auto 任务运行中实时显示已爬/待爬缓存数; 人工停止后继续爬取验证

- engine: CrawlJob 增加 _auto_visited/_auto_pending 运行中实时状态 (含发现新链接入队后刷新)
- app.py 详情接口: 任务运行中优先读内存实时值, 不再显示上次运行结束的旧数字
- 端到端验证: 人工停止(已爬3,缓存18) → 继续爬取(跳过seed,爬完18页,缓存0) ✅
- 运行中采样: visited 1→2 实时增长, pending 15→14 同步递减 ✅

.gitignore

@@ -5,3 +5,4 @@ data/cookies_*.json
 logs/
 out/
 data/exports/
+data/auto_state/

app.py

@@ -308,11 +308,19 @@ def api_task_detail(tid):
         result["run_page"] = page
         result["run_pages"] = pages
         result["run_page_size"] = page_size
-    # auto 状态数量 (独立文件)
+    # auto 状态数量: 运行中优先读内存实时值, 否则读状态文件
     if task.get("mode") == "auto":
+        live_v = live_p = None
+        with JOBS_LOCK:
+            job = JOBS.get(tid)
+            if job:
+                live_v = getattr(job, "_auto_visited", None)
+                live_p = getattr(job, "_auto_pending", None)
         st = store.load_auto_state(tid, task)
-        result["auto_pending_count"] = len(st.get("pending", []))
-        result["auto_visited_count"] = len(st.get("visited", []))
+        result["auto_pending_count"] = (
+            live_p if live_p is not None else len(st.get("pending", [])))
+        result["auto_visited_count"] = (
+            live_v if live_v is not None else len(st.get("visited", [])))
     return jsonify(result)
 
 

data/auto_state/t_99e73a95c81f.json

@@ -1,11 +0,0 @@
-{
-  "pending": [],
-  "visited": [
-    "http://127.0.0.1:38081/p2",
-    "http://127.0.0.1:38081/p5",
-    "http://127.0.0.1:38081/p3",
-    "http://127.0.0.1:38081/",
-    "http://127.0.0.1:38081/p1",
-    "http://127.0.0.1:38081/p4"
-  ]
-}

engine.py

@@ -226,6 +226,9 @@ class CrawlJob:
         self.persist = persist          # callable(task_id, run)
         self._stop = threading.Event()
         self._pause = threading.Event()
+        # 运行中的 auto 实时状态 (供详情接口读取; 任务结束时由状态文件兜底)
+        self._auto_visited = None
+        self._auto_pending = None
         self._cfg_lock = threading.RLock()
         self.thread = None
 
@@ -630,6 +633,8 @@ class CrawlJob:
             queued.add(normalize_url(u))
 
         run["progress"]["total"] = len(queue)
+        self._auto_visited = len(visited)
+        self._auto_pending = len(queue)
         self._persist()
 
         p, browser, ctx, page, cookie_file = self._open_browser()
@@ -650,6 +655,8 @@ class CrawlJob:
                 if key in visited and url != seed:     # 起始网址不去重, 其余已爬跳过
                     continue
                 visited.add(key)
+                self._auto_visited = len(visited)
+                self._auto_pending = len(queue)
                 idx = len(visited)
                 run["progress"]["current_url"] = url
                 run["progress"]["done"] = len(visited)
@@ -666,6 +673,7 @@ class CrawlJob:
                         if lk not in visited and lk not in queued:
                             queued.add(lk)
                             queue.append((link, depth + 1, url))
+                    self._auto_pending = len(queue)   # 新链接入队后实时刷新
                 if entry["status"] == "OK":
                     self._delay()
             run["progress"]["total"] = len(visited)