"""xhamster.com — natywny server-side HLS extractor.

2026-06-08: re-test Chrome DevTools + cross-IP NAPRAWIA założenie z `_vps_blocked_fallback`.
Wcześniej `xhamstercom` szedł przez WebView fallback (założenie: Cloudflare blokuje Hetzner
IP). Re-test pokazał:
  1. VPS pobiera scene page (HTTP 200, BEZ Cloudflare challenge — blok się zdjął).
  2. Master HLS URL jest w SSR HTML plain: `video-nss.xhcdn.com/<token>,<expiry>/media=hls4/
     multi=.../...m3u8`. `<expiry>` to UNIX ts → token TIME-BOUND, nie IP-bound.
  3. Cross-IP test (VPS Hetzner): master m3u8 → 200, wariant playlist → 200, segment .m4s
     → 206 video/mp4. Cały łańcuch PORTABLE — mobile gra HLS direct z residential IP,
     zero VPS proxy bandwidth.

Dlatego resolvujemy SERVER-SIDE jak porntrex/freshporno: fetch page (curl_cffi chrome) →
wyłuskaj master m3u8 → oddaj jako type='m3u8' mobile_direct. ExoPlayer robi adaptive
multi-quality z jednego master URL.

NB `sources.standard.av1/h264` w HTML to ZASZYFROWANE hex-bloby (player deszyfruje w JS),
bezużyteczne server-side — dlatego bierzemy HLS, nie mp4.
"""
from __future__ import annotations

import logging
import re

from app.extractors._fetch import _DEFAULT_IMPERSONATE, _DEFAULT_UA, _HAS_CURL_CFFI, fetch_tube_html
from app.extractors._models import HosterDead, StreamSource

log = logging.getLogger(__name__)

_BASE = "https://xhamster.com"

# Master HLS na xhcdn (video-nss.xhcdn.com / fallback inne sub-domeny). JSON w HTML
# escape'uje slashe (`https:\/\/...`), więc unescape przed matchowaniem.
_M3U8_RE = re.compile(r"https://[a-z0-9.\-]*xhcdn\.com/[^\"'\\ ]+?\.m3u8", re.IGNORECASE)

# Markery skasowanej sceny (strona istnieje, ale bez wideo) → HosterDead.
_DEAD_MARKERS = (
    "this video has been deleted",
    "this video was deleted",
    "video is no longer available",
    "has been removed",
)


def extract(page_url: str, *, timeout: float = 60.0) -> list[StreamSource] | None:
    html = ""
    if _HAS_CURL_CFFI:
        from curl_cffi import requests as _cf_requests

        session = _cf_requests.Session(impersonate=_DEFAULT_IMPERSONATE)
        try:
            resp = session.get(
                page_url,
                headers={"User-Agent": _DEFAULT_UA, "Accept": "text/html,application/xhtml+xml"},
                timeout=timeout,
                allow_redirects=True,
            )
            html = resp.text if resp.status_code < 400 else ""
        except Exception as e:
            log.info("xhamster: page fetch failed %s: %s", page_url, e)
            html = ""
    if not html:
        # fetch_tube_html podnosi TubePageError dla 404/410 (caller → dead_at).
        html = fetch_tube_html(page_url, timeout=timeout)

    # JSON-escaped slashe → plain, żeby regex złapał master URL.
    unescaped = html.replace("\\/", "/")
    m = _M3U8_RE.search(unescaped)
    if not m:
        low = unescaped.lower()
        if any(marker in low for marker in _DEAD_MARKERS):
            raise HosterDead(f"xhamster: scene deleted {page_url}")
        log.info("xhamster: no HLS master URL on %s", page_url)
        return None

    master = m.group(0)
    return [
        StreamSource(
            link=master,
            type="m3u8",
            quality=None,  # HLS master = adaptive multi-quality (ExoPlayer wybiera)
            referer=_BASE + "/",
            # Master + warianty + segmenty są time-bound (nie IP/cookie-bound),
            # zweryfikowane cross-IP 2026-06-08 → mobile gra direct, zero VPS proxy.
            raw={"mobile_direct_ok": True},
        )
    ]