From 920740b76fad2ec6d648d813bdfb0fd22c117905 Mon Sep 17 00:00:00 2001
From: jtrzupek <jtrzupek@gmail.com>
Date: Mon, 1 Jun 2026 10:36:33 +0200
Subject: [PATCH] fix(pornditt): server-side KVS resolve; extract shared _kvs
 helper
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

pornditt is the same kt_player KVS engine as yespornvip: flashvars carry
function/0/-obfuscated get_file urls + license_code, and the VPS reaches it
(HTTP 200). It was on _vps_blocked_fallback (WebView), where the scrape grabbed
the VAST preroll ad (trafostatic) instead of content (user bug "pornditt łapie
reklamę zamiast video").

Extracted the verified yespornvip logic into app/extractors/tubes/_kvs.py
(resolve_kvs: fetch page → decode function/0 get_file via kt_player algo → follow
302 in-session → portable CDN, multi-quality). yespornvip.py and new pornditt.py
are now thin wrappers. Registry: porndittcom _vps_blocked_fallback → pornditt.extract.

Verified on prod: pornditt → 720p/480p on twa.tgprn.com (portable, fresh-session
206 video/mp4); yespornvip still → 1080/720/480p on cdn5 (refactor intact).
Backend-only, no OTA — mobile plays mp4+mobile_direct_ok natively with quality
picker, zero WebView/ads.

Note: a real-browser residential load shows MEDIA_ERR on the content (the page's
own player flow / ad gating); server-side decode+follow sidesteps the player
entirely, which is why it resolves cleanly. The original bug scene (40f118e1) has
its video deleted on pornditt — verified on a live scene (156091).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 app/extractors/__init__.py         |   6 +-
 app/extractors/tubes/_kvs.py       | 163 ++++++++++++++++++++++++++++
 app/extractors/tubes/pornditt.py   |  23 ++++
 app/extractors/tubes/yespornvip.py | 168 ++---------------------------
 4 files changed, 198 insertions(+), 162 deletions(-)
 create mode 100644 app/extractors/tubes/_kvs.py
 create mode 100644 app/extractors/tubes/pornditt.py

