feat(scrapers): extract tags + description from porndish scene pages
porndish-only scenes had no tags and no description — the scraper only derived a
title from the URL slug. The scene page (g1/bimber WP theme) carries both: a
<p class="entry-tags"> list of /video2/<slug>/ links (the "#" tags the user sees,
categories + co-performers) and a prose description <p> in .entry-content.
Override _fetch_scene_metadata in PornDishScraper to pull both from one page
fetch. Extend the base hook to accept an optional 4th return element
(description) and thread it into RawScene.description — backward compatible with
the existing 3-tuple (pornhat). Strips leading embed-button labels
("Video Player N", "Server N") from the prose. Verified on live scenes: clean
tag lists + real descriptions.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
jtrzupek
parent
77323d23e6
commit
210aec0536
2 changed files with 104 additions and 1 deletions
|
|
@ -198,13 +198,18 @@ class BaseSearchScraper(BaseDirectTubeScraper):
|
|||
studio: RawStudio | None = None
|
||||
extra_performers: list[RawPerformer] = []
|
||||
tags: list[RawTag] = []
|
||||
description: str | None = None
|
||||
try:
|
||||
meta = self._fetch_scene_metadata(scene_url)
|
||||
except Exception as e:
|
||||
log.debug("%s metadata fetch failed for %s: %s", self.sitetag, scene_url, e)
|
||||
meta = None
|
||||
if meta is not None:
|
||||
studio, extra_performers, tags = meta
|
||||
# Back-compat: subclass może zwrócić 3-tuple (studio, performers, tags)
|
||||
# LUB 4-tuple z dodatkowym `description` (porndish). Unpack defensywnie.
|
||||
studio, extra_performers, tags = meta[0], meta[1], meta[2]
|
||||
if len(meta) > 3:
|
||||
description = meta[3]
|
||||
|
||||
# Performer z query zawsze obecny (driver scraping). Extra performers
|
||||
# z detail page dorzucamy — dedupe po slug/name w resolverze.
|
||||
|
|
@ -213,6 +218,7 @@ class BaseSearchScraper(BaseDirectTubeScraper):
|
|||
yield RawScene(
|
||||
external_id=f"{self.sitetag}:{scene_url}",
|
||||
title=title,
|
||||
description=description,
|
||||
url=scene_url,
|
||||
playback_sources=[
|
||||
RawPlaybackSource(
|
||||
|
|
|
|||
|
|
@ -2,12 +2,52 @@
|
|||
|
||||
Search: `https://porndish.com/page/<n>/?s=<q>`.
|
||||
Scene URL: `https://porndish.com/<slug>/`.
|
||||
|
||||
Scene detail page (g1/bimber WordPress theme) zawiera:
|
||||
- `<p class="entry-tags"><a class="entry-tag entry-tag-N" href=".../video2/<slug>/">Name</a>…`
|
||||
— lista tagów (kategorie + performerzy wymieszani, tak jak porndish je pokazuje
|
||||
jako „#" hashtagi). Bierzemy wszystkie jako RawTag (resolver dedupuje; performer
|
||||
z query i tak dochodzi osobno).
|
||||
- prozę opisu w `<p>` wewnątrz `.entry-content` (przed `entry-tags`, po embed-JS).
|
||||
Bez `_fetch_scene_metadata` overrides scena z samego porndish miała 0 tagów i brak
|
||||
description (bug-report od Jana 2026-06-06: „nie ma tagów (# na stronie) ani description").
