"""Per-origin extractor check: dla 1 sample sceny z każdego tube origin,
wywołaj try_extract i sklasyfikuj wynik (direct mp4/m3u8 vs WebView hoster vs fail).
Uruchamiać na VPS: docker compose exec -T api python scripts/check_all_hosters.py
"""
from app.db import SessionLocal
from sqlalchemy import text
from app.extractors import try_extract


def main():
    with SessionLocal() as s:
        rows = s.execute(text("""
            SELECT DISTINCT ON (ps.origin)
                ps.origin, ps.page_url, sc.title
            FROM playback_sources ps
            JOIN scenes sc ON sc.id = ps.scene_id
            WHERE ps.dead_at IS NULL AND ps.origin LIKE 'tube:%'
              AND ps.page_url IS NOT NULL
            ORDER BY ps.origin, sc.created_at DESC
        """)).all()

    print(f"{'origin':<26} {'result':<48} verdict")
    print("-" * 95)
    for r in rows:
        sitetag = r.origin.replace("tube:", "")
        try:
            sources = try_extract(sitetag, r.page_url)
        except Exception as e:
            print(f"{r.origin:<26} EXC: {str(e)[:42]:<48} ERROR")
            continue
        if not sources:
            print(f"{r.origin:<26} {'None (no sources)':<48} FAIL")
            continue
        # Klasyfikacja po type pierwszego źródła
        types = [getattr(x, "type", "?") for x in sources]
        first = sources[0]
        t = getattr(first, "type", "?")
        link = (getattr(first, "link", "") or "")[:40]
        if t == "hoster":
            verdict = "WEBVIEW (page → ad risk)"
        elif t in ("mp4", "m3u8", "hls", "mpd"):
            verdict = "DIRECT (native ExoPlayer)"
        else:
            verdict = f"OTHER({t})"
        n = len(sources)
        print(f"{r.origin:<26} {f'{t} x{n} {link}':<48} {verdict}")


if __name__ == "__main__":
    main()