feat: Dashboard + History — AI quality metrics, run tracking, collaboration

runtime/api/main.py

@@ -148,6 +148,101 @@ def submit_feedback(payload: dict) -> dict:
     return {"status": "ok", "saved_to": str(fname)}
 
 
+@app.get("/history")
+def list_history() -> dict:
+    """List past test runs from workspace."""
+    import json as _json
+
+    ws = get_settings().workspace_dir
+    runs: list[dict] = []
+
+    # Scan workspace/_demo and workspace/执行日志 for run outputs
+    for scan_dir in [ws / "_demo", ws / "执行日志"]:
+        if not scan_dir.exists():
+            continue
+        for f in sorted(scan_dir.rglob("*.json"), reverse=True):
+            try:
+                data = _json.loads(f.read_text(encoding="utf-8"))
+                if isinstance(data, dict) and "run_id" in data:
+                    runs.append({
+                        "run_id": data.get("run_id", f.stem),
+                        "target": data.get("target", data.get("target_type", f.stem)),
+                        "date": data.get("date", data.get("timestamp", "")),
+                        "total": data.get("total", 0),
+                        "passed": data.get("succeeded", data.get("passed", 0)),
+                        "failed": data.get("failed", 0),
+                        "duration_s": data.get("duration_s", data.get("duration_ms", 0) / 1000 if "duration_ms" in data else 0),
+                        "confidence": data.get("confidence", 0),
+                    })
+            except Exception:
+                continue
+
+    return {"runs": runs[:50]}
+
+
+@app.get("/dashboard")
+def get_dashboard() -> dict:
+    """Aggregate quality metrics from all runs."""
+    import json as _json
+
+    ws = get_settings().workspace_dir
+    all_runs: list[dict] = []
+    expert_fails: dict[str, int] = {}
+
+    for scan_dir in [ws / "_demo", ws / "执行日志"]:
+        if not scan_dir.exists():
+            continue
+        for f in scan_dir.rglob("*.json"):
+            try:
+                data = _json.loads(f.read_text(encoding="utf-8"))
+                if isinstance(data, dict) and "total" in data:
+                    all_runs.append(data)
+                if "results" in data and isinstance(data["results"], dict):
+                    for node_id, r in data["results"].items():
+                        if not r.get("ok") and r.get("name"):
+                            name = r["name"]
+                            expert_fails[name] = expert_fails.get(name, 0) + 1
+            except Exception:
+                continue
+
+    total = len(all_runs)
+    if total == 0:
+        return {
+            "total_runs": 0, "avg_pass_rate": 0, "avg_confidence": 0,
+            "total_test_cases": 0, "recent_runs": [], "top_failures": [],
+        }
+
+    pass_rates = [(r.get("succeeded", r.get("passed", 0)) / max(r.get("total", 1), 1)) for r in all_runs]
+    confidences = [r.get("confidence", 0) for r in all_runs if isinstance(r.get("confidence"), (int, float))]
+    total_cases = sum(r.get("total", 0) for r in all_runs)
+
+    top = sorted(expert_fails.items(), key=lambda x: -x[1])[:10]
+
+    recent = sorted(all_runs, key=lambda r: str(r.get("date", r.get("timestamp", ""))), reverse=True)[:10]
+    recent_summaries = [
+        {
+            "run_id": r.get("run_id", ""),
+            "target": r.get("target", r.get("target_type", "")),
+            "date": str(r.get("date", r.get("timestamp", ""))),
+            "total": r.get("total", 0),
+            "passed": r.get("succeeded", r.get("passed", 0)),
+            "failed": r.get("failed", 0),
+            "confidence": r.get("confidence", 0),
+            "duration_s": r.get("duration_s", 0),
+        }
+        for r in recent
+    ]
+
+    return {
+        "total_runs": total,
+        "avg_pass_rate": sum(pass_rates) / total,
+        "avg_confidence": sum(confidences) / len(confidences) if confidences else 0,
+        "total_test_cases": total_cases,
+        "recent_runs": recent_summaries,
+        "top_failures": [{"expert": name, "fail_count": cnt} for name, cnt in top],
+    }
+
+
 def _run_in_background(run_id: str, decision) -> None:
     try:
         summary = _kernel.execute_sync(run_id, decision)

runtime/web/src/App.tsx

@@ -1,5 +1,5 @@
 import { Outlet, NavLink } from "react-router-dom";
-import { Beaker, Upload, BookOpen, Settings, Stethoscope, MessageSquare } from "lucide-react";
+import { Beaker, Upload, BookOpen, Settings, Stethoscope, MessageSquare, BarChart3, Clock } from "lucide-react";
 
 export default function App() {
   return (
@@ -25,6 +25,18 @@ export default function App() {
                   Catalog
                 </NavLink>
               </li>
+              <li>
+                <NavLink to="/dashboard" className={({ isActive }) => (isActive ? "font-semibold" : "")}>
+                  <BarChart3 className="inline w-4 h-4 mr-1" aria-hidden="true" />
+                  Dashboard
+                </NavLink>
+              </li>
+              <li>
+                <NavLink to="/history" className={({ isActive }) => (isActive ? "font-semibold" : "")}>
+                  <Clock className="inline w-4 h-4 mr-1" aria-hidden="true" />
+                  History
+                </NavLink>
+              </li>
               <li>
                 <NavLink to="/doctor" className={({ isActive }) => (isActive ? "font-semibold" : "")}>
                   <Stethoscope className="inline w-4 h-4 mr-1" aria-hidden="true" />

