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

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

app.py

@@ -104,13 +104,16 @@ def persist_cb(task_id, run):
         print(f"[db] 同步失败: {e}", flush=True)
 
 
-def start_run(task):
+def start_run(task, skip_seed=False):
-    """为任务启动一次爬取, 返回 (run, error)"""
+    """为任务启动一次爬取, 返回 (run, error)
+    skip_seed: auto 任务继续爬取模式, 跳过起始网址直接从缓存队列消费
+    """
     with JOBS_LOCK:
         job = JOBS.get(task["id"])
         if job and job.is_running():
             return None, "该任务已有正在运行的爬取"
         run = make_run(task)
+        run["skip_seed"] = bool(skip_seed)
         store.add_run(task["id"], run)
         job = CrawlJob(task, run, persist_cb)
         JOBS[task["id"]] = job
@@ -305,11 +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_pending_count"] = (
-        result["auto_visited_count"] = len(st.get("visited", []))
+            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)
 
 
@@ -424,6 +435,23 @@ def api_trash_clear():
 
 # ---------------- API: 运行控制 ----------------
 
+@app.route("/api/tasks/<tid>/continue", methods=["POST"])
+def api_continue(tid):
+    """继续爬取: auto 任务从待爬缓存队列接着爬 (跳过起始网址, 保留已爬集合)"""
+    task = store.get_task(tid)
+    if not task:
+        return jsonify({"error": "任务不存在"}), 404
+    if task.get("mode") != "auto":
+        return jsonify({"error": "仅自动爬取任务支持继续爬取"}), 400
+    st = store.load_auto_state(tid, task)
+    if not st.get("pending"):
+        return jsonify({"error": "没有待爬缓存链接，无需继续"}), 400
+    run, err = start_run(task, skip_seed=True)
+    if err:
+        return jsonify({"error": err}), 409
+    return jsonify(run)
+
+
 @app.route("/api/tasks/<tid>/start", methods=["POST"])
 def api_start(tid):
     task = store.get_task(tid)

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
 
@@ -598,9 +601,10 @@ class CrawlJob:
     def _crawl_auto(self):
         """自动爬取: 持续递归 爬取->发现链接->爬取... 直到无新链接可爬或达到上限
         - max_depth: 0=无限制, N=只爬 N 层
-        - max_pages: 0=无限制, N=安全上限
+        - max_pages: 0=无限制, N=安全上限 (继续爬取模式按本次新增页数重新计算)
         - 提取但未爬取的链接持久化到 auto.pending, 已爬集合持久化到 auto.visited
         - 停止后再次运行从缓存队列继续爬; 起始网址每次运行都重新爬(不去重)
+        - skip_seed(继续爬取): 跳过起始网址直接消费缓存队列, 页数上限按本次新增重新计算
         """
         run = self.run
         auto = self.task.get("auto", {})
@@ -615,8 +619,13 @@ class CrawlJob:
         state = store.load_auto_state(self.task["id"], self.task)
         visited = set(state.get("visited", []) or [])
         pending = state.get("pending", []) or []
+        visited_base = len(visited)
         # 起始网址每次运行都爬(不做去重), 缓存队列继续消费
-        queue = [(seed, 0, "")]
+        # 继续爬取模式(skip_seed): 有缓存队列时跳过起始网址, 直接从待爬队列接着爬
+        skip_seed = bool(self.run.get("skip_seed"))
+        queue = []
+        if not (skip_seed and pending):
+            queue.append((seed, 0, ""))
         if pending:
             queue.extend((p["url"], p.get("depth", 0), p.get("source", "")) for p in pending)
         queued = set(visited)
@@ -624,20 +633,30 @@ 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()
         try:
             while queue and not self._stop.is_set():
                 self._wait_if_paused()
-                if max_pages > 0 and len(visited) >= max_pages:
+                if max_pages > 0:
-                    self._log("info", f"达到最大页数上限 {max_pages}, 停止")
+                    if skip_seed:
-                    break
+                        # 继续爬取模式: 页数上限按本次新增页数重新计算 (累计 visited 不阻塞继续)
+                        if len(visited) - visited_base >= max_pages:
+                            self._log("info", f"本次继续爬取达到页数上限 {max_pages}, 停止")
+                            break
+                    elif len(visited) >= max_pages:
+                        self._log("info", f"达到最大页数上限 {max_pages}, 停止")
+                        break
                 url, depth, src = queue.pop(0)
                 key = normalize_url(url)
                 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)
@@ -654,6 +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)

static/app.js

@@ -602,7 +602,10 @@ function renderDetail() {
       ${t.running ? `
         <button class="btn" onclick="actTask('${t.id}','pause')">⏸ 暂停</button>
         <button class="btn danger" onclick="actTask('${t.id}','stop')">⏹ 终止</button>`
-      : `<button class="btn primary" onclick="actTask('${t.id}','start')">▶ 立即执行</button>`}
+      : `
+        <button class="btn primary" onclick="actTask('${t.id}','start')">▶ 立即执行</button>
+        ${t.mode === "auto" && (t.auto_pending_count || 0) > 0 ? `
+          <button class="btn" onclick="contTask('${t.id}')">⏩ 继续爬取（缓存 ${t.auto_pending_count} 条）</button>` : ""}`}
       ${t.mode === "auto" ? `<button class="btn" onclick="probeTask('${t.id}')">🧪 试爬取</button>
         <button class="btn" ${t.running ? "disabled" : ""} onclick="clearCache('${t.id}')">🧹 清空缓存</button>` : ""}
       <button class="btn" onclick="openEdit('${t.id}')">✏️ 编辑配置</button>
@@ -734,6 +737,15 @@ function stopLogPoll() {
   if (state.logTimer) { clearInterval(state.logTimer); state.logTimer = null; }
 }
 
+async function contTask(tid) {
+  try {
+    const r = await api(`/api/tasks/${tid}/continue`, { method: "POST" });
+    toast("已开始继续爬取缓存队列");
+    loadTasks();
+    if (state.detail.task && state.detail.task.id === tid) openDetail(tid);
+  } catch (e) { toast(e.message, true); }
+}
+
 /* ---------------- 打包导出 ---------------- */
 async function exportZip(tid) {
   const btn = window.event && window.event.target;