goon/app/connectors/direct_scrapers/hqporner_browse.py

"""HQPornerBrowseScraper, latest-vids browse dla hqporner.com.

Dlaczego: hqporner był TYLKO w `ALL_DIRECT_SCRAPERS` (search po nazwisku performera),
więc nowe sceny wchodziły wyłącznie wtedy, gdy worker akurat szukał pasującego
performera. Efekt: 0 świeżych scen w 7 dni (audit 2026-07-26, user: "nie widzę nowych
scen z hqporner"), mimo że tube publikuje codziennie. Browse dokłada ścieżkę
niezależną od stanu performerów.

Listing: `https://hqporner.com/` (page 1 = newest), kolejne `?p=N`, ~46-50 scen/stronę.
Detail (`/hdporn/<id>-<slug>.html`) ma komplet metadanych:
  - tytuł        → `<title>` (Title Case; `<h1>` jest lowercase, więc gorszy),
  - długość      → meta description "Video duration is 46min 58sec",
  - performerki  → `<a href="/actress/<slug>">Display Name</a>` (display, nie slug),
  - tagi         → `/category/<slug>`,
  - miniaturka   → `fastporndelivery.hqporner.com/imgs/.../_main.jpg`.

`external_id` MUSI być identyczny jak w `HQPornerScraper` (search): `hqpornercom:<url>`.
Inaczej ta sama scena z obu ścieżek zrobiłaby się dwoma wpisami.

Miniaturki fastporndelivery wymagają Referera (bez niego 403), backend i tak owija je
w `/proxy/img/`, a phash liczymy z jawnym refererem.
"""
from __future__ import annotations

import html as _html
import logging
import re

from app.connectors.base import (
    RawFingerprint,
    RawPerformer,
    RawPlaybackSource,
    RawScene,
    RawTag,
)
from app.connectors.direct_scrapers._browse_base import (
    BaseBrowseScraper,
    compute_thumbnail_phash,
)
from app.normalize.text import slugify

log = logging.getLogger(__name__)

_BASE = "https://hqporner.com"

_SCENE_HREF_RE = re.compile(r'/hdporn/(\d+)-([^"\.]+)\.html')
_TITLE_RE = re.compile(r"<title>(.*?)</title>", re.IGNORECASE | re.DOTALL)
# meta description: "... Video duration is 46min 58sec. Tags related ..."
_DUR_RE = re.compile(r"duration is\s*(?:(\d+)\s*min)?\s*(?:(\d+)\s*sec)?", re.IGNORECASE)
_ACTRESS_RE = re.compile(
    r'<a[^>]+href=[\'"]/actress/([a-z0-9\-_]+)[\'"][^>]*>(.*?)</a>',
    re.IGNORECASE | re.DOTALL,
)
_CATEGORY_RE = re.compile(r'href=[\'"]/category/([a-z0-9\-_]+)[\'"]', re.IGNORECASE)
_THUMB_RE = re.compile(r"fastporndelivery\.hqporner\.com/imgs/[^\"'\s]+_main\.jpg", re.IGNORECASE)
_TAG_STRIP = re.compile(r"<[^>]+>")

# hqporner wrzuca do /category/ także jakość ("1080p-porn", "4k-porn"). To nie jest tag
# treści, tylko atrybut pliku, i zaśmieca chmurę tagów (user: "śmieciowe tagi").
_QUALITY_TAG_RE = re.compile(r"^(?:\d{3,4}p|4k|8k|hd|uhd|full-?hd)(?:-porn)?$", re.IGNORECASE)


def _clean(text: str) -> str:
    return _html.unescape(_TAG_STRIP.sub("", text)).strip()


class HQPornerBrowseScraper(BaseBrowseScraper):
    sitetag = "hqpornercom"

    def _listing_url(self, page: int) -> str:
        return _BASE + "/" if page <= 1 else f"{_BASE}/?p={page}"

    def _extract_scene_urls(self, listing_html: str) -> list[str]:
        seen: set[str] = set()
        out: list[str] = []
        for m in _SCENE_HREF_RE.finditer(listing_html):
            url = f"{_BASE}/hdporn/{m.group(1)}-{m.group(2)}.html"
            if url in seen:
                continue
            seen.add(url)
            out.append(url)
        return out

    def _parse_detail(self, scene_url: str, detail_html: str) -> RawScene | None:
        tm = _TITLE_RE.search(detail_html)
        title = _clean(tm.group(1)) if tm else ""
        # "<Tytuł> - HQporner.com" → utnij sufiks strony.
        title = re.sub(r"\s*-\s*HQporner\.com\s*$", "", title, flags=re.IGNORECASE).strip()
        if not title:
            return None

        duration_sec: int | None = None
        dm = _DUR_RE.search(detail_html)
        if dm and (dm.group(1) or dm.group(2)):
            duration_sec = int(dm.group(1) or 0) * 60 + int(dm.group(2) or 0)
            duration_sec = duration_sec or None

        performers: list[RawPerformer] = []
        seen_perf: set[str] = set()
        for slug, disp in _ACTRESS_RE.findall(detail_html):
            name = _clean(disp) or slug.replace("-", " ").title()
            key = slugify(name)
            if not key or key in seen_perf:
                continue
            seen_perf.add(key)
            performers.append(
                RawPerformer(external_id=f"{self.sitetag}:performer:{key}", name=name)
            )

        tags: list[RawTag] = []
        seen_tag: set[str] = set()
        for slug in _CATEGORY_RE.findall(detail_html):
            slug = slug.lower()
            if slug in seen_tag or _QUALITY_TAG_RE.match(slug):
                continue
            seen_tag.add(slug)
            tags.append(
                RawTag(
                    external_id=f"{self.sitetag}:tag:{slug}",
                    name=slug.replace("-", " ").title(),
                    slug=slug,
                )
            )

        thumbnail_url: str | None = None
        thm = _THUMB_RE.search(detail_html)
        if thm:
            thumbnail_url = "https://" + thm.group(0)

        fingerprints: list[RawFingerprint] = []
        if thumbnail_url:
            ph = compute_thumbnail_phash(thumbnail_url, referer=_BASE + "/")
            if ph:
                fingerprints.append(RawFingerprint(kind="phash", value=ph))

        return RawScene(
            # Ten sam format co HQPornerScraper.search, inaczej duplikat sceny.
            external_id=f"{self.sitetag}:{scene_url}",
            title=title,
            duration_sec=duration_sec,
            url=scene_url,
            performers=performers,
            tags=tags,
            fingerprints=fingerprints,
            playback_sources=[
                RawPlaybackSource(
                    origin=f"tube:{self.sitetag}",
                    page_url=scene_url,
                    duration_sec=duration_sec,
                    thumbnail_url=thumbnail_url,
                )
            ],
            raw={"source": "direct_scraper:hqporner_browse", "url": scene_url},
        )