import argparse
import hashlib
import json
from pathlib import Path
import fitz

parser = argparse.ArgumentParser()
parser.add_argument("pdf", type=Path)
args = parser.parse_args()
root = Path(__file__).resolve().parents[1]
source = args.pdf.resolve()
if not source.is_relative_to(root):
    raise SystemExit("PDF musi znajdować się w Vilmal.")
sha = hashlib.sha256(source.read_bytes()).hexdigest()
out = root / "data" / "media" / "derived" / "pdf-reviews" / sha
out.mkdir(parents=True, exist_ok=True)
with fitz.open(source) as document:
    pages = []
    for index, page in enumerate(document):
        image = out / f"page-{index + 1}.png"
        if not image.exists():
            page.get_pixmap(matrix=fitz.Matrix(1.5, 1.5), alpha=False).save(image)
        pages.append({"page": index + 1, "image": str(image), "text": page.get_text()})
    print(json.dumps({"sha256": sha, "pages": pages}, ensure_ascii=True))