"""One-shot: backfill `release_date` for freshporno scenes that were scraped before
the `itemprop="uploadDate"` regex was added.

Tło: bug-report 2026-05-20 ("brak Brazzers Exxtra po 15-05") wymusił dodanie
`release_date` extracta z `itemprop="uploadDate"` w freshporno connector. Stare
scenes (z przed tego patcha) mają `release_date = NULL`, przez co scene_resolver
nie liczy date-overlap signal → score < 0.92 → orphan zamiast merged z TPDB
canonical.

10468 orphan freshporno scenes (vs 4789 canonical) — 99% bez release_date.
Po backfill resolver auto-merge przy następnym bulk-dedup tick.

Idempotent: update tylko gdy aktualne `release_date IS NULL` i `uploadDate`
ekstrakcja się powiedzie.
"""
from __future__ import annotations

import logging
import re
from datetime import UTC, date, datetime

import httpx
from sqlalchemy import select

from app.db import session_scope
from app.models import Scene
from app.models.playback_source import PlaybackSource

log = logging.getLogger(__name__)

USER_AGENT = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/140.0.0.0"
_UPLOAD_DATE_RE = re.compile(
    r'itemprop="uploadDate"[^>]+content="(\d{4}-\d{2}-\d{2})',
)


def main() -> int:
    logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")

    with session_scope() as session:
        rows = session.execute(
            select(Scene.id, PlaybackSource.page_url)
            .join(PlaybackSource, PlaybackSource.scene_id == Scene.id)
            .where(PlaybackSource.origin == "tube:freshpornoorg")
            .where(Scene.release_date.is_(None))
        ).all()
        log.info("freshporno scenes without release_date: %d", len(rows))

    client = httpx.Client(
        timeout=15.0,
        follow_redirects=True,
        headers={"User-Agent": USER_AGENT},
    )

    updated = 0
    skipped = 0
    errors = 0

    for scene_id, page_url in rows:
        try:
            r = client.get(page_url)
            if r.status_code != 200:
                if r.status_code in (404, 410):
                    skipped += 1
                else:
                    errors += 1
                continue
            m = _UPLOAD_DATE_RE.search(r.text)
            if not m:
                skipped += 1
                continue
            try:
                rd = date.fromisoformat(m.group(1))
            except ValueError:
                skipped += 1
                continue

            with session_scope() as s:
                scene = s.get(Scene, scene_id)
                if scene is None or scene.release_date is not None:
                    continue
                scene.release_date = rd
            updated += 1
            if updated % 100 == 0:
                log.info(
                    "progress: updated=%d skipped=%d errors=%d (%d/%d)",
                    updated, skipped, errors,
                    updated + skipped + errors, len(rows),
                )
        except Exception as e:
            errors += 1
            if errors <= 5:
                log.warning("scene=%s url=%s failed: %s", scene_id, page_url, e)

    client.close()
    log.info(
        "DONE: candidates=%d updated=%d skipped=%d errors=%d",
        len(rows), updated, skipped, errors,
    )
    return 0


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