"""Payment Timing Optimizer — identifies early-payment discount opportunities."""
import os
import re
import pandas as pd
from models import PaymentRecommendation, PaymentOptimizationReport
def optimize_payment_timing(data_dir: str, available_cash: float = 0.0) -> dict:
"""Analyze vendor terms and recommend early-payment discounts."""
df = pd.read_csv(os.path.join(data_dir, "vendors.csv"))
recommendations = []
total_discounts = 0.0
total_early_spend = 0.0
for _, row in df.iterrows():
raw = row.get("discount_terms", "")
if pd.isna(raw) or not str(raw).strip():
continue
terms = str(raw).strip()
match = re.match(r"(\d+\.?\d*)/(\d+)\s+net\s+(\d+)", terms, re.IGNORECASE)
if not match:
continue
disc_pct = float(match.group(1))
disc_days = int(match.group(2))
net_days = int(match.group(3))
amount = float(row["outstanding_amount"])
days_accel = net_days - disc_days
if days_accel <= 0 or amount <= 0:
continue
discount_amt = round(amount * disc_pct / 100, 2)
ann_return = round((disc_pct / (100 - disc_pct)) * (365 / days_accel) * 100, 2)
rec_type = "pay_early" if ann_return > 10 else "hold"
reasoning = (f"Taking {disc_pct}% discount by paying {days_accel} days early "
f"yields {ann_return:.1f}% annualized return, "
f"{'well above' if rec_type == 'pay_early' else 'below'} "
f"10% cost of capital threshold.")
recommendations.append(PaymentRecommendation(
vendor_name=str(row["vendor_name"]),
invoice_amount=amount,
due_date=str(row.get("due_date", "")),
discount_terms=terms,
discount_amount=discount_amt,
annualized_return_pct=ann_return,
recommendation=rec_type,
reasoning=reasoning,
))
total_discounts += discount_amt
if rec_type == "pay_early":
total_early_spend += amount
recommendations.sort(key=lambda x: x.annualized_return_pct, reverse=True)
return PaymentOptimizationReport(
recommendations=recommendations,
total_discount_available=round(total_discounts, 2),
total_annualized_savings=round(total_discounts * 12, 2),
cash_required_for_early_payments=round(total_early_spend, 2),
).model_dump()
