samples/generate_samples.py

"""Generate synthetic bank statements + ground-truth JSON.

Produces:
  samples/sample_digital.pdf        -- real reportlab table (exercises pdfplumber)
  samples/sample_scan.png           -- rendered + rotated + noised (exercises MiniCPM-V)
  samples/sample_digital_truth.json -- ground-truth transactions (test fixture)
  samples/sample_scan_truth.json    -- same data, used for vision-accuracy check

No real bank data is used (privacy + licensing). Narrations are deliberately
messy to mimic real Indian bank statements.

Run:  python samples/generate_samples.py
"""

import json
import os
import random

HERE = os.path.dirname(os.path.abspath(__file__))

OPENING_BALANCE = 152340.50

# (narration, debit, credit) -- one tuple per transaction row.
# Narrations are intentionally messy and cover most ledger heads so the
# rules layer + reconciliation both get a real workout.
RAW = [
    ("UPI/CR/512345678901/RAMESH TRADERS/PAYTM/Payment for goods", None, 45000.00),
    ("NEFT-HDFC0001234-ACME PVT LTD-INV 4521", None, 128500.00),
    ("UPI/DR/423451234567/SHARMA RENT/GPAY/April office rent", 35000.00, None),
    ("ACH/NACH/SBILIFE INSURANCE/PREMIUM/POL 9921", 12500.00, None),
    ("SALARY APR 2026 - STAFF PAYROLL BATCH 04", 86000.00, None),
    ("ATM CSH WDL/AXIS/KANPUR MALL ROAD/04APR", 10000.00, None),
    ("UPI/CR/612345098765/SUNIL ENTERPRISES/PHONEPE/sale", None, 22750.00),
    ("GST PMT-CBIC-26AAAPL1234C1Z5-GSTR3B APR", 18900.00, None),
    ("NEFT-ICIC0004567-MEGHA SUPPLIERS-PO 7781", 64200.00, None),
    ("INT.PD-SAVINGS A/C-QTR ENDING", None, 1842.00),
    ("BANK CHRG-SMS CHGS APR 2026", 23.60, None),
    ("UPI/DR/734512345098/BSNL BROADBAND/INTERNET BILL", 1499.00, None),
    ("RTGS-UTIB0000123-DELTA EXPORTS-SETTLEMENT", None, 215000.00),
    ("ACH/NACH/HDFC HOME LOAN/EMI/LN44219", 28750.00, None),
    ("UPI/DR/845612340987/KESHAV ELECTRICALS/utility KESCO bill", 6540.00, None),
    ("CSH DEP/BRANCH/KANPUR/CASH DEPOSIT BY CLIENT", None, 50000.00),
    ("TDS PMT-CPC-TDS Q4 26AAAPL1234C", 9400.00, None),
    ("UPI/CR/923451234560/VIKAS AGENCY/sale receipt", None, 17800.00),
    ("NEFT-PUNB0123456-RAVI CONSULTING-PROF FEES", 25000.00, None),
    ("IRCTC TRAVEL/RAIL TICKET/PNR 8841200012", 3450.00, None),
    ("UPI/DR/103451234561/SWIGGY/staff lunch conveyance", 1280.00, None),
    ("ELECTRICITY KESCO BILL PMT/CONS 4412", 8920.00, None),
    ("UPI/CR/213451234562/ANJALI STORES/PHONEPE/sale", None, 9650.00),
    ("AMC CHGS-DEBIT CARD ANNUAL FEE", 590.00, None),
    ("NEFT-SBIN0009988-OWN ACCOUNT TRANSFER-SELF", 40000.00, None),
    ("UPI/DR/313451234563/GARG REPAIRS/AC servicing maintenance", 4200.00, None),
    ("INCOME TAX PMT-CBDT-ADVANCE TAX 26AAAPL", 35000.00, None),
    ("UPI/CR/413451234564/MOHIT TRADING/sale", None, 31200.00),
    ("LIC OF INDIA/ACH/PREMIUM/POLICY 5512", 9800.00, None),
    ("UPI/DR/513451234565/JIO RECHARGE/mobile postpaid", 799.00, None),
    ("NEFT-YESB0001122-GLOBE IMPORTS-PURCHASE", 52400.00, None),
    ("DRAWINGS-PROPRIETOR-PERSONAL WITHDRAWAL", 30000.00, None),
    ("UPI/CR/613451234566/CAPITAL INTRODUCED BY PARTNER", None, 200000.00),
    ("INT CREDIT-FIXED DEPOSIT INTEREST", None, 4521.00),
    ("UPI/DR/713451234567/PETROL HPCL/fuel conveyance", 2500.00, None),
    ("NEFT-KKBK0007788-ZENITH PVT LTD-INV 9921", None, 76300.00),
    ("ATM CSH WDL/HDFC/CIVIL LINES/22APR", 15000.00, None),
    ("UPI/DR/813451234568/PRINTWELL/office stationery purchase", 3120.00, None),
    ("INTEREST PAID ON OVERDRAFT-OD A/C", 5640.00, None),
    ("RTGS-HDFC0009090-NEXUS SOLUTIONS-FINAL PMT", None, 98000.00),
]