diff --git a/app/extractors/__init__.py b/app/extractors/__init__.py
index 15ae652..6e93a17 100644
--- a/app/extractors/__init__.py
+++ b/app/extractors/__init__.py
@@ -33,6 +33,7 @@ from app.extractors.tubes import (
     latestpornvideo,
     paradisehill,
     porn00,
+    pornditt,
     pornhat,
     porntrex,
     pornxp,
@@ -82,7 +83,10 @@ _REGISTRY: dict[str, Callable[[str], list[StreamSource] | None]] = {
     # IP (potwierdzone Chrome DevTools MCP 2026-05-15). Mobile WebView + INJECTED_JS
     # (PlayerScreen.tsx:805) skanuje <video>.src + XHR — łapie URL po decode-ie player JS.
     "xhamstercom": _vps_blocked_fallback.extract,
-    "porndittcom": _vps_blocked_fallback.extract,
+    # pornditt — KVS jak yespornvip (function/0 + license). VPS dociera → resolve
+    # server-side (decode + follow 302 → portable twa.tgprn.com CDN). Wcześniej WebView
+    # fallback łapał VAST preroll (trafostatic) zamiast contentu. Patrz pornditt.py/_kvs.py.
+    "porndittcom": pornditt.extract,
     "fpoxxx": _vps_blocked_fallback.extract,
     "sxylandcom": _vps_blocked_fallback.extract,
     # Aggregator tubes — generic embed-iframe → hoster unpacker
diff --git a/app/extractors/tubes/_kvs.py b/app/extractors/tubes/_kvs.py
new file mode 100644
index 0000000..0bc7270
--- /dev/null
+++ b/app/extractors/tubes/_kvs.py
@@ -0,0 +1,163 @@
+"""Współdzielony resolver dla KVS (kt_player) tube'ów z `function/0/` + `license_code`.
+
+Wzorzec (yespornvip, pornditt, ... — wszystkie KVS na tym samym silniku):
+  flashvars: `video_url` / `video_alt_url` / `video_alt_url2` = `function/0/https://<host>/
+  get_file/<srv>/<HASH>/.../<id>.mp4/` (480/720/1080p) + `license_code: '$...'`.
+  kt_player dekoduje HASH (permutacja pierwszych 32 znaków algorytmem license_code —
+  algo zgodny z yt-dlp KVS `_kvs_get_real_url`, zweryfikowany 2026-05-31 że odtwarza
+  output kt_player).
+
+Zdekodowany `get_file` 302-redirectuje do CDN (cdn5.yesporn.vip / twa.tgprn.com / ...)
+który serwuje wideo (206). Finalny URL jest **time-bound signed, NIE IP/cookie-bound** —
+portable cross-IP. get_file token jest session-bound, więc 302 MUSI być rozwiązany w tej
+samej sesji curl_cffi co fetch strony. Oddajemy finalny portable CDN url per jakość →
+mobile gra direct natywnie, multi-quality, zero WebView/reklam/proxy.
+
+Dlaczego server-side (nie WebView): WebView ładuje ad-heavy stronę + VAST preroll, a
+scrape łapie reklamę zamiast contentu. Direct get_file→CDN omija player/reklamę.
+"""
+from __future__ import annotations
+
+import logging
+import re
+import time
+import urllib.parse as _up
+
+from app.extractors._fetch import _DEFAULT_IMPERSONATE, _DEFAULT_UA, _HAS_CURL_CFFI
+from app.extractors._models import StreamSource
+
+log = logging.getLogger(__name__)
+
+_URL_RE = re.compile(
+    r"(video(?:_alt)?_url\d*)\s*:\s*[\"'](function/0/[^\"']+)[\"']",
+    re.IGNORECASE,
+)
+_TEXT_RE = re.compile(
+    r"(video(?:_alt)?_url\d*)_text\s*:\s*[\"']([^\"']*)[\"']",
+    re.IGNORECASE,
+)
+_LICENSE_RE = re.compile(r"license_code\s*:\s*[\"'](\$[^\"']+)[\"']", re.IGNORECASE)
+
+_HASH_LENGTH = 32
+
+
+def _license_token(license_code: str) -> list[int]:
+    license_code = license_code.replace("$", "")
+    license_values = [int(c) for c in license_code]
+    modlicense = license_code.replace("0", "1")
+    center = len(modlicense) // 2
+    modlicense = str(4 * abs(int(modlicense[:center + 1]) - int(modlicense[center:])))[:center + 1]
+    return [
+        (license_values[index + offset] + current) % 10
+        for index, current in enumerate(int(c) for c in modlicense)
+        for offset in range(4)
+    ]
+
+
+def real_url(video_url: str, license_code: str) -> str:
+    """Dekoduje `function/0/...get_file/N/<HASH>/...` permutując pierwsze 32 znaki HASH."""
+    if not video_url.startswith("function/0/"):
+        return video_url
+    parsed = _up.urlparse(video_url[len("function/0/"):])
+    lt = _license_token(license_code)
+    parts = parsed.path.split("/")
+    h = parts[3][:_HASH_LENGTH]
+    idx = list(range(_HASH_LENGTH))
+    acc = 0
+    for src in reversed(range(_HASH_LENGTH)):
+        acc += lt[src]
+        dest = (src + acc) % _HASH_LENGTH
+        idx[src], idx[dest] = idx[dest], idx[src]
+    parts[3] = "".join(h[i] for i in idx) + parts[3][_HASH_LENGTH:]
+    return _up.urlunparse(parsed._replace(path="/".join(parts)))
+
+
+def _quality_rank(label: str | None) -> int:
+    if not label:
+        return -1
+    m = re.search(r"(\d{3,4})\s*p", label, re.IGNORECASE)
+    return int(m.group(1)) if m else -1
+
+
+def _resolve_get_file(session, base_url: str, get_file_url: str, timeout: float) -> str | None:
+    sep = "&" if "?" in get_file_url else "?"
+    url = f"{get_file_url}{sep}rnd={int(time.time() * 1000)}"
+    try:
+        r = session.get(
+            url, timeout=timeout, allow_redirects=True, stream=True,
+            headers={"Referer": base_url + "/", "Range": "bytes=0-1"},
+        )
+        final = str(r.url)
+        status = r.status_code
+        r.close()
+    except Exception as e:
+        log.info("kvs: get_file resolve failed (%s): %s", get_file_url[:60], e)
+        return None
+    if status >= 400 or "/get_file/" in final:
+        log.info("kvs: get_file resolve bad status=%s final=%s", status, final[:70])
+        return None
+    return final
+
+
+def resolve_kvs(page_url: str, *, base_url: str, timeout: float = 60.0) -> list[StreamSource] | None:
+    """Fetch KVS page → decode function/0 get_file (per jakość) → follow 302 → portable CDN.
+
+    Zwraca StreamSource'y posortowane malejąco po jakości, lub None gdy nic nie wyszło.
+    base_url: scheme+host hosta (np. 'https://yesporn.vip') — do Referera i logów.
+    """
+    if not _HAS_CURL_CFFI:
+        log.info("kvs: curl_cffi unavailable — cannot resolve %s", page_url)
+        return None
+    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("kvs: page fetch failed %s: %s", page_url, e)
+        return None
+    if not html:
+        log.info("kvs: empty page %s", page_url)
+        return None
+
+    licm = _LICENSE_RE.search(html)
+    if not licm:
+        log.info("kvs: no license_code on %s", page_url)
+        return None
+    license_code = licm.group(1)
+
+    quality_by_var: dict[str, str] = {}
+    for m in _TEXT_RE.finditer(html):
+        quality_by_var[m.group(1).lower()] = m.group(2).strip()
+
+    seen_dec: set[str] = set()
+    result: list[StreamSource] = []
+    for m in _URL_RE.finditer(html):
+        var_name = m.group(1).lower()
+        decoded = real_url(m.group(2), license_code)
+        if decoded in seen_dec:
+            continue
+        seen_dec.add(decoded)
+        final = _resolve_get_file(session, base_url, decoded, timeout)
+        if not final:
+            continue
+        result.append(
+            StreamSource(
+                link=final,
+                type="mp4",
+                quality=quality_by_var.get(var_name) or None,
+                referer=base_url + "/",
+                raw={"mobile_direct_ok": True},
+            )
+        )
+
+    if not result:
+        log.info("kvs: no resolvable get_file on %s", page_url)
+        return None
+
+    result.sort(key=lambda s: _quality_rank(s.quality), reverse=True)
+    return result
diff --git a/app/extractors/tubes/pornditt.py b/app/extractors/tubes/pornditt.py
new file mode 100644
index 0000000..0029827
--- /dev/null
+++ b/app/extractors/tubes/pornditt.py
@@ -0,0 +1,23 @@
+"""pornditt.com — KVS (kt_player) direct stream extractor. Patrz app/extractors/tubes/_kvs.py.
+
+User bug 2026-05-31 (scene 40f118e1): "Pornditt łapie reklamę zamiast video". pornditt
+był na _vps_blocked_fallback (WebView), gdzie scrape łapał VAST preroll (trafostatic) zamiast
+contentu. Identyczny silnik jak yespornvip: flashvars `video_url`/`video_alt_url` =
+`function/0/...get_file/...` + `license_code`; VPS dociera (HTTP 200). Resolve server-side:
+decode + follow 302 → portable CDN (twa.tgprn.com, time-bound, NIE IP/cookie-bound —
+zweryfikowane cross-IP 2026-06-01 fresh session → 206 video/mp4). Native, multi-quality,
+zero WebView/reklam.
+
+NB: runtime `window.flashvars.video_url` pokazuje już ZDEKODOWANY plain get_file, ale raw
+HTML (server-fetch) ma formę `function/0/...` + license — dekodujemy sami (_kvs.real_url).
+"""
+from __future__ import annotations
+
+from app.extractors._models import StreamSource
+from app.extractors.tubes import _kvs
+
+_BASE = "https://v.pornditt.com"
+
+
+def extract(page_url: str, *, timeout: float = 60.0) -> list[StreamSource] | None:
+    return _kvs.resolve_kvs(page_url, base_url=_BASE, timeout=timeout)
diff --git a/app/extractors/tubes/yespornvip.py b/app/extractors/tubes/yespornvip.py
index 1007d67..d0fcdf2 100644
--- a/app/extractors/tubes/yespornvip.py
+++ b/app/extractors/tubes/yespornvip.py
@@ -1,171 +1,17 @@
-"""yesporn.vip — KVS (kt_player) engine direct stream extractor.
+"""yesporn.vip — KVS (kt_player) direct stream extractor. Patrz app/extractors/tubes/_kvs.py.
 
