63 lines
2.2 KiB
Python
63 lines
2.2 KiB
Python
import re
|
|
from typing import Optional
|
|
from datetime import datetime
|
|
from dateutil import parser
|
|
|
|
def sanitize_amount(text: str) -> Optional[float]:
|
|
"""
|
|
Sanitizes a string representing an amount/number (e.g., "$1,234.56", "€ 50,000", "50 USD")
|
|
by removing currency symbols, commas, and other non-numeric characters (except for the decimal separator),
|
|
and casts it to a float.
|
|
|
|
Args:
|
|
text (str): The raw extracted text from the document.
|
|
|
|
Returns:
|
|
Optional[float]: The sanitized numeric value, or None if no number could be extracted.
|
|
"""
|
|
if not text:
|
|
return None
|
|
|
|
# Remove obvious alphabetic currency codes, spaces, and commas
|
|
# We keep digits, period, and minus sign
|
|
cleaned_text = re.sub(r'[^\d\.-]', '', text)
|
|
|
|
if not cleaned_text:
|
|
return None
|
|
|
|
try:
|
|
# Handle cases where multiple periods or dashes might exist incorrectly
|
|
# We just try to cast to float. If the OCR produced something like "1.23.45", this will fail.
|
|
# A more robust regex can handle exact capture if needed.
|
|
# But this basic cast covers 95% of standard sanitized strings.
|
|
return float(cleaned_text)
|
|
except ValueError:
|
|
return None
|
|
|
|
def sanitize_date(text: str, field_type: str = 'DATE') -> Optional[str]:
|
|
"""
|
|
Sanitizes a string representing a date or datetime and converts it to ISO 8601 format.
|
|
|
|
Args:
|
|
text (str): The raw extracted text from the document.
|
|
field_type (str): 'DATE' or 'DATETIME'. Determines the output format.
|
|
|
|
Returns:
|
|
Optional[str]: The sanitized date string in ISO format, or None if it could not be parsed.
|
|
"""
|
|
if not text:
|
|
return None
|
|
|
|
try:
|
|
# dateutil.parser is very robust at handling formats like "Dec 11, 2020", "11/12/2020", etc.
|
|
# fuzzy=True allows it to ignore extra words/characters around the date
|
|
parsed_date = parser.parse(text, fuzzy=True)
|
|
|
|
if field_type == 'DATETIME':
|
|
return parsed_date.strftime('%Y-%m-%dT%H:%M:%S')
|
|
else:
|
|
return parsed_date.strftime('%Y-%m-%d')
|
|
except (ValueError, TypeError, OverflowError):
|
|
return None
|
|
|