# -*- coding: utf-8 -*-
"""
Ekstrak data kurikulum (CPL, CPMK, MK, Sub-CPMK, matriks) yang ter-embed di
dashboard pemetaan, lalu simpan sebagai seed/kurikulum.json.

Jalankan ulang bila data dashboard diperbarui:
    python tools/extract_seed.py [path_index_html]
"""
import json
import pathlib
import sys

BASE = pathlib.Path(__file__).resolve().parent.parent
DEFAULT_SRC = BASE.parent / "dashboard-deploy" / "index.html"
OUT = BASE / "seed" / "kurikulum.json"


def main() -> None:
    src = pathlib.Path(sys.argv[1]) if len(sys.argv) > 1 else DEFAULT_SRC
    html = src.read_text(encoding="utf-8")
    i = html.find('{"cpl"')
    if i < 0:
        raise SystemExit(f"Blok JSON kurikulum tidak ditemukan di {src}")
    data, _ = json.JSONDecoder().raw_decode(html[i:])

    OUT.parent.mkdir(parents=True, exist_ok=True)
    OUT.write_text(json.dumps(data, ensure_ascii=False, indent=1), encoding="utf-8")

    n_sub = sum(
        len(c["subs"]) for mk in data["sub"].values() for c in mk["cpmks"].values()
    )
    print(
        f"[OK] {OUT.name}: {len(data['cpl'])} CPL, {len(data['cpmk'])} CPMK, "
        f"{len(data['mk'])} MK, {n_sub} Sub-CPMK (sumber: {src})"
    )


if __name__ == "__main__":
    main()