A R R R Associates
108 lines
4.7 KiB
Python
108 lines
4.7 KiB
Python
from __future__ import annotations
|
|
|
|
import re
|
|
from pathlib import Path
|
|
|
|
import pandas as pd
|
|
import pdfplumber
|
|
|
|
from .base import BaseParser, StatementMeta, amount, extract_text, finalize, norm
|
|
from .common import date_iso, find, infer_mode
|
|
|
|
|
|
class CityUnionBankParser(BaseParser):
|
|
bank_name = "City Union Bank"
|
|
parser_name = "CityUnionBankParser"
|
|
|
|
@classmethod
|
|
def detect(cls, text: str) -> float:
|
|
upper = (text or "").upper()
|
|
score = 0.0
|
|
if "CITY UNION BANK" in upper:
|
|
score += 0.62
|
|
if "CIUB0" in upper or "IFSC :CIUB" in upper or "IFSC:CIUB" in upper:
|
|
score += 0.20
|
|
if all(token in upper for token in ("DATE", "DESCRIPTION", "CHEQUE NO", "DEBIT", "CREDIT", "BALANCE")):
|
|
score += 0.16
|
|
if "END OF STATEMENT - FROM INTERNET BANKING" in upper:
|
|
score += 0.02
|
|
return min(score, 0.99)
|
|
|
|
@staticmethod
|
|
def _reference_number(narration: str, cheque_no: str) -> str:
|
|
if norm(cheque_no):
|
|
return norm(cheque_no)
|
|
for pattern in (
|
|
r"\bUTR[:/\s-]*([A-Z0-9]{10,30})\b",
|
|
r"\bCHQ\s*NO\s*([0-9]{1,12})\b",
|
|
r"\b([A-Z]{4}H\d{10,})\b",
|
|
):
|
|
match = re.search(pattern, narration, re.I)
|
|
if match:
|
|
return match.group(1)
|
|
return ""
|
|
|
|
def parse(self, path, text=None):
|
|
pdf_path = Path(path)
|
|
text = text or extract_text(pdf_path)
|
|
meta = StatementMeta(
|
|
bank_name=self.bank_name,
|
|
source_file=pdf_path.name,
|
|
parser_name=self.parser_name,
|
|
confidence="High",
|
|
)
|
|
meta.customer_name = norm(find(r"CUSTOMER DETAILS\s*:\s*([^\n]+)", text, flags=re.I))
|
|
meta.account_number = find(r"ACCOUNT NO\s*\(15 DIGIT\)\s*:\s*([0-9]+)", text, flags=re.I)
|
|
if not meta.account_number:
|
|
meta.account_number = find(r"ACCOUNT NO\s*:\s*([A-Z0-9-]+)", text, flags=re.I)
|
|
meta.ifsc = find(r"IFSC\s*:\s*([A-Z0-9]+)", text, flags=re.I)
|
|
period = re.search(r"STATEMENT OF ACCOUNT\s+from\s+(\d{2}/\d{2}/\d{4})\s+to\s+(\d{2}/\d{2}/\d{4})", text, re.I)
|
|
if period:
|
|
meta.period_from = date_iso(period.group(1))
|
|
meta.period_to = date_iso(period.group(2))
|
|
|
|
rows: list[dict] = []
|
|
with pdfplumber.open(str(pdf_path)) as pdf:
|
|
for page_no, page in enumerate(pdf.pages, start=1):
|
|
for table in page.extract_tables() or []:
|
|
if not table:
|
|
continue
|
|
header = [norm(cell).upper() for cell in table[0]]
|
|
if not all(token in header for token in ("DATE", "DESCRIPTION", "DEBIT", "CREDIT", "BALANCE")):
|
|
continue
|
|
for raw in table[1:]:
|
|
cells = list(raw) + [None] * (6 - len(raw))
|
|
txn_date, description, cheque_no, debit, credit, balance = cells[:6]
|
|
txn_date = norm(txn_date)
|
|
if not re.fullmatch(r"\d{2}/\d{2}/\d{4}", txn_date):
|
|
continue
|
|
debit_value = amount(debit)
|
|
credit_value = amount(credit)
|
|
balance_value = amount(balance)
|
|
if debit_value is None and credit_value is None:
|
|
continue
|
|
narration = norm(description)
|
|
rows.append({
|
|
"transaction_date": txn_date,
|
|
"value_date": txn_date,
|
|
"narration": narration,
|
|
"reference_no": self._reference_number(narration, norm(cheque_no)),
|
|
"debit": debit_value,
|
|
"credit": credit_value,
|
|
"balance": balance_value,
|
|
"source_page": page_no,
|
|
"mode": infer_mode(narration),
|
|
})
|
|
|
|
data = pd.DataFrame(rows)
|
|
if not data.empty:
|
|
first = data.iloc[0]
|
|
first_balance = float(first["balance"])
|
|
first_debit = float(first["debit"]) if pd.notna(first["debit"]) else 0.0
|
|
first_credit = float(first["credit"]) if pd.notna(first["credit"]) else 0.0
|
|
meta.opening_balance = round(first_balance + first_debit - first_credit, 2)
|
|
meta.total_debit = round(float(pd.to_numeric(data["debit"], errors="coerce").fillna(0).sum()), 2)
|
|
meta.total_credit = round(float(pd.to_numeric(data["credit"], errors="coerce").fillna(0).sum()), 2)
|
|
meta.closing_balance = round(float(data.iloc[-1]["balance"]), 2)
|
|
return meta, finalize(data, meta)
|