DATES = None


def build_transactions():
    """Return the ground-truth transaction list with running balances."""
    txns = []
    balance = OPENING_BALANCE
    # Dates: spread across April 2026, monotonically non-decreasing.
    day = 1
    for i, (narration, debit, credit) in enumerate(RAW):
        # advance the day every ~1-2 rows, stay within April
        if i > 0 and i % 2 == 0:
            day = min(day + 1, 30)
        date_iso = f"2026-04-{day:02d}"
        balance = round(balance - (debit or 0) + (credit or 0), 2)
        ref = "".join([c for c in narration if c.isdigit()][:6]) or None
        txns.append({
            "date": date_iso,
            "narration": narration,
            "ref_no": ref,
            "debit": debit,
            "credit": credit,
            "balance": balance,
        })
    return txns


def _printed_date(iso):
    """ISO -> DD/MM/YYYY as printed on the statement."""
    y, m, d = iso.split("-")
    return f"{d}/{m}/{y}"


def _fmt_amt(v):
    if v is None:
        return ""
    # Indian-style grouping is hard to parse; keep it plain with thousands commas
    return f"{v:,.2f}"


def render_pdf(txns, path):
    from reportlab.lib import colors
    from reportlab.lib.pagesizes import A4
    from reportlab.lib.units import mm
    from reportlab.platypus import (SimpleDocTemplate, Table, TableStyle,
                                    Paragraph, Spacer)
    from reportlab.lib.styles import getSampleStyleSheet

    styles = getSampleStyleSheet()
    doc = SimpleDocTemplate(path, pagesize=A4,
                            leftMargin=10 * mm, rightMargin=10 * mm,
                            topMargin=12 * mm, bottomMargin=12 * mm)
    story = []
    story.append(Paragraph("STATE BANK STYLE DEMO — Account Statement", styles["Title"]))
    story.append(Paragraph("Account: XXXXXX4412    Period: 01/04/2026 - 30/04/2026",
                           styles["Normal"]))
    story.append(Paragraph(f"Opening Balance: {_fmt_amt(OPENING_BALANCE)}", styles["Normal"]))
    story.append(Spacer(1, 6 * mm))

    header = ["Date", "Narration", "Ref No", "Debit", "Credit", "Balance"]
    data = [header]
    narr_style = styles["BodyText"]
    narr_style.fontSize = 7
    narr_style.leading = 8
    for t in txns:
        data.append([
            _printed_date(t["date"]),
            Paragraph(t["narration"], narr_style),
            t["ref_no"] or "",
            _fmt_amt(t["debit"]),
            _fmt_amt(t["credit"]),
            _fmt_amt(t["balance"]),
        ])

