diff --git a/app.py b/app.py index 7f9c38f..de44f58 100644 --- a/app.py +++ b/app.py @@ -260,10 +260,10 @@ def save_smart_config(cfg): # ===== 网址缓存 + 失败重试 配置 ===== -# cache: enabled=是否启用网址缓存(命中直接返回历史, 不重新提取); ttl_seconds=缓存有效期(秒) -# retry: max_retries=提取失败后的重试次数(默认2, 即最多尝试3次); delay_seconds=每次重试间隔(秒) -DEFAULT_CACHE_CONFIG = {"enabled": True, "ttl_seconds": 3600} -DEFAULT_RETRY_CONFIG = {"max_retries": 2, "delay_seconds": 2} +# cache: enabled=是否启用网址缓存(命中直接返回历史, 不重新提取); ttl_seconds=缓存有效期(秒, 默认7天, -1=永久有效) +# retry: max_retries=提取失败后的重试次数(默认2, 即最多尝试3次); delay_seconds=每次重试间隔(秒, 默认15) +DEFAULT_CACHE_CONFIG = {"enabled": True, "ttl_seconds": 7 * 24 * 3600} # 7天 +DEFAULT_RETRY_CONFIG = {"max_retries": 2, "delay_seconds": 15} def load_capture_settings(): @@ -310,17 +310,28 @@ def save_capture_settings(settings): def find_cache_hit(url, action, ttl_seconds): """ 按 url+action 在提取历史中查缓存：取最近一条成功记录，且创建时间在 TTL 内。 + ttl_seconds < 0 表示永久有效（不按时间过滤）。 命中返回 dict(记录)，未命中返回 None。 """ - from datetime import timedelta - cutoff = (datetime.now() - timedelta(seconds=max(1, int(ttl_seconds)))).strftime("%Y-%m-%d %H:%M:%S") conn = get_db() - row = conn.execute( - "SELECT id, url, title, backend, file_path, content, html_path, created_at FROM captures " - "WHERE url=? AND action=? AND status='success' AND created_at >= ? " - "ORDER BY id DESC LIMIT 1", - (url, action, cutoff) - ).fetchone() + ttl = int(ttl_seconds) + if ttl < 0: + # 永久有效 + row = conn.execute( + "SELECT id, url, title, backend, file_path, content, html_path, created_at FROM captures " + "WHERE url=? AND action=? AND status='success' " + "ORDER BY id DESC LIMIT 1", + (url, action) + ).fetchone() + else: + from datetime import timedelta + cutoff = (datetime.now() - timedelta(seconds=max(1, ttl))).strftime("%Y-%m-%d %H:%M:%S") + row = conn.execute( + "SELECT id, url, title, backend, file_path, content, html_path, created_at FROM captures " + "WHERE url=? AND action=? AND status='success' AND created_at >= ? " + "ORDER BY id DESC LIMIT 1", + (url, action, cutoff) + ).fetchone() conn.close() return dict(row) if row else None @@ -1585,7 +1596,7 @@ def capture(): "action": "screenshot" | "html" | "text" | "smart", "refresh": false, // 可选：true=强制实时提取(绕过网址缓存) "cache": true, // 可选：false=本次请求禁用缓存 - "cache_ttl_seconds": 3600, // 可选：本次请求的缓存时效(秒)，覆盖全局设置；命中判定也用它 + "cache_ttl_seconds": 604800, // 可选：本次请求的缓存时效(秒, -1=永久)，覆盖全局设置；命中判定也用它 "scroll_times": 0, "scroll_delay": 1000, "full_page": false, @@ -1639,10 +1650,11 @@ def capture(): cache_cfg = settings.get("cache", DEFAULT_CACHE_CONFIG) retry_cfg = settings.get("retry", DEFAULT_RETRY_CONFIG) force_refresh = bool(data.get("refresh") or data.get("force")) - # per-request 缓存参数：cache_ttl_seconds=本次缓存时效(覆盖全局)，cache=false / no_cache=true=本次禁用缓存 + # per-request 缓存参数：cache_ttl_seconds=本次缓存时效(秒, -1=永久, 覆盖全局)，cache=false / no_cache=true=本次禁用缓存 no_cache = data.get("cache") is False or str(data.get("no_cache")).lower() in ("true", "1", "yes") try: - cache_ttl_override = max(1, int(data.get("cache_ttl_seconds"))) if data.get("cache_ttl_seconds") is not None else None + v = int(data.get("cache_ttl_seconds")) + cache_ttl_override = v if v != 0 else None except Exception: cache_ttl_override = None diff --git a/templates/index.html b/templates/index.html index baad52a..9da6134 100644 --- a/templates/index.html +++ b/templates/index.html @@ -748,8 +748,8 @@

重要提示：