feat(latest): 最近抓取改为最近24小时每次运行各一张卡片

原先每源只显示最新一场,补跑+定时连跑时前一场被顶下,用户误以为覆盖。
现同源多场并列(按运行时间倒序,上限80张),卡片带运行类型(续采/修复);
文章明细提取拆为 _articles_from_data 纯函数供多场复用。

web/app.py

@@ -1400,12 +1400,9 @@ def dashboard():
                            fail_top=sorted(fail_top.items(), key=lambda kv: -kv[1])[:6])
 
 
-def _latest_run_articles(name):
-    """源最近一次运行的文章明细(标题/链接/状态/失败原因),供最近抓取页与调度中心详情抽屉共用。"""
-    runs = get_history_runs(name, include_data=True)
-    if not runs:
-        return []
-    biz, _legacy = get_business_result(runs[0]["data"])
+def _articles_from_data(data):
+    """从一次运行的归一化 JSON 提取文章明细(标题/链接/状态/失败原因)。"""
+    biz, _legacy = get_business_result(data)
     if not biz:
         return []
     return [{
@@ -1417,44 +1414,43 @@ def _latest_run_articles(name):
     } for a in biz.get("articles", [])]
 
 
+def _latest_run_articles(name):
+    """源最近一次运行的文章明细,供调度中心详情弹窗用。"""
+    runs = get_history_runs(name, include_data=True)
+    return _articles_from_data(runs[0]["data"]) if runs else []
+
+
 @app.route("/latest")
 @require_auth
 def latest_content():
-    """最近抓取:每个源最近一次运行抓到的文章标题与状态,快速盘点内容。"""
+    """最近抓取:最近 24 小时内每次运行各一张卡片(同源多场都展示,按时间倒序)。"""
     skills = pl.skill_sources()
     names = sorted((set(skills) | set(pl.list_schedules())) - {"digest"})
     names = [n for n in names
              if skills.get(n, {}).get("enabled", False)
              or pl.schedule_state(n) == "active"]
+    cutoff = datetime.now() - timedelta(hours=24)
     cards = []
     for n in names:
-        runs = get_history_runs(n, include_data=True)
-        if not runs:
-            cards.append({"name": n, "display": skills.get(n, {}).get("name") or n,
-                          "run_at": None, "articles": [], "succeeded": None,
-                          "failed": None, "exhausted": False})
-            continue
-        run = runs[0]
-        biz, _legacy = get_business_result(run["data"])
-        articles = []
-        if biz:
-            for a in biz.get("articles", []):
-                articles.append({
-                    "title": a.get("title") or a.get("url", ""),
-                    "url": a.get("url", ""),
-                    "status": a.get("status", ""),
-                })
-        cards.append({
-            "name": n,
-            "display": skills.get(n, {}).get("name") or n,
-            "run_at": run["run_at"],
-            "articles": articles,
-            "succeeded": run["succeeded_count"],
-            "failed": run["failed_count"],
-            "exhausted": bool(biz.get("source_exhausted")) if biz else False,
-        })
-    cards.sort(key=lambda c: c["run_at"] or "", reverse=True)
-    return render_template("latest.html", cards=cards)
+        display = skills.get(n, {}).get("name") or n
+        for run in get_history_runs(n, include_data=True):
+            try:
+                run_dt = datetime.strptime(run["run_at"], "%Y-%m-%d %H:%M:%S")
+            except ValueError:
+                continue
+            if run_dt < cutoff:
+                break  # 按时间倒序,再往前的都不在窗口内
+            cards.append({
+                "name": n, "display": display,
+                "run_at": run["run_at"], "mode": run["mode"],
+                "articles": _articles_from_data(run["data"]),
+                "succeeded": run["succeeded_count"],
+                "failed": run["failed_count"],
+                "exhausted": (lambda b: bool(b.get("source_exhausted")) if b else False)(
+                    get_business_result(run["data"])[0]),
+            })
+    cards.sort(key=lambda c: c["run_at"], reverse=True)
+    return render_template("latest.html", cards=cards[:80])
 
 
 @app.route("/schedules")

web/templates/latest.html

@@ -17,7 +17,7 @@
 
 <h1>最近抓取</h1>
 <p style="color:#666;font-size:13px">
-  每个源最近一次运行抓到的文章标题与状态,按运行时间倒序;点击标题打开原文链接。
+  最近 24 小时内每次运行各一张卡片(同源多场并列,如补跑+定时各一张),按运行时间倒序;点击标题打开原文链接。
   带已取尽标记=该源近期无新文章(正常),空列表=本次未交付(看上次运行详情排查)。
 </p>
 
@@ -26,7 +26,7 @@
   <summary>
     <code>{{ c.name }}</code>
     {% if c.display and c.display != c.name %}<span style="color:#888">{{ c.display }}</span>{% endif %}
-    <span class="runinfo">{{ c.run_at or '从未运行' }}</span>
+    <span class="runinfo">{{ c.run_at or '从未运行' }}{% if c.mode and c.mode != '常规采集' %} · {{ c.mode }}{% endif %}</span>
     {% if c.succeeded is not none %}
       <span class="badge ok">✓ {{ c.succeeded }}</span>
       {% if c.failed %}<span class="badge fail">✗ {{ c.failed }}</span>{% endif %}