    col_widths = [20 * mm, 78 * mm, 16 * mm, 22 * mm, 22 * mm, 24 * mm]
    table = Table(data, colWidths=col_widths, repeatRows=1)
    table.setStyle(TableStyle([
        ("BACKGROUND", (0, 0), (-1, 0), colors.HexColor("#1a3c6e")),
        ("TEXTCOLOR", (0, 0), (-1, 0), colors.white),
        ("FONTSIZE", (0, 0), (-1, -1), 7),
        ("FONTNAME", (0, 0), (-1, 0), "Helvetica-Bold"),
        ("ALIGN", (3, 0), (5, -1), "RIGHT"),
        ("GRID", (0, 0), (-1, -1), 0.4, colors.grey),
        ("ROWBACKGROUNDS", (0, 1), (-1, -1), [colors.white, colors.HexColor("#eef3fa")]),
        ("VALIGN", (0, 0), (-1, -1), "MIDDLE"),
    ]))
    story.append(table)
    doc.build(story)


def render_scan(txns, path):
    """Render a statement page to an image, then rotate + add noise to mimic a scan."""
    from PIL import Image, ImageDraw, ImageFont

    W, H = 1240, 1754  # ~A4 at 150dpi
    img = Image.new("RGB", (W, H), "white")
    draw = ImageDraw.Draw(img)
    try:
        font = ImageFont.truetype("arial.ttf", 16)
        font_b = ImageFont.truetype("arialbd.ttf", 20)
        font_s = ImageFont.truetype("arial.ttf", 13)
    except Exception:
        font = ImageFont.load_default()
        font_b = font
        font_s = font

    draw.text((40, 30), "DEMO BANK — Account Statement (scanned)", fill="black", font=font_b)
    draw.text((40, 64), "Account: XXXXXX4412   Period: 01/04/2026 - 30/04/2026",
              fill="black", font=font_s)
    draw.text((40, 84), f"Opening Balance: {_fmt_amt(OPENING_BALANCE)}", fill="black", font=font_s)

    # column x positions
    cols = [40, 140, 660, 760, 920, 1080]
    y = 120
    headers = ["Date", "Narration", "Ref", "Debit", "Credit", "Balance"]
    for x, h in zip(cols, headers):
        draw.text((x, y), h, fill="black", font=font)
    y += 26
    draw.line((40, y, W - 40, y), fill="black", width=1)
    y += 6

    for t in txns:
        draw.text((cols[0], y), _printed_date(t["date"]), fill="black", font=font_s)
        narr = t["narration"]
        if len(narr) > 70:
            narr = narr[:70]
        draw.text((cols[1], y), narr, fill="black", font=font_s)
        draw.text((cols[2], y), (t["ref_no"] or ""), fill="black", font=font_s)
        draw.text((cols[3], y), _fmt_amt(t["debit"]), fill="black", font=font_s)
        draw.text((cols[4], y), _fmt_amt(t["credit"]), fill="black", font=font_s)
        draw.text((cols[5], y), _fmt_amt(t["balance"]), fill="black", font=font_s)
        y += 30

    # Rotate 1.5 deg and add mild gaussian-ish noise to simulate a scan.
    img = img.rotate(1.5, expand=False, fillcolor="white")
    px = img.load()
    rnd = random.Random(42)
    for _ in range(45000):
        x = rnd.randint(0, W - 1)
        yy = rnd.randint(0, H - 1)
        v = rnd.randint(-25, 25)
        r, g, b = px[x, yy]
        px[x, yy] = (max(0, min(255, r + v)),
                     max(0, min(255, g + v)),
                     max(0, min(255, b + v)))
    img.save(path, "PNG")


def main():
    txns = build_transactions()

    pdf_path = os.path.join(HERE, "sample_digital.pdf")
    scan_path = os.path.join(HERE, "sample_scan.png")
    truth_digital = os.path.join(HERE, "sample_digital_truth.json")
    truth_scan = os.path.join(HERE, "sample_scan_truth.json")

    render_pdf(txns, pdf_path)
    render_scan(txns, scan_path)
    with open(truth_digital, "w", encoding="utf-8") as f:
        json.dump(txns, f, indent=2)
    with open(truth_scan, "w", encoding="utf-8") as f:
        json.dump(txns, f, indent=2)

    print(f"Wrote {len(txns)} transactions")
    print(f"  {pdf_path}")
    print(f"  {scan_path}")
    print(f"  {truth_digital}")
    print(f"  {truth_scan}")


if __name__ == "__main__":
    main()