数据新鲜度校验: 所有模块加上日期/时效检查

1. search_cache.json → 添加 _metadata.updated_at 时间戳
   web_search.search_from_cache() 跳过超过36h的旧缓存
   防止某查询失败时残留旧数据

2. metrics_feedback.json → collector 检查mtime,超过24h不采用

3. trends.json → 已有 date==TODAY 校验(load_trends)

4. collector_ai_advice → DB每日覆盖,时序安全
   creator→topic状态位避免重复生成
   optimizer→文章状态位避免重复审查

scripts/collector.py

@@ -398,6 +398,11 @@ class SustainabilityCollector:
         metrics_file = DATA_DIR / "metrics_feedback.json"
         if metrics_file.exists():
             try:
+                mtime = datetime.datetime.fromtimestamp(metrics_file.stat().st_mtime)
+                age = (datetime.datetime.now() - mtime).total_seconds()
+                if age > 86400:  # 超过24h的数据不采用
+                    logger.debug("metrics_feedback 过时(%.0fh)，跳过", age / 3600)
+                else:
                     feedback = json.loads(metrics_file.read_text(encoding='utf-8'))
                     top_domains = feedback.get("top_domains", [])
                     if top_domains:

scripts/opencode_search.py

@@ -108,10 +108,13 @@ def refresh_cache():
             "AI工具 人工智能 效率提升 2026",
         ]
 
-    cache = {}
+    cache = {"_metadata": {"updated_at": datetime.datetime.now().isoformat()}}
     if SEARCH_CACHE_FILE.exists():
         try:
-            cache = json.loads(SEARCH_CACHE_FILE.read_text(encoding="utf-8"))
+            old = json.loads(SEARCH_CACHE_FILE.read_text(encoding="utf-8"))
+            for k, v in old.items():
+                if not k.startswith("_"):
+                    cache.setdefault(k, v)
         except Exception:
             pass
 

scripts/web_search.py

@@ -7,7 +7,7 @@
 2. Bing Web Search API（设 BING_API_KEY）
 3. Bing 网页抓取（服务器环境常反爬拦截）
 """
-import json, logging, os, re
+import json, logging, os, re, datetime
 from pathlib import Path
 from typing import List, Dict, Optional
 from urllib.parse import quote_plus
@@ -92,11 +92,21 @@ def search_scrape(query: str, max_results: int = 5) -> List[Dict]:
 
 
 def search_from_cache(query: str, max_results: int = 5) -> List[Dict]:
-    """从 opencode webfetch 预填充的缓存中读取"""
+    """从 opencode webfetch 预填充的缓存中读取（跳过超过36小时的缓存）"""
     if not SEARCH_CACHE_FILE.exists():
         return []
     try:
         cache = json.loads(SEARCH_CACHE_FILE.read_text(encoding="utf-8"))
+        meta = cache.get("_metadata", {})
+        updated = meta.get("updated_at", "")
+        if updated:
+            try:
+                age = (datetime.datetime.now() - datetime.datetime.fromisoformat(updated)).total_seconds()
+                if age > 129600:  # 36h
+                    logger.warning("搜索缓存过时(%dh)，跳过", int(age // 3600))
+                    return []
+            except Exception:
+                pass
         results = cache.get(query, [])
         return results[:max_results]
     except Exception: