"""Hentaimama: ogon w tytulach i brakujace miniatury (2026-10-15).

1. Tytuly: od ~31.08.2026 og:title mial ogon "TYTUL - Watch Online Free in HD - Hentaimama"
   i tak weszlo ok. 100 scen. Scraper juz go ucina, tu poprawiamy istniejace (title i
   title_normalized).
2. Miniatury: ok. 2100 zrodel z pierwszego przebiegu (12.08) nie ma thumbnail_url, a te sceny
   nie maja innego zrodla z obrazkiem. Strony maja og:image, pobieramy je po kolei.

    python scripts/fix_hentaimama_titles_thumbs.py            # podglad
    python scripts/fix_hentaimama_titles_thumbs.py --apply [--limit N] [--sleep 0.5]

Dlugie, wiec na prodzie przez `docker compose run -d --rm worker python3 scripts/...`
(z pliku, nie przez stdin).
"""
from __future__ import annotations

import argparse
import html as html_mod
import logging
import time

import sqlalchemy as sa

from app.connectors.direct_scrapers.hentaimama import _OG_IMAGE_RE, _TITLE_TAIL_RE
from app.db import session_scope
from app.extractors._fetch import fetch_tube_html
from app.normalize.text import normalize

log = logging.getLogger(__name__)


def fix_titles(apply: bool) -> int:
    n = 0
    with session_scope() as session:
        rows = session.execute(
            sa.text("SELECT id, title FROM scenes WHERE title ~* 'hentaimama\\s*$'")
        ).all()
        for scene_id, title in rows:
            new = _TITLE_TAIL_RE.sub("", title).strip()
            if not new or new == title:
                continue
            n += 1
            if n <= 5:
                print(f"{title!r} -> {new!r}")
            if apply:
                session.execute(
                    sa.text("UPDATE scenes SET title=:t, title_normalized=:n WHERE id=:id"),
                    {"t": new[:512], "n": normalize(new)[:512], "id": scene_id},
                )
        if apply:
            session.commit()
    return n


def fix_thumbs(apply: bool, limit: int, sleep: float) -> dict[str, int]:
    c = {"checked": 0, "filled": 0, "no_image": 0, "fetch_fail": 0}
    with session_scope() as session:
        rows = session.execute(
            sa.text(
                "SELECT id, page_url FROM playback_sources WHERE origin='tube:hentaimama' "
                "AND thumbnail_url IS NULL AND dead_at IS NULL ORDER BY id LIMIT :lim"
            ),
            {"lim": limit},
        ).all()
    for pb_id, page_url in rows:
        c["checked"] += 1
        try:
            page = fetch_tube_html(page_url, timeout=30.0)
        except Exception as e:
            c["fetch_fail"] += 1
            log.info("fetch %s: %s", page_url, e)
            time.sleep(sleep)
            continue
        m = _OG_IMAGE_RE.search(page)
        if not m:
            c["no_image"] += 1
        else:
            img = html_mod.unescape(m.group(1)).strip()
            if apply:
                with session_scope() as session:
                    session.execute(
                        sa.text("UPDATE playback_sources SET thumbnail_url=:u WHERE id=:id"),
                        {"u": img[:2048], "id": pb_id},
                    )
                    session.commit()
            c["filled"] += 1
            if c["filled"] <= 3:
                print(page_url, "->", img)
        if c["checked"] % 100 == 0:
            log.info("thumbs %s", c)
        time.sleep(sleep)
    return c


def main() -> int:
    logging.basicConfig(level=logging.INFO)
    p = argparse.ArgumentParser()
    p.add_argument("--apply", action="store_true")
    p.add_argument("--limit", type=int, default=5)
    p.add_argument("--sleep", type=float, default=0.5)
    p.add_argument("--skip-titles", action="store_true")
    args = p.parse_args()
    if not args.skip_titles:
        print("titles:", fix_titles(args.apply))
    print("thumbs:", fix_thumbs(args.apply, args.limit, args.sleep))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())