""" make_hard_tests.py — adversarial PDFs to try to BREAK the unedited converter. Each function builds one statement that stresses a different real-world failure mode. We then run converter.py on each and see what survives. """ from __future__ import annotations import random from pathlib import Path from reportlab.lib import colors from reportlab.lib.pagesizes import A4 from reportlab.lib.styles import getSampleStyleSheet from reportlab.lib.units import mm from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle HERE = Path(__file__).resolve().parent OUT = HERE / "hard" OUT.mkdir(exist_ok=True) random.seed(7) DESCS = ["AMAZON MKTPLACE", "WOOLWORTHS 1234", "SALARY ACME CORP", "BPAY ASIC", "UBER TRIP", "AGL ENERGY DD", "NETFLIX SUB", "INTEREST PAID", "TFR RENT 062000", "DIVIDEND VAS ETF", "COLES 0456", "OPTUS MOBILE"] def gen(n=15, opening_c=1_000_000): rows, running = [], opening_c d = 1 for i in range(n): d += random.randint(1, 2) is_cr = random.random() < 0.45 c = random.randint(500, 400000) running += c if is_cr else -c rows.append({ "date": f"{min(d,28):02d}/03/2025", "desc": f"{random.choice(DESCS)} REF{random.randint(1000,9999)}", "debit_c": 0 if is_cr else c, "credit_c": c if is_cr else 0, "bal_c": running, }) return rows, opening_c, running def usd(c): return f"${c/100:,.2f}" def eur(c): return f"€{c/100:,.2f}".replace(",", "X").replace(".", ",").replace("X", ".") # 1.234,56 def build(path, header, data_rows, meta_lines, grid=True): styles = getSampleStyleSheet() doc = SimpleDocTemplate(str(path), pagesize=A4, topMargin=16*mm, bottomMargin=16*mm, leftMargin=14*mm, rightMargin=14*mm) story = [Paragraph("Test Bank Statement", styles["Title"]), Spacer(1, 6)] for m in meta_lines: story.append(Paragraph(m, styles["Normal"])) story.append(Spacer(1, 10)) tbl = Table([header] + data_rows, repeatRows=1) style = [("FONTSIZE", (0, 0), (-1, -1), 7), ("ALIGN", (2, 0), (-1, -1), "RIGHT")] if grid: style += [("GRID", (0, 0), (-1, -1), 0.4, colors.grey), ("BACKGROUND", (0, 0), (-1, 0), colors.HexColor("#1f3a5f")), ("TEXTCOLOR", (0, 0), (-1, 0), colors.white)] tbl.setStyle(TableStyle(style)) story.append(tbl) doc.build(story) def std_meta(op_c, cl_c, op_label="Opening Balance", cl_label="Closing Balance"): return ["Account: 111222333 - TEST ACCOUNT", "Account Holder: TEST USER PTY LTD", "Statement Period: 01/03/2025 to 31/03/2025", f"{op_label}: {usd(op_c)}", f"{cl_label}: {usd(cl_c)}"] # ---- T1: borderless (no gridlines) ---- def t1_no_gridlines(): rows, op, cl = gen() data = [[r["date"], r["desc"], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for r in rows] build(OUT/"t1_no_gridlines.pdf", ["Date","Description","Debit","Credit","Balance"], data, std_meta(op, cl), grid=False) # ---- T2: scanned / image-only (no text layer) ---- def t2_scanned(): try: from PIL import Image, ImageDraw, ImageFont from reportlab.pdfgen import canvas except Exception as e: print(" (skip t2 scanned — PIL missing:", e, ")"); return def font(sz): for p in ("/System/Library/Fonts/Supplemental/Arial.ttf", "/Library/Fonts/Arial.ttf", "/System/Library/Fonts/Helvetica.ttc"): try: return ImageFont.truetype(p, sz) except Exception: pass return ImageFont.load_default(size=sz) rows, op, cl = gen() W, H = 1654, 2339 # A4 @ ~200dpi, crisp TrueType text = realistic clean scan img = Image.new("RGB", (W, H), "white"); dr = ImageDraw.Draw(img) f, fb = font(30), font(42) y = 60 dr.text((60, y), "Test Bank Statement", fill="black", font=fb); y += 80 for line in std_meta(op, cl): dr.text((60, y), line, fill="black", font=f); y += 46 y += 24 cols = [60, 320, 780, 1040, 1300] for x, h in zip(cols, ["Date", "Description", "Debit", "Credit", "Balance"]): dr.text((x, y), h, fill="black", font=f) y += 54 for r in rows: cells = [r["date"], r["desc"][:22], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for x, cell in zip(cols, cells): dr.text((x, y), cell, fill="black", font=f) y += 44 png = OUT/"t2_scan.png"; img.save(png) c = canvas.Canvas(str(OUT/"t2_scanned.pdf"), pagesize=A4) c.drawImage(str(png), 0, 0, width=A4[0], height=A4[1]); c.save() # ---- T3: renamed headers ---- def t3_renamed_headers(): rows, op, cl = gen() data = [[r["date"], r["desc"], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for r in rows] build(OUT/"t3_renamed_headers.pdf", ["Transaction Date","Details","Withdrawals","Deposits","Running Balance"], data, std_meta(op, cl)) # ---- T4: single signed Amount column ---- def t4_signed_amount(): rows, op, cl = gen() data = [[r["date"], r["desc"], usd(r["credit_c"]) if r["credit_c"] else f"-{usd(r['debit_c'])}", f"{usd(r['bal_c'])} CR"] for r in rows] build(OUT/"t4_signed_amount.pdf", ["Date","Description","Amount","Balance"], data, std_meta(op, cl)) # ---- T5: European number format + euro symbol ---- def t5_european_format(): rows, op, cl = gen() data = [[r["date"], r["desc"], eur(r["debit_c"]) if r["debit_c"] else "", eur(r["credit_c"]) if r["credit_c"] else "", f"{eur(r['bal_c'])} CR"] for r in rows] meta = ["Account: 111222333 - TEST ACCOUNT", "Account Holder: TEST USER", "Statement Period: 01/03/2025 to 31/03/2025", f"Opening Balance: {eur(op)}", f"Closing Balance: {eur(cl)}"] build(OUT/"t5_european_format.pdf", ["Date","Description","Debit","Credit","Balance"], data, meta) # ---- T6: two accounts in one PDF ---- def t6_multi_account(): styles = getSampleStyleSheet() doc = SimpleDocTemplate(str(OUT/"t6_multi_account.pdf"), pagesize=A4, topMargin=16*mm, bottomMargin=16*mm, leftMargin=14*mm, rightMargin=14*mm) story = [Paragraph("Test Bank Statement", styles["Title"]), Spacer(1, 6)] for acc in ("Account: AAA-111", "Account: BBB-222"): rows, op, cl = gen() for m in [acc, f"Opening Balance: {usd(op)}", f"Closing Balance: {usd(cl)}"]: story.append(Paragraph(m, styles["Normal"])) story.append(Spacer(1, 6)) data = [["Date","Description","Debit","Credit","Balance"]] + [ [r["date"], r["desc"], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for r in rows] t = Table(data, repeatRows=1) t.setStyle(TableStyle([("GRID",(0,0),(-1,-1),0.4,colors.grey),("FONTSIZE",(0,0),(-1,-1),7)])) story.append(t); story.append(Spacer(1, 14)) doc.build(story) # ---- T7: relabeled balance lines (no "Opening Balance:") ---- def t7_diff_balance_labels(): rows, op, cl = gen() data = [[r["date"], r["desc"], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for r in rows] build(OUT/"t7_diff_balance_labels.pdf", ["Date","Description","Debit","Credit","Balance"], data, std_meta(op, cl, "Balance Brought Forward", "Balance Carried Forward")) # ---- T8: text-month dates (should still survive) ---- def t8_text_dates(): rows, op, cl = gen() months = {3: "Mar"} data = [[f"{i+1:02d} Mar 2025", r["desc"], usd(r["debit_c"]) if r["debit_c"] else "", usd(r["credit_c"]) if r["credit_c"] else "", f"{usd(r['bal_c'])} CR"] for i, r in enumerate(rows)] build(OUT/"t8_text_dates.pdf", ["Date","Description","Debit","Credit","Balance"], data, std_meta(op, cl)) if __name__ == "__main__": for fn in (t1_no_gridlines, t2_scanned, t3_renamed_headers, t4_signed_amount, t5_european_format, t6_multi_account, t7_diff_balance_labels, t8_text_dates): fn() print("built:", fn.__name__)