goon/app/connectors/direct_scrapers/superporn.py

"""superporn.com — latest-vids browse scraper (przez Bright Data residential proxy).

Dołączony 2026-06-10 (user request). superporn twardo blokuje VPS IP Cloudflarem
(403 na KAŻDEJ impersonacji TLS — chrome/safari/firefox), więc ingest HTML idzie
przez Bright Data residential proxy (`settings.brightdata_proxy_url`). Gdy proxy
nieskonfigurowane → scraper no-op (pusty iterator, log warning).

Proxy używamy TYLKO do scrape HTML. Playback NIE idzie przez proxy: `<source>` mp4
(cdnst*.superporn.com) ma token IP-bound do fetchera (403 cross-IP), więc resolve
musi nastąpić po stronie telefonu — extractor `superporncom` → `_vps_blocked_fallback`
(mobile WebView ładuje stronę z residential IP telefonu, INJECTED_JS bierze video.src).
Thumbnaile (img*.superporn.com) schodzą z VPS bez proxy (image proxy działa).

Sygnały (SSR HTML):
  - JSON-LD VideoObject: name, description, uploadDate, thumbnailUrl, duration
    ISO 8601 (`P0DT0H38M48S`). `author` = uploader (NIE performer — pomijamy).
  - Performerzy: chip `<a class="chip-link" href=".../pornstar/<slug>"><span>Name</span>`
  - Kategorie: chip `<a class="chip-link" href="/<slug>"><span>Name</span>` (bez
    `/pornstar/`) — w bloku `#collapse-categories`.
  - Duration backup: `<video ... data-video-duration="2328">` (sekundy).

Listing: `?page=N` (newest-first; sitemap lastmod jest z 2024, bezużyteczny).
"""
from __future__ import annotations

import html
import logging
import re

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

log = logging.getLogger(__name__)

_BASE = "https://www.superporn.com"

_SCENE_URL_RE = re.compile(r'href="(https://www\.superporn\.com/video/[a-z0-9\-]+)"', re.IGNORECASE)
# Chip pornstar: href .../pornstar/<slug> (abs lub rel), nazwa w <span>.
_PERF_CHIP_RE = re.compile(
    r'<a[^>]+class="chip-link"[^>]+href="[^"]*/pornstar/[^"]*"[^>]*>.*?<span>([^<]+)</span>',
    re.IGNORECASE | re.DOTALL,
)
# Chip kategorii: href="/<slug>" (NIE /pornstar/, NIE /video/, NIE absolutny http).
_CAT_CHIP_RE = re.compile(
    r'<a[^>]+class="chip-link"[^>]+href="/(?!pornstar/|video/)([a-z0-9\-]+)"[^>]*>.*?<span>([^<]+)</span>',
    re.IGNORECASE | re.DOTALL,
)
_DATA_DUR_RE = re.compile(r'data-video-duration="(\d+)"')
# ISO 8601 z dniami: `P0DT0H38M48S`.
_ISO_DUR_FULL_RE = re.compile(
    r"P(?:(\d+)D)?T?(?:(\d+)H)?(?:(\d+)M)?(?:(\d+)S)?", re.IGNORECASE
)


def _clean(text: str) -> str:
    """superporn dwukrotnie HTML-koduje encje w JSON-LD (`&amp;#39;` zamiast `'`).
    Unescape iteracyjnie (max 2) — bez over-decode dla zwykłych nazw."""
    for _ in range(2):
        new = html.unescape(text)
        if new == text:
            break
        text = new
    return text.strip()


def _parse_iso_duration_days(value: str | None) -> int | None:
    """`P0DT0H38M48S` → 2328. None gdy zero/parse fail."""
    if not value:
        return None
    m = _ISO_DUR_FULL_RE.match(value.strip())
    if not m:
        return None
    d, h, mn, s = (int(g or 0) for g in m.groups())
    total = d * 86400 + h * 3600 + mn * 60 + s
    return total or None


class SuperpornScraper(BaseBrowseScraper):
    sitetag = "superporncom"

    def __init__(self) -> None:
        super().__init__()
        # Bright Data residential — bez niego superporn jest nieosiągalny z VPS.
        self._proxy = get_settings().brightdata_proxy_url
        if not self._proxy:
            log.warning("superporn: BRIGHTDATA_PROXY_URL unset — scraper disabled")

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

    def crawl_page(self, page: int):
        # Bez proxy nie ma sensu uderzać (gwarantowany CF 403) — sygnalizuj
        # "exhausted" (pusta lista), żeby deep-crawl nie retry'ował w kółko.
        if not self._proxy:
            return []
        return super().crawl_page(page)

    def _extract_scene_urls(self, listing_html: str) -> list[str]:
        seen: set[str] = set()
        out: list[str] = []
        for m in _SCENE_URL_RE.finditer(listing_html):
            url = m.group(1)
            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:
        video = _extract_video_object(detail_html) or {}

        title = _clean(video.get("name") or "")
        if not title:
            return None

        slug_m = re.search(r"/video/([a-z0-9\-]+)", scene_url, re.IGNORECASE)
        scene_slug = slug_m.group(1) if slug_m else None

        duration_sec = _parse_iso_duration_days(video.get("duration"))
        if duration_sec is None:
            dm = _DATA_DUR_RE.search(detail_html)
            if dm:
                duration_sec = int(dm.group(1)) or None

        release_date = _parse_iso_date(video.get("uploadDate"))
        description = _clean(video.get("description") or "") or None
        thumbnail_url = (video.get("thumbnailUrl") or "").strip() or None

        performers: list[RawPerformer] = []
        seen_perf: set[str] = set()
        for m in _PERF_CHIP_RE.finditer(detail_html):
            name = _clean(m.group(1))
            slug = slugify(name)
            if not slug or slug in seen_perf:
                continue
            seen_perf.add(slug)
            performers.append(
                RawPerformer(external_id=f"{self.sitetag}:performer:{slug}", name=name)
            )

        tags: list[RawTag] = []
        seen_tag: set[str] = set()
        for m in _CAT_CHIP_RE.finditer(detail_html):
            slug, name = m.group(1).strip(), _clean(m.group(2))
            if not name or slug in seen_tag:
                continue
            seen_tag.add(slug)
            tags.append(RawTag(external_id=f"{self.sitetag}:tag:{slug}", name=name, slug=slug))

        # Phash z thumbnaila — proxy NIE potrzebny (img*.superporn.com działa z VPS).
        fingerprints: list[RawFingerprint] = []
        if thumbnail_url:
            ph = compute_thumbnail_phash(thumbnail_url, referer=_BASE + "/")
            if ph:
                fingerprints.append(RawFingerprint(kind="phash", value=ph))

        playback_sources = [
            RawPlaybackSource(
                origin=f"tube:{self.sitetag}",
                page_url=scene_url,
                duration_sec=duration_sec,
                thumbnail_url=thumbnail_url,
            )
        ]

        return RawScene(
            external_id=f"{self.sitetag}:{scene_slug or scene_url}",
            title=title,
            description=description,
            release_date=release_date,
            duration_sec=duration_sec,
            url=scene_url,
            performers=performers,
            tags=tags,
            fingerprints=fingerprints,
            playback_sources=playback_sources,
        )