-REWRITE 2026-05-31 (user "yespornvip otwiera reklamy / 410 / fallback do strony"):
-VPS znów dociera do yesporn.vip (HTTP 200 — odblokowane jak porntrex 2026-05-22),
-więc resolvujemy SERVER-SIDE zamiast WebView fallback (który pokazywał stronę +
-preroll-reklamę i scrape łapał reklamę zamiast wideo).
-
-KVS flashvars: `video_url` / `video_alt_url` / `video_alt_url2` (480/720/1080p), każdy w
-formie `function/0/https://yesporn.vip/get_file/<srv>/<HASH>/...mp4/` + `license_code`.
-kt_player dekoduje HASH algorytmem license_code (permutacja pierwszych 32 znaków —
-identyczny algo jak yt-dlp KVS; zweryfikowany 2026-05-31 że odtwarza output kt_player).
-
-Zdekodowany `get_file` 302 → `cdn5.yesporn.vip/remote_control.php?time=&cv=` który
-serwuje wideo bezpośrednio (206 video/mp4). Ten finalny URL jest **time-bound signed,
-NIE IP/cookie-bound** — zweryfikowane cross-IP (VPS resolve → fetch z innego IP = 206).
-Resolvujemy 302 w tej samej sesji co fetch strony (get_file token bywa session-bound),
-oddajemy finalny portable CDN url per jakość. Mobile gra direct natywnie, zero WebView,
-zero reklam, zero VPS proxy bandwidth.
+VPS dociera do yesporn.vip (HTTP 200, odblokowane jak porntrex 2026-05-22) → resolvujemy
+SERVER-SIDE (decode function/0 get_file + follow 302 → portable cdn5 url) zamiast WebView
+fallback który pokazywał stronę + preroll-reklamę. Native, multi-quality, zero reklam.
+Zweryfikowane 2026-05-31: prod 3×cdn5, emulator 1080p picker → native decode.
 """
 from __future__ import annotations
 
