src/doc_verify/samples.py

"""Fictional invoices with separately specified expected discrepancy labels."""

import copy
import json
from pathlib import Path

from reportlab.lib import colors
from reportlab.lib.pagesizes import A4
from reportlab.lib.styles import getSampleStyleSheet
from reportlab.platypus import PageBreak, Paragraph, SimpleDocTemplate, Spacer, Table, TableStyle

CASES = [
    ("Matching documents", []),
    ("Price discrepancy", ["price_mismatch"]),
    ("Quantity discrepancy", ["quantity_mismatch"]),
    ("Incorrect total", ["total_mismatch"]),
    ("Missing invoice number", ["missing_field"]),
    ("Different currency", ["currency_mismatch"]),
    (
        "Ambiguous purchase lines",
        ["ambiguous_match", "unmatched_purchase_line", "subtotal_mismatch"],
    ),
    ("Incorrect line amount", ["line_total_mismatch", "subtotal_mismatch"]),
    ("Missing date", ["missing_field"]),
    ("Unmatched product", ["unmatched_line", "unmatched_purchase_line"]),
]


def sample_data(index):
    case = index % 10
    inv = {
        "vendor": "Northwind Demo Supplies",
        "number": f"INV-{index + 1:03}",
        "date": "2026-09-01",
        "currency": "INR",
        "subtotal": "1200.00",
        "tax": "120.00",
        "total": "1320.00",
        "lines": [
            {
                "sku": "MON-01",
                "description": "Desktop monitor stand",
                "quantity": "10",
                "unit_price": "120.00",
                "amount": "1200.00",
            }
        ],
        "sources": [],
        "notes": [],
    }
    po = copy.deepcopy(inv)
    po["number"] = f"PO-{index + 1:03}"
    if case == 1:
        inv["lines"][0].update(unit_price="125.00", amount="1250.00")
        inv.update(subtotal="1250.00", total="1370.00")
    elif case == 2:
        inv["lines"][0].update(quantity="12", amount="1440.00")
        inv.update(subtotal="1440.00", total="1560.00")
    elif case == 3:
        inv["total"] = "1400.00"
    elif case == 4:
        inv["number"] = None
    elif case == 5:
        inv["currency"] = "USD"
    elif case == 6:
        po["lines"].append(copy.deepcopy(po["lines"][0]))
    elif case == 7:
        inv["lines"][0]["amount"] = "1190.00"
    elif case == 8:
        inv["date"] = None
    elif case == 9:
        inv["lines"][0].update(sku="CHAIR-01", description="Office chair")
    return inv, po


def pdf(path, doc, kind, multipage=False):
    styles = getSampleStyleSheet()
    story = [
        Paragraph("NORTHWIND DEMO SUPPLIES", styles["Title"]),
        Spacer(1, 12),
        Paragraph(kind, styles["Heading1"]),
        Paragraph("Fictional portfolio sample. Not a payment request.", styles["Normal"]),
        Spacer(1, 20),
    ]
    for field in ("vendor", "number", "date", "currency"):
        if doc[field] is not None:
            story.append(Paragraph(f"{field.title()}: {doc[field]}", styles["Normal"]))
    if multipage:
        story += [
            Spacer(1, 20),
            Paragraph("Line items and totals continue on page 2.", styles["Normal"]),
            PageBreak(),
        ]
    story.append(Spacer(1, 20))
    rows = [["SKU", "Description", "Quantity", "Unit price", "Amount"]]
    rows += [
        [line[key] for key in ("sku", "description", "quantity", "unit_price", "amount")]
        for line in doc["lines"]
    ]
    table = Table(rows, colWidths=[65, 185, 60, 80, 80])
    table.setStyle(
        TableStyle(
            [
                ("BACKGROUND", (0, 0), (-1, 0), colors.HexColor("#234e70")),
                ("TEXTCOLOR", (0, 0), (-1, 0), colors.white),
                ("FONTSIZE", (0, 0), (-1, -1), 9),
                ("BOTTOMPADDING", (0, 0), (-1, -1), 12),
                ("TOPPADDING", (0, 0), (-1, -1), 12),
                ("LINEBELOW", (0, 0), (-1, -1), 0.5, colors.HexColor("#c8d5df")),
            ]
        )
    )
    story += [table, Spacer(1, 25)]
    for field in ("subtotal", "tax", "total"):
        story.append(Paragraph(f"{field.title()}: {doc[field]}", styles["Heading3"]))
    SimpleDocTemplate(str(path), pagesize=A4, leftMargin=50, rightMargin=50).build(story)


def generate(root=Path("data/samples")):
    root.mkdir(parents=True, exist_ok=True)
    labels = []
    for index in range(20):
        inv, po = sample_data(index)
        inv_path, po_path = (
            root / f"invoice-{index + 1:02}.pdf",
            root / f"purchase-{index + 1:02}.pdf",
        )
        pdf(inv_path, inv, "Invoice", multipage=index == 18)
        pdf(po_path, po, "Purchase order", multipage=index == 18)
        if index in {10, 11, 12, 19}:
            from PIL import ImageFilter

            from .extraction import render

            image = render(inv_path.read_bytes())
            if index == 19:
                image = image.resize((image.width // 2, image.height // 2)).filter(
                    ImageFilter.GaussianBlur(0.6)
                )
            inv_path = inv_path.with_suffix(".png")
            image.save(inv_path)
        labels.append(
            {
                "id": index + 1,
                "name": CASES[index % 10][0] + (" (scan)" if inv_path.suffix == ".png" else ""),
                "invoice_file": inv_path.name,
                "purchase_file": po_path.name,
                "invoice": inv,
                "purchase": po,
                "expected_codes": CASES[index % 10][1],
                "split": "tuning" if index < 10 else "heldout",
            }
        )
    (root / "labels.json").write_text(json.dumps(labels, indent=2), encoding="utf-8")
    return labels