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
import html as html_mod
|
||||
import logging
|
||||
import re
|
||||
|
||||
from app.connectors.base import RawPerformer, RawStudio, RawTag
|
||||
from app.connectors.direct_scrapers._search_base import BaseSearchScraper
|
||||
from app.extractors import browser_get
|
||||
|
||||
log = logging.getLogger(__name__)
|
||||
|
||||
_ENTRY_TAG_RE = re.compile(
|
||||
r'<a[^>]+href="[^"]*/video2/(?P<slug>[^"/]+)/"[^>]*class="[^"]*entry-tag[^"]*"[^>]*>'
|
||||
r'(?P<name>[^<]+)</a>',
|
||||
re.IGNORECASE,
|
||||
)
|
||||
_ENTRY_CONTENT_RE = re.compile(
|
||||
r'<div[^>]*class="[^"]*entry-content[^"]*"[^>]*>(?P<body>.*?)</article>',
|
||||
re.IGNORECASE | re.DOTALL,
|
||||
)
|
||||
_SCRIPT_STYLE_RE = re.compile(r"<script\b.*?</script>|<style\b.*?</style>", re.IGNORECASE | re.DOTALL)
|
||||
_P_RE = re.compile(r"<p\b[^>]*>(?P<inner>.*?)</p>", re.IGNORECASE | re.DOTALL)
|
||||
_TAG_STRIP_RE = re.compile(r"<[^>]+>")
|
||||
_WS_RE = re.compile(r"\s+")
|
||||
_SLUG_RE = re.compile(r"[^a-z0-9]+")
|
||||
|
||||
|
||||
def _slugify(name: str) -> str:
|
||||
return _SLUG_RE.sub("-", name.lower()).strip("-") or "tag"
|
||||
|
||||
|
||||
def _clean_text(fragment: str) -> str:
|
||||
txt = _TAG_STRIP_RE.sub(" ", fragment)
|
||||
txt = html_mod.unescape(txt)
|
||||
return _WS_RE.sub(" ", txt).strip()
|
||||
|
||||
|
||||
class PornDishScraper(BaseSearchScraper):
|
||||
|
|
@ -17,3 +57,60 @@ class PornDishScraper(BaseSearchScraper):
|
|||
r'href="(?P<url>https://porndish\.com/(?P<slug>[a-z0-9][a-z0-9\-]+))/"',
|
||||
re.IGNORECASE,
|
||||
)
|
||||
|
||||
def _fetch_scene_metadata(
|
||||
self, scene_url: str
|
||||
) -> tuple[RawStudio | None, list[RawPerformer], list[RawTag], str | None] | None:
|
||||
"""Fetch scene page → (studio=None, performers=[], tags, description).
|
||||
|
||||
4-elementowy zwrot (base unpacka opcjonalny `description`). porndish nie
|
||||
wyróżnia studia, a performer z query dochodzi w base — tu tylko tagi + opis.
|
||||
"""
|
||||
try:
|
||||
r = browser_get(scene_url, timeout=self._timeout)
|
||||
except Exception as e:
|
||||
log.debug("porndish meta fetch failed for %s: %s", scene_url, e)
|
||||
return None
|
||||
if r.status_code != 200 or not r.text:
|
||||
return None
|
||||
html = r.text
|
||||
|
||||
# Tagi: entry-tag anchors (slug z /video2/<slug>/ + display name).
|
||||
tags: list[RawTag] = []
|
||||
seen: set[str] = set()
|
||||
for m in _ENTRY_TAG_RE.finditer(html):
|
||||
name = html_mod.unescape(m.group("name")).strip()
|
||||
slug = (m.group("slug") or "").strip().lower() or _slugify(name)
|
||||
if not name or len(name) > 40 or slug in seen:
|
||||
continue
|
||||
seen.add(slug)
|
||||
tags.append(RawTag(external_id=f"porndishcom:tag:{slug}", name=name, slug=slug))
|
||||
|
||||
# Description: najdłuższy prozowy <p> w .entry-content (bez entry-tags / embed-JS).
|
||||
description: str | None = None
|
||||
mc = _ENTRY_CONTENT_RE.search(html)
|
||||
body = mc.group("body") if mc else html
|
||||
body = _SCRIPT_STYLE_RE.sub(" ", body)
|
||||
best = ""
|
||||
for pm in _P_RE.finditer(body):
|
||||
inner = pm.group("inner")
|
||||
if "entry-tag" in inner:
|
||||
continue
|
||||
txt = _clean_text(inner)
|
||||
# Pomijamy resztki JS / boilerplate „Watch … porn video" / przyciski serwerów.
|
||||
if not txt or "getElementById" in txt or "addEventListener" in txt:
|
||||
continue
|
||||
low = txt.lower()
|
||||
if low.startswith("watch ") and low.endswith("porn video"):
|
||||
continue
|
||||
if len(txt) > len(best):
|
||||
best = txt
|
||||
# Strip wiodące etykiety przycisków embedu („Video Player 1 Video Player 2 …",
|
||||
# czasem „Server N") które wpadają na początek prozy.
|
||||
best = re.sub(r"^(?:Video Player \d+\s*|Server \d+\s*|Download\s*)+", "", best, flags=re.IGNORECASE).strip()
|
||||
if len(best) >= 40:
|
||||
description = best
|
||||
|
||||
if not tags and description is None:
|
||||
return None
|
||||
return (None, [], tags, description)
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue