"""pornbusy.com — dispatch po hosterze z `itemprop="embedURL"`.

Katalog jest rozproszony po hosterach (próbka 120 scen): ~41% loadvid,
~20% rodzina seekplayer (av.ezplayer.me / av.seeks.cloud / 4k.upn.one /
4k.player4me.vip — objęte poszerzonym `_HOST_RE` silnika), ~8% zpi.cx,
~12% upload18 (token z wbitym `i=<ip>/24` = IP-bound), reszta ogon.

Obsługujemy trzy pierwsze (~69% katalogu):
  - loadvid  → manifest przez POST, marker `manifest_producer` (patrz hosters/loadvid.py)
  - seekplayer → istniejący silnik (ten sam AES key/IV)
  - zpi.cx   → `embedURL` JEST plikiem wideo (mimo `.webm` w nazwie to mp4);
               zweryfikowane: Range 206 `video/mp4`, bez Referera, bez tokena

upload18 i ogon zwracamy jako None — resolve na VPS wymusiłby proxowanie CAŁEGO
wideo (token IP-bound), co łamie zasadę no-video-proxy.
"""
from __future__ import annotations

import logging
import re
from urllib.parse import urlparse

from app.extractors._fetch import fetch_tube_html
from app.extractors._models import StreamSource
from app.extractors.hosters import loadvid, seekplayer_engine

log = logging.getLogger(__name__)

_BASE = "https://pornbusy.com"
_EMBED_RE = re.compile(r'itemprop="embedURL"\s+content="([^"]+)"', re.IGNORECASE)
_IFRAME_RE = re.compile(r'<iframe[^>]+src="([^"]+)"', re.IGNORECASE)
_ZPI_HOST_RE = re.compile(r"^(?:[a-z0-9]+\.)?zpi\.cx$", re.IGNORECASE)


def extract(page_url: str, *, timeout: float = 60.0) -> list[StreamSource] | None:
    html_text = fetch_tube_html(page_url, timeout=timeout)
    m = _EMBED_RE.search(html_text) or _IFRAME_RE.search(html_text)
    if not m:
        log.info("pornbusy: no embedURL/iframe on %s", page_url)
        return None
    embed = m.group(1).strip()
    if embed.startswith("//"):
        embed = "https:" + embed

    if loadvid.matches(embed):
        return loadvid.extract(embed, timeout=timeout)

    if seekplayer_engine.matches(embed):
        sources = seekplayer_engine.extract(embed, timeout=timeout)
        if not sources:
            return None
        player_origin = f"https://{urlparse(embed).hostname}/"
        out: list[StreamSource] = []
        for s in sources:
            raw = dict(s.raw or {})
            raw["proxy_no_verify"] = True
            out.append(
                StreamSource(
                    link=s.link,
                    type=s.type or "m3u8",
                    quality=s.quality,
                    # Referer = origin PLAYERA, nie strony (jak przy galaxyporn:
                    # z Refererem strony CDN zwraca 403).
                    referer=s.referer or player_origin,
                    raw=raw,
                )
            )
        return out

    host = (urlparse(embed).hostname or "").lower()
    if _ZPI_HOST_RE.match(host):
        # embedURL to bezpośrednio plik (rozszerzenie `.webm` myli — to mp4).
        return [
            StreamSource(
                link=embed,
                type="mp4",
                raw={"mobile_direct_ok": True},
            )
        ]

    log.info("pornbusy: unsupported hoster %s (%s)", host, page_url)
    return None