-import logging
-import re
-import time
-import urllib.parse as _up
-
-from app.extractors._fetch import _DEFAULT_IMPERSONATE, _DEFAULT_UA, _HAS_CURL_CFFI, fetch_tube_html
 from app.extractors._models import StreamSource
-
-log = logging.getLogger(__name__)
+from app.extractors.tubes import _kvs
 
 _BASE = "https://yesporn.vip"
 
-# flashvars: `video_url: 'function/0/https://.../get_file/...'` (+ alt/alt2 dla wyższych q).
-_URL_RE = re.compile(
-    r"(video(?:_alt)?_url\d*)\s*:\s*'(function/0/[^']+)'",
-    re.IGNORECASE,
-)
-_TEXT_RE = re.compile(
-    r"(video(?:_alt)?_url\d*)_text\s*:\s*'([^']*)'",
-    re.IGNORECASE,
-)
-_LICENSE_RE = re.compile(r"license_code\s*:\s*'(\$[^']+)'", re.IGNORECASE)
-
-_HASH_LENGTH = 32
-
-
-def _kvs_license_token(license_code: str) -> list[int]:
-    """kt_player license token (algo zgodny z yt-dlp KVS)."""
-    license_code = license_code.replace("$", "")
-    license_values = [int(c) for c in license_code]
-    modlicense = license_code.replace("0", "1")
-    center = len(modlicense) // 2
-    modlicense = str(4 * abs(int(modlicense[:center + 1]) - int(modlicense[center:])))[:center + 1]
-    return [
-        (license_values[index + offset] + current) % 10
-        for index, current in enumerate(int(c) for c in modlicense)
-        for offset in range(4)
-    ]
-
-
-def _kvs_real_url(video_url: str, license_code: str) -> str:
-    """Dekoduje `function/0/...get_file/N/<HASH>/...` permutując pierwsze 32 znaki HASH."""
-    if not video_url.startswith("function/0/"):
-        return video_url
-    parsed = _up.urlparse(video_url[len("function/0/"):])
-    lt = _kvs_license_token(license_code)
-    parts = parsed.path.split("/")
-    h = parts[3][:_HASH_LENGTH]
-    idx = list(range(_HASH_LENGTH))
-    acc = 0
-    for src in reversed(range(_HASH_LENGTH)):
-        acc += lt[src]
-        dest = (src + acc) % _HASH_LENGTH
-        idx[src], idx[dest] = idx[dest], idx[src]
-    parts[3] = "".join(h[i] for i in idx) + parts[3][_HASH_LENGTH:]
-    return _up.urlunparse(parsed._replace(path="/".join(parts)))
-
-
-def _quality_rank(label: str | None) -> int:
-    if not label:
-        return -1
-    m = re.search(r"(\d{3,4})\s*p", label, re.IGNORECASE)
-    return int(m.group(1)) if m else -1
-
-
-def _resolve_get_file(session, get_file_url: str, timeout: float) -> str | None:
-    """Follow get_file 302 → finalny portable CDN url (w sesji z cookies strony)."""
-    sep = "&" if "?" in get_file_url else "?"
-    url = f"{get_file_url}{sep}rnd={int(time.time() * 1000)}"
-    try:
-        r = session.get(
-            url, timeout=timeout, allow_redirects=True, stream=True,
-            headers={"Referer": _BASE + "/", "Range": "bytes=0-1"},
-        )
-        final = str(r.url)
-        status = r.status_code
-        r.close()
-    except Exception as e:
-        log.info("yespornvip: get_file resolve failed (%s): %s", get_file_url[:60], e)
-        return None
-    if status >= 400 or "/get_file/" in final:
-        log.info("yespornvip: get_file resolve bad status=%s final=%s", status, final[:70])
-        return None
-    return final
-
 
 def extract(page_url: str, *, timeout: float = 60.0) -> list[StreamSource] | None:
-    # Wspólna sesja: get_file token jest session-bound, więc 302 MUSI lecieć w tej samej
-    # sesji curl_cffi co fetch strony.
-    if not _HAS_CURL_CFFI:
-        # Bez curl_cffi nie zrobimy same-session resolve — brak fallbacku (WebView fallback
-        # pokazywał reklamy; lepiej None niż reklama).
-        log.info("yespornvip: curl_cffi unavailable — cannot resolve")
-        return None
-
-    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("yespornvip: page fetch failed %s: %s", page_url, e)
-        return None
-    if not html:
-        log.info("yespornvip: empty page %s", page_url)
-        return None
-
-    licm = _LICENSE_RE.search(html)
-    if not licm:
-        log.info("yespornvip: no license_code on %s", page_url)
-        return None
-    license_code = licm.group(1)
-
-    quality_by_var: dict[str, str] = {}
-    for m in _TEXT_RE.finditer(html):
-        quality_by_var[m.group(1).lower()] = m.group(2).strip()
-
-    seen_dec: set[str] = set()
-    result: list[StreamSource] = []
-    for m in _URL_RE.finditer(html):
-        var_name = m.group(1).lower()
-        decoded = _kvs_real_url(m.group(2), license_code)
-        if decoded in seen_dec:
-            continue
-        seen_dec.add(decoded)
-        final = _resolve_get_file(session, decoded, timeout)
-        if not final:
-            continue
-        result.append(
-            StreamSource(
-                link=final,
-                type="mp4",
-                quality=quality_by_var.get(var_name) or None,
-                referer=_BASE + "/",
-                # cdn5 remote_control.php url: time-bound signed, NIE IP/cookie-bound
-                # (zweryfikowane cross-IP) → mobile gra direct, zero proxy.
-                raw={"mobile_direct_ok": True},
-            )
-        )
-
-    if not result:
-        log.info("yespornvip: no resolvable get_file on %s", page_url)
-        return None
-
-    result.sort(key=lambda s: _quality_rank(s.quality), reverse=True)
-    return result
+    return _kvs.resolve_kvs(page_url, base_url=_BASE, timeout=timeout)