runtime/web/src/main.tsx

@@ -10,6 +10,8 @@ import CatalogPage from "./pages/CatalogPage";
 import SettingsPage from "./pages/SettingsPage";
 import DoctorPage from "./pages/DoctorPage";
 import FeedbackPage from "./pages/FeedbackPage";
+import DashboardPage from "./pages/DashboardPage";
+import HistoryPage from "./pages/HistoryPage";
 import "./index.css";
 
 const queryClient = new QueryClient({
@@ -28,6 +30,8 @@ ReactDOM.createRoot(document.getElementById("root")!).render(
             <Route path="/catalog" element={<CatalogPage />} />
             <Route path="/settings" element={<SettingsPage />} />
             <Route path="/doctor" element={<DoctorPage />} />
+            <Route path="/dashboard" element={<DashboardPage />} />
+            <Route path="/history" element={<HistoryPage />} />
             <Route path="/feedback" element={<FeedbackPage />} />
             <Route path="*" element={<Navigate to="/" replace />} />
           </Route>

runtime/web/src/pages/DashboardPage.tsx

@@ -0,0 +1,116 @@
+import { useEffect, useState } from "react";
+import { BarChart3, TrendingUp, AlertTriangle, CheckCircle2, Clock, Activity } from "lucide-react";
+
+interface RunSummary {
+  run_id: string;
+  target: string;
+  date: string;
+  total: number;
+  passed: number;
+  failed: number;
+  confidence: number;
+  duration_s: number;
+}
+
+interface DashboardData {
+  total_runs: number;
+  avg_pass_rate: number;
+  avg_confidence: number;
+  total_test_cases: number;
+  recent_runs: RunSummary[];
+  top_failures: { expert: string; fail_count: number }[];
+}
+
+const BASE = (import.meta as any).env?.VITE_API_BASE || "http://localhost:8800";
+
+export default function DashboardPage() {
+  const [data, setData] = useState<DashboardData | null>(null);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState("");
+
+  useEffect(() => {
+    fetch(`${BASE}/dashboard`)
+      .then((r) => { if (!r.ok) throw new Error("No data yet"); return r.json(); })
+      .then(setData)
+      .catch((e) => setError(e.message))
+      .finally(() => setLoading(false));
+  }, []);
+
+  if (loading) return <div className="p-6"><Activity className="w-5 h-5 animate-spin inline mr-2" />Loading dashboard...</div>;
+  if (error || !data) {
+    return (
+      <div className="max-w-4xl mx-auto p-6 space-y-4">
+        <h2 className="text-xl font-semibold flex items-center gap-2"><BarChart3 className="w-5 h-5" /> AI Quality Dashboard</h2>
+        <div className="p-8 text-center border rounded-lg bg-slate-50">
+          <BarChart3 className="w-12 h-12 text-slate-300 mx-auto mb-3" />
+          <p className="text-slate-500">No test runs yet. Start a test to populate quality metrics.</p>
+          <p className="text-xs text-slate-400 mt-2">Metrics are collected from completed test runs in your workspace.</p>
+        </div>
+      </div>
+    );
+  }
+
+  const passRate = data.avg_pass_rate * 100;
+  const confRate = data.avg_confidence * 100;
+
+  return (
+    <div className="max-w-4xl mx-auto p-6 space-y-6">
+      <h2 className="text-xl font-semibold flex items-center gap-2">
+        <BarChart3 className="w-5 h-5" /> AI Quality Dashboard
+      </h2>
+
+      {/* KPI Cards */}
+      <div className="grid grid-cols-2 md:grid-cols-4 gap-4">
+        {[
+          { label: "Total Runs", value: data.total_runs, icon: <Activity className="w-4 h-4" />, color: "bg-blue-50 border-blue-200" },
+          { label: "Pass Rate", value: `${passRate.toFixed(1)}%`, icon: <CheckCircle2 className="w-4 h-4" />, color: passRate >= 80 ? "bg-green-50 border-green-200" : "bg-yellow-50 border-yellow-200" },
+          { label: "Avg Confidence", value: `${confRate.toFixed(1)}%`, icon: <TrendingUp className="w-4 h-4" />, color: confRate >= 70 ? "bg-green-50 border-green-200" : "bg-yellow-50 border-yellow-200" },
+          { label: "Test Cases", value: data.total_test_cases, icon: <Clock className="w-4 h-4" />, color: "bg-purple-50 border-purple-200" },
+        ].map((kpi) => (
+          <div key={kpi.label} className={`p-4 rounded-lg border ${kpi.color}`}>
+            <div className="flex items-center gap-1.5 text-xs text-slate-500 mb-1">{kpi.icon}{kpi.label}</div>
+            <div className="text-2xl font-bold">{kpi.value}</div>
+          </div>
+        ))}
+      </div>
+
+      {/* Recent Runs */}
+      <section>
+        <h3 className="text-sm font-medium text-slate-600 mb-3">Recent Test Runs</h3>
+        <div className="space-y-2">
+          {data.recent_runs.map((run) => (
+            <div key={run.run_id} className="flex items-center justify-between px-4 py-3 border rounded-lg hover:bg-slate-50">
+              <div className="flex-1 min-w-0">
+                <div className="font-medium text-sm truncate">{run.target || run.run_id}</div>
+                <div className="text-xs text-slate-400">{run.date}</div>
+              </div>
+              <div className="flex items-center gap-4 text-sm">
+                <span className="text-green-600">{run.passed} pass</span>
+                {run.failed > 0 && <span className="text-red-600">{run.failed} fail</span>}
+                <span className="text-slate-400">{(run.confidence * 100).toFixed(0)}% conf</span>
+                <a href={`#/runs/${run.run_id}/report`} className="text-blue-600 hover:underline text-xs">Report →</a>
+              </div>
+            </div>
+          ))}
+        </div>
+      </section>
+
+      {/* Top Failures */}
+      {data.top_failures.length > 0 && (
+        <section>
+          <h3 className="text-sm font-medium text-slate-600 mb-3 flex items-center gap-1.5">
+            <AlertTriangle className="w-4 h-4 text-amber-500" /> Top Failing Areas
+          </h3>
+          <div className="space-y-1.5">
+            {data.top_failures.map((f) => (
+              <div key={f.expert} className="flex items-center justify-between px-3 py-2 border rounded text-sm">
+                <span>{f.expert}</span>
+                <span className="text-red-600 font-medium">{f.fail_count} failures</span>
+              </div>
+            ))}
+          </div>
+        </section>
+      )}
+    </div>
+  );
+}

