"""Build the synthetic E2E fixture PDF at ``tests/fixtures/synthetic_giro.pdf``.

Re-runnable on demand. Output bytes are stable across runs in page
content, layout, and text — only the PDF's embedded timestamps change,
which pipeline tests don't read. The committed fixture is what CI
consumes; re-run this script locally if you change the ground truth.

Contents: one A4 portrait page with six known strings placed at fixed
positions near the top. The goal is reproducible ground truth, not a
realistic bank statement. The pipeline's fake OCR client is seeded with
those same strings (at plausible bboxes) so the E2E test can assert
exact matches.

Usage::

    uv run python scripts/create_fixture_pdf.py
"""

from __future__ import annotations

from pathlib import Path

import fitz  # PyMuPDF

OUT_PATH = (
    Path(__file__).resolve().parent.parent / "tests" / "fixtures" / "synthetic_giro.pdf"
)

LINES: list[str] = [
    "DKB",
    "IBAN: DE89370400440532013000",
    "Statement period: 01.03.2026 - 31.03.2026",
    "Opening balance: 1234.56 EUR",
    "Closing balance: 1450.22 EUR",
    "Statement date: 31.03.2026",
]


def build() -> None:
    doc = fitz.open()
    # A4 @ 72 dpi -> 595 x 842 points.
    page = doc.new_page(width=595, height=842)
    y = 72.0
    for line in LINES:
        page.insert_text(
            (72.0, y),
            line,
            fontsize=12,
            fontname="helv",
        )
        y += 24.0
    OUT_PATH.parent.mkdir(parents=True, exist_ok=True)
    # deflate=False + garbage=0 keeps the output byte-stable.
    doc.save(
        str(OUT_PATH),
        deflate=False,
        deflate_images=False,
        garbage=0,
        clean=False,
    )
    doc.close()


if __name__ == "__main__":
    build()
    print(f"wrote {OUT_PATH}")