some fix for nude save db with history with pic

games/nude.py

@@ -1,3 +1,5 @@
+import asyncio
+import json
 import logging
 import random
 from pathlib import PurePosixPath
@@ -6,6 +8,8 @@ from urllib.parse import urlparse
 import aiohttp
 from aiogram import types
 
+import config
+
 
 logger = logging.getLogger(__name__)
 
@@ -17,7 +21,6 @@ CAPTIONS = [
 
 LESAINT_POSTS_API = "https://lesaintdesseins.fr/wp-json/wp/v2/posts"
 LESAINT_REFERER = "https://lesaintdesseins.fr/"
-LATEST_POSTS_LIMIT = 30
 REQUEST_HEADERS = {
     "User-Agent": (
         "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
@@ -29,9 +32,12 @@ REQUEST_HEADERS = {
     "Referer": LESAINT_REFERER,
 }
 
+_NUDE_HISTORY_LOCK = asyncio.Lock()
+
 
 def _extract_photo_candidates(posts: list[dict]) -> list[dict]:
     candidates = []
+    seen_urls = set()
 
     for post in posts:
         embedded = post.get("_embedded") or {}
@@ -48,9 +54,10 @@ def _extract_photo_candidates(posts: list[dict]) -> list[dict]:
                 or media.get("source_url")
             )
 
-            if not image_url:
+            if not image_url or image_url in seen_urls:
                 continue
 
+            seen_urls.add(image_url)
             candidates.append(
                 {
                     "image_url": image_url,
@@ -73,25 +80,130 @@ def _guess_filename(image_url: str) -> str:
     return filename
 
 
-async def _fetch_latest_photo_candidates(session: aiohttp.ClientSession) -> list[dict]:
+def _load_nude_history() -> dict:
+    path = config.NUDE_HISTORY_PATH
+    if not path.exists():
+        return {"sent_images": []}
+
+    try:
+        with path.open("r", encoding="utf-8") as fh:
+            data = json.load(fh)
+    except (OSError, json.JSONDecodeError):
+        logger.exception("Failed to load nude history")
+        return {"sent_images": []}
+
+    if not isinstance(data, dict):
+        return {"sent_images": []}
+
+    sent_images = data.get("sent_images")
+    if not isinstance(sent_images, list):
+        sent_images = []
+
+    cleaned = [item for item in sent_images if isinstance(item, str) and item.strip()]
+    return {"sent_images": cleaned[-config.NUDE_HISTORY_MAX_SIZE :]}
+
+
+def _save_nude_history(data: dict) -> None:
+    path = config.NUDE_HISTORY_PATH
+    path.parent.mkdir(parents=True, exist_ok=True)
+    tmp_path = path.with_name(f"{path.name}.tmp")
+    with tmp_path.open("w", encoding="utf-8") as fh:
+        json.dump(data, fh, ensure_ascii=False, indent=2)
+    tmp_path.replace(path)
+
+
+def _remember_sent_image(history: dict, image_url: str) -> None:
+    sent_images = history.get("sent_images")
+    if not isinstance(sent_images, list):
+        sent_images = []
+
+    sent_images = [item for item in sent_images if item != image_url]
+    sent_images.append(image_url)
+    history["sent_images"] = sent_images[-config.NUDE_HISTORY_MAX_SIZE :]
+
+
+async def _fetch_posts_page(
+    session: aiohttp.ClientSession, page: int, per_page: int
+) -> tuple[list[dict], int | None]:
     params = {
-        "per_page": str(LATEST_POSTS_LIMIT),
+        "page": str(page),
+        "per_page": str(per_page),
         "orderby": "date",
         "order": "desc",
         "_embed": "wp:featuredmedia",
     }
 
     async with session.get(LESAINT_POSTS_API, params=params, headers=REQUEST_HEADERS) as resp:
+        if resp.status == 400 and page > 1:
+            body = await resp.text()
+            if "rest_post_invalid_page_number" in body:
+                return [], None
+            raise RuntimeError(f"API returned 400: {body[:300]}")
+
         if resp.status != 200:
             body = await resp.text()
             raise RuntimeError(f"API returned {resp.status}: {body[:300]}")
 
+        total_pages_header = resp.headers.get("X-WP-TotalPages")
+        total_pages = (
+            int(total_pages_header)
+            if total_pages_header and total_pages_header.isdigit()
+            else None
+        )
         posts = await resp.json(content_type=None)
 
     if not isinstance(posts, list):
         raise RuntimeError("Unexpected API response format")
 
-    return _extract_photo_candidates(posts)
+    return posts, total_pages
+
+
+async def _fetch_latest_photo_candidates(
+    session: aiohttp.ClientSession, sent_urls: set[str]
+) -> list[dict]:
+    per_page = max(1, min(config.NUDE_POSTS_PER_PAGE, 100))
+    max_scan = max(config.NUDE_POSTS_POOL_TARGET, config.NUDE_MAX_POSTS_TO_SCAN)
+    max_pages = max(1, (max_scan + per_page - 1) // per_page)
+
+    candidates = []
+    seen_urls = set()
+    unseen_count = 0
+    total_pages = None
+
+    for page in range(1, max_pages + 1):
+        if total_pages is not None and page > total_pages:
+            break
+
+        posts, reported_total_pages = await _fetch_posts_page(session, page, per_page)
+        if reported_total_pages is not None:
+            total_pages = reported_total_pages
+        if not posts:
+            break
+
+        for candidate in _extract_photo_candidates(posts):
+            image_url = candidate["image_url"]
+            if image_url in seen_urls:
+                continue
+
+            seen_urls.add(image_url)
+            candidates.append(candidate)
+
+            if image_url not in sent_urls:
+                unseen_count += 1
+
+        if len(candidates) >= config.NUDE_POSTS_POOL_TARGET and unseen_count >= config.NUDE_MIN_UNSEEN_POOL:
+            break
+        if len(candidates) >= max_scan:
+            break
+
+    return candidates
+
+
+def _pick_unsent_candidate(candidates: list[dict], sent_urls: set[str]) -> dict | None:
+    unseen = [candidate for candidate in candidates if candidate["image_url"] not in sent_urls]
+    if not unseen:
+        return None
+    return random.choice(unseen)
 
 
 async def _download_photo(
@@ -120,24 +232,39 @@ async def handle_nude_cmd(message: types.Message):
             "🔥 <b>Справка по команде /nude:</b>\n\n"
             "Берет случайную фотку из последних постов сайта le saint des seins.\n"
             "• Источник: <code>https://lesaintdesseins.fr/</code>\n"
-            "• Достает последние посты через WordPress REST API\n"
+            "• Пагинацией тянет 1000+ последних постов через WordPress REST API\n"
+            "• Хранит историю отправок и старается не повторять фотки\n"
             "• Выбирает случайную картинку и отправляет ее прямо в Telegram\n\n"
             "<i>Пример:</i> <code>/nude</code>"
         )
         await message.reply(help_msg, parse_mode="HTML")
         return
 
-    timeout = aiohttp.ClientTimeout(total=30)
+    timeout = aiohttp.ClientTimeout(total=60)
+    history_snapshot = _load_nude_history()
+    sent_urls_snapshot = set(history_snapshot["sent_images"])
 
     try:
         async with aiohttp.ClientSession(timeout=timeout) as session:
-            candidates = await _fetch_latest_photo_candidates(session)
+            candidates = await _fetch_latest_photo_candidates(session, sent_urls_snapshot)
 
             if not candidates:
                 await message.reply("Не удалось найти свежие фотки :(")
                 return
 
+            async with _NUDE_HISTORY_LOCK:
+                history = _load_nude_history()
+                sent_urls = set(history["sent_images"])
+                selected = _pick_unsent_candidate(candidates, sent_urls)
+
+                if selected is None:
+                    logger.info("Nude history exhausted current pool, resetting history")
+                    history = {"sent_images": []}
                     selected = random.choice(candidates)
+
+                _remember_sent_image(history, selected["image_url"])
+                _save_nude_history(history)
+
             caption = random.choice(CAPTIONS)
             image_bytes, filename = await _download_photo(session, selected["image_url"])