goon/app/extractors/tubes/hqporner.py

"""hqporner.com — direct stream extractor.

Page → iframe (mydaddy.cc lub hqwo.cc — hosting się zmienia w czasie) → wyciągnij
mp4 URL-e z `<source>` tagów lub innych miejsc w HTML/JS playera.

Dwie generacje hostera (oba aktywne dla różnych scen):

1. **Stara: mydaddy.cc/video/<hash>/** — FluidPlayer wrapper z `<source>` tagami
   bezpośrednio w HTML iframe:
   `<source src="//s12.bigcdn.cc/.../360.mp4" title="360p">` + 720p + 1080p.

2. **Nowa: hqwo.cc/player/<hash>?img=<base64>** — `<source>` tagi są wewnątrz
   JavaScript string literal (`$("#jw").html("<video>...<source src=\"...\">")`).
   Quotes są escaped (`\"`), więc plain regex na `<source[^>]+src="..."`
   nie matchuje. Trzeba odescape'ować HTML przed regex match.

   URL pattern: `https://hqwo.cc/pubs/<pub_id>/<quality>.mp4` gdzie pub_id jest
   inny niż player_hash w iframe URL — generowany serwerem per request.

Fallback gdy oba zawiodą: hoster type → mobile otworzy w WebView (FluidPlayer
JS wyciągnie URL po user click).
"""
from __future__ import annotations

import logging
import re

from app.extractors._fetch import _DEFAULT_UA, browser_get, fetch_tube_html
from app.extractors._models import StreamSource
from app.extractors.hoster import extract_stream_from_hoster

log = logging.getLogger(__name__)


_IFRAME_RE = re.compile(
    r'<div[^>]+id=["\']?playerWrapper["\']?[^>]*>.*?<iframe[^>]+src=["\']([^"\']+)',
    re.IGNORECASE | re.DOTALL,
)

# Match `<source src="...mp4" title="...">` z opcjonalnym title. Po unescape
# (`\"` → `"`) ten regex łapie zarówno raw HTML (mydaddy.cc) jak i JS-embedded
# HTML (hqwo.cc).
_SOURCE_RE = re.compile(
    r'<source[^>]+src=["\']((?://|https?://)[^"\']+\.mp4[^"\']*)["\'](?:[^>]+title=["\']([^"\']+))?',
    re.IGNORECASE,
)


def extract(page_url: str, *, timeout: float = 60.0) -> list[StreamSource] | None:
    page_html = fetch_tube_html(page_url, timeout=timeout)
    m = _IFRAME_RE.search(page_html)
    if not m:
        log.warning("hqporner: no iframe in %s", page_url)
        return None
    iframe_src = m.group(1).strip()
    if iframe_src.startswith("//"):
        iframe_src = "https:" + iframe_src
    elif iframe_src.startswith("/"):
        iframe_src = f"https://hqporner.com{iframe_src}"

    headers = {
        "User-Agent": _DEFAULT_UA,
        "Accept": "text/html,application/xhtml+xml",
        "Accept-Language": "en-US,en;q=0.9",
        "Referer": "https://hqporner.com/",
    }
    try:
        r = browser_get(iframe_src, headers=headers, timeout=timeout, follow_redirects=True)
        r.raise_for_status()
    except Exception as e:
        log.warning("hqporner iframe fetch %s failed: %s", iframe_src, e)
        return None

    # Hqwo.cc embeds `<source>` tags inside `$.html("<video>...<source src=\"...\">")`
    # JS string literals — quotes are escaped. Plain HTML in mydaddy.cc has raw quotes.
    # Unescape commonly-escaped sequences so the same regex handles both shapes.
    iframe_html = (
        r.text.replace('\\"', '"').replace("\\'", "'").replace("\\\\", "\\")
    )

    # CDN-y (bigcdn.cc, hqwo.cc) bindują URL do Referera embed iframe'a (host hqwo.cc /
    # mydaddy.cc), nie hqporner.com. Trzymamy referer = host iframe'a dla proxy.
    from urllib.parse import urlparse as _urlparse
    iframe_host = _urlparse(iframe_src).hostname or ""
    iframe_referer = f"https://{iframe_host}/" if iframe_host else iframe_src

    # De-dup by URL: hqwo.cc emits `<source>` tags twice (adblock + non-adblock branches).
    seen_urls: set[str] = set()
    sources: list[StreamSource] = []
    for sm in _SOURCE_RE.finditer(iframe_html):
        url = sm.group(1).strip()
        if url.startswith("//"):
            url = "https:" + url
        if url in seen_urls:
            continue
        seen_urls.add(url)
        title = (sm.group(2) or "").strip()
        # `force_proxy=True` (2026-05-20): CDN-y bigcdn.cc/flyflv IP-bound + flyflv ma
        # `ip=46.62.219.154` w URL path. Mobile direct = 404/403 → fallback proxy
        # generuje flicker. Force_proxy wymusza mobile użycie proxied od razu.
        # Bug-report e8ddd8d4: "kliknięcie otwiera reklamę" gdy _vps_blocked_fallback
        # (hqporner page ads). Force_proxy + native mp4 = quality picker + natywny.
        sources.append(StreamSource(
            link=url, quality=title or None, type="mp4", referer=iframe_referer,
            raw={"force_proxy": True},
        ))

    if sources:
        return sources

    # Fallback 1: niektóre mydaddy.cc iframes używają packed JS (JWPlayer).
    stream_url = extract_stream_from_hoster(
        iframe_src, referer="https://hqporner.com/", timeout=timeout,
    )
    if stream_url:
        type_hint = "m3u8" if ".m3u8" in stream_url.lower() else "mp4"
        return [StreamSource(link=stream_url, type=type_hint, referer=iframe_referer)]

    # Fallback 2: oddaj iframe URL jako hoster type — mobile otworzy w WebView,
    # FluidPlayer JS sam wyciągnie URL po user click / przejściu adblock check.
    log.info("hqporner: using hoster fallback for %s", iframe_src)
    return [StreamSource(link=iframe_src, type="hoster")]