Files
scripts/cleaners/nationwide-cleaner.py~
2026-02-11 13:31:50 +00:00

105 lines
2.6 KiB
Python

#!/usr/bin/env python3
import csv
from decimal import Decimal, InvalidOperation
import sys
from pathlib import Path
from datetime import datetime
def clean_money(value: str) -> Decimal | None:
if not value:
return None
value = (
value.replace("£", "")
.replace(",", "")
.strip()
)
if not value:
return None
try:
return Decimal(value)
except InvalidOperation:
return None
def parse_date(value: str) -> str | None:
"""
Convert '01 Dec 2025''12/01/2025'
"""
value = value.strip()
if not value:
return None
try:
dt = datetime.strptime(value, "%d %b %Y")
return dt.strftime("%m/%d/%Y")
except ValueError:
return None
def convert(in_path: Path, out_path: Path):
with open(in_path, newline="", encoding="latin-1") as f_in, \
open(out_path, "w", newline="", encoding="utf-8") as f_out:
reader = csv.reader(f_in)
# Skip metadata lines until we hit the header
for row in reader:
if row and row[0] == "Date":
header = row
break
else:
raise RuntimeError("Could not find Nationwide CSV header row")
dict_reader = csv.DictReader(f_in, fieldnames=header)
fieldnames = ["Date", "Payee", "Description", "Amount"]
writer = csv.DictWriter(f_out, fieldnames=fieldnames)
writer.writeheader()
for row in dict_reader:
raw_date = row.get("Date") or ""
date = parse_date(raw_date)
if not date:
continue
desc = (row.get("Description") or "").strip()
payee = desc
paid_out = clean_money(row.get("Paid out", ""))
paid_in = clean_money(row.get("Paid in", ""))
if paid_out is not None:
amount = -paid_out
elif paid_in is not None:
amount = paid_in
else:
continue
writer.writerow({
"Date": date,
"Payee": payee,
"Description": desc,
"Amount": f"{amount:.2f}",
})
def main():
if len(sys.argv) < 2:
print("Usage: python nationwide-cleaner.py input.csv [output.csv]")
sys.exit(1)
in_path = Path(sys.argv[1])
out_path = Path(sys.argv[2]) if len(sys.argv) >= 3 else in_path.with_name(
in_path.stem + "_actual.csv"
)
convert(in_path, out_path)
print(f"Written cleaned file to {out_path}")
if __name__ == "__main__":
main()