body.split('n')
    cleaned_lines = []
    found_signature = False

    # Start from the bottom and work upward to detect signature block
    for i in range(len(lines) - 1, -1, -1):
        line = lines[i].strip()

        # Stop if we've already found a signature
        if found_signature:
            continue

        # Check if line matches any known signature delimiter
        if any(p.match(line) for p in signature_patterns):
            found_signature = True
            continue

        # Heuristics: Skip phone/email/contact lines
        if re.search(r'(phone|mobile|email|website|linkedin|twitter|facebook|contact)', line, re.IGNORECASE):
            continue

        # Heuristic: Skip if line is name-like (short and capitalized words)
        if re.match(r'^([A-Z][a-z]+s?){1,3}$', line):
            continue

        # Once we decide to keep lines, prepend to cleaned list
        cleaned_lines.insert(0, line)

    return 'n'.join(cleaned_lines).strip()

# Example usage
if __name__ == "__main__":
    email_text = """
Hi John,

Thanks for the update. I’ll take a look and get back to you.

Best regards,  
Sarah Thompson  
Senior Developer  
ACME Corp  
Phone: 123-456-7890  
Email: sarah@acme.com
"""
    print("Cleaned Email:n")
    print(clean_email_signature(email_text))

Key Features:

• Recognizes common signature delimiters and phrases.

• Filters lines with contact info keywords.

• Ignores name-like patterns often seen in signatures.

• Processes text from the bottom up to detect trailing signatures.

You can further refine this by incorporating ML-based approaches or integrating with libraries like talon or replyto if needed.