add revolut

nlb-export/nlb.py

@@ -2,6 +2,7 @@ import csv
 import re
 import subprocess
 import sys
+from decimal import Decimal
 from pathlib import Path
 # extract transaction from izpisek pdf
 TRANSACTION_START = re.compile(r"^\d{2}\.\d{2}\.\d{2}\s")
@@ -24,6 +25,14 @@ def pdf_to_text(pdf_path):
     return result.stdout
 
 
+def to_number(text):
+    """Statements write 1.234,56 and mark a negative balance with a trailing minus."""
+    text = text.strip()
+    if text.endswith("-"):
+        text = "-" + text[:-1]
+    return Decimal(text.replace(".", "").replace(",", "."))
+
+
 def parse_transaction(line):
     m = TRANSACTION_LINE.match(line.rstrip())
     account = ""
@@ -33,15 +42,12 @@ def parse_transaction(line):
             account = f
         else:
             desc_parts.append(f)
-    balance = m.group("balance")
-    if balance.endswith("-"):
-        balance = "-" + balance[:-1]
     return {
         "date": m.group("date"),
         "desc": " ".join(desc_parts),
         "account": account,
-        "amount": m.group("amount"),
-        "balance": balance,
+        "amount": f"{to_number(m.group('amount')):+.2f}",
+        "balance": f"{to_number(m.group('balance')):.2f}",
         "currency": "EUR",
     }
 
@@ -70,13 +76,25 @@ def extract_transactions(pdf_path):
     return rows
 
 
+def check_balances(rows, path):
+    """Every balance must be the previous one plus the transaction amount."""
+    for prev, row in zip(rows, rows[1:]):
+        expected = Decimal(prev["balance"]) + Decimal(row["amount"])
+        if expected != Decimal(row["balance"]):
+            print(f"  warning: {path.name} {row['date']} {row['desc'][:40]!r} "
+                  f"balance {row['balance']} != {expected}")
+
+
 def main():
     pdf_dir = Path(sys.argv[1])
     out_path = pdf_dir.with_suffix(".csv")
 
     all_rows = []
     for pdf_path in sorted(pdf_dir.glob("*.pdf")):
-        all_rows.extend(extract_transactions(pdf_path))
+        rows = extract_transactions(pdf_path)
+        # consecutive izpiski of one account, so the chain carries over the file boundary
+        check_balances(all_rows[-1:] + rows, pdf_path)
+        all_rows.extend(rows)
 
     with open(out_path, "w", newline="") as f:
         writer = csv.DictWriter(f, fieldnames=["date", "desc", "account", "amount", "balance", "currency"])