"""Reaper martwych klipów: twórcy kasują klipy, RedGifs oddaje wtedy 404/410 na pliku.

Sprawdza HEAD `hd_url` (albo `sd_url`) dla próbki żywych klipów: najpierw te najdawniej
sprawdzane (`updated_at`), bo ingest `latest` i tak odświeża świeże. 404/410/403 →
`dead_at` + `dead_reason`; błędy sieci nie oznaczają śmierci (patrz notatka o dwóch
kontrolach przy przemiale źródeł). Klip, który źródło znów poda w listingu, upsert
odżywia (zdejmuje `dead_at`).

Uruchomienie (kontener worker):
  python -m scripts.clips_reap_dead --limit 500          # dry-run
  python -m scripts.clips_reap_dead --limit 500 --yes    # zapisz
"""
from __future__ import annotations

import argparse
import logging
import time

import httpx
from sqlalchemy import select, text

from app.db import session_scope
from app.models.clip import Clip

log = logging.getLogger(__name__)
_UA = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/152.0.0.0 Safari/537.36"}
_DEAD_CODES = {403, 404, 410}


def run(*, limit: int, apply: bool) -> dict[str, int]:
    checked = dead = errors = 0
    dead_ids: list = []
    with session_scope() as session:
        rows = session.execute(
            select(Clip.id, Clip.hd_url, Clip.sd_url)
            .where(Clip.dead_at.is_(None))
            .order_by(Clip.updated_at.asc())
            .limit(limit)
        ).all()
    with httpx.Client(headers=_UA, timeout=20, follow_redirects=True) as http:
        for cid, hd, sd in rows:
            url = hd or sd
            if not url:
                continue
            checked += 1
            try:
                r = http.head(url)
            except httpx.HTTPError:
                errors += 1
                continue
            if r.status_code in _DEAD_CODES:
                dead += 1
                dead_ids.append((cid, f"http {r.status_code}"))
            time.sleep(0.2)
    if apply:
        with session_scope() as session:
            for cid, reason in dead_ids:
                session.execute(
                    text("UPDATE clips SET dead_at = now(), dead_reason = :r WHERE id = :id AND dead_at IS NULL"),
                    {"r": reason, "id": cid},
                )
            # Sprawdzone i zywe: odswiez updated_at, zeby kolejny run bral inne.
            alive = [cid for cid, _hd, _sd in rows if cid not in {d[0] for d in dead_ids}]
            if alive:
                session.execute(
                    text("UPDATE clips SET updated_at = now() WHERE id = ANY(:ids)"), {"ids": alive}
                )
    out = {"checked": checked, "dead": dead, "errors": errors, "applied": int(apply)}
    print(out)
    return out


def main() -> None:
    logging.basicConfig(level=logging.INFO, format="%(message)s")
    ap = argparse.ArgumentParser()
    ap.add_argument("--limit", type=int, default=500)
    ap.add_argument("--yes", action="store_true", help="zapisz dead_at (bez tego dry-run)")
    a = ap.parse_args()
    run(limit=a.limit, apply=a.yes)


if __name__ == "__main__":
    main()