runtime/web/src/pages/HistoryPage.tsx

@@ -0,0 +1,131 @@
+import { useEffect, useState } from "react";
+import { Clock, Search, Eye, Download, Share2 } from "lucide-react";
+
+interface RunMeta {
+  run_id: string;
+  target: string;
+  date: string;
+  total: number;
+  passed: number;
+  failed: number;
+  duration_s: number;
+  confidence: number;
+}
+
+const BASE = (import.meta as any).env?.VITE_API_BASE || "http://localhost:8800";
+
+export default function HistoryPage() {
+  const [runs, setRuns] = useState<RunMeta[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [search, setSearch] = useState("");
+  const [selected, setSelected] = useState<Set<string>>(new Set());
+
+  useEffect(() => {
+    fetch(`${BASE}/history`)
+      .then((r) => r.json())
+      .then((data) => setRuns(data.runs || []))
+      .catch(() => {})
+      .finally(() => setLoading(false));
+  }, []);
+
+  const filtered = runs.filter(
+    (r) => !search || r.target?.toLowerCase().includes(search.toLowerCase()) || r.run_id.includes(search)
+  );
+
+  const toggleSelect = (id: string) => {
+    const next = new Set(selected);
+    if (next.has(id)) next.delete(id); else next.add(id);
+    if (next.size > 2) return; // max 2 for compare
+    setSelected(next);
+  };
+
+  const exportBundle = async () => {
+    const toExport = Array.from(selected);
+    const bundle: Record<string, any> = {};
+    for (const id of toExport) {
+      try {
+        const r = await fetch(`${BASE}/report/${id}`);
+        if (r.ok) bundle[id] = await r.json();
+      } catch {}
+    }
+    const blob = new Blob([JSON.stringify(bundle, null, 2)], { type: "application/json" });
+    const url = URL.createObjectURL(blob);
+    const a = document.createElement("a");
+    a.href = url;
+    a.download = `tagent-bundle-${toExport.join("-")}.json`;
+    a.click();
+    URL.revokeObjectURL(url);
+  };
+
+  return (
+    <div className="max-w-4xl mx-auto p-6 space-y-4">
+      <div className="flex items-center justify-between">
+        <h2 className="text-xl font-semibold flex items-center gap-2">
+          <Clock className="w-5 h-5" /> Test History
+        </h2>
+        <div className="flex gap-2">
+          {selected.size === 2 && (
+            <button onClick={exportBundle} className="flex items-center gap-1.5 px-3 py-1.5 text-sm border rounded-lg hover:bg-slate-50">
+              <Share2 className="w-4 h-4" /> Export & Compare
+            </button>
+          )}
+        </div>
+      </div>
+
+      {/* Search */}
+      <div className="relative">
+        <Search className="w-4 h-4 absolute left-3 top-2.5 text-slate-400" />
+        <input
+          type="text"
+          value={search}
+          onChange={(e) => setSearch(e.target.value)}
+          placeholder="Search by target or run ID..."
+          className="w-full pl-9 pr-4 py-2 border rounded-lg text-sm"
+        />
+      </div>
+
+      {loading ? (
+        <p className="text-sm text-slate-400">Loading history...</p>
+      ) : filtered.length === 0 ? (
+        <div className="p-8 text-center border rounded-lg bg-slate-50">
+          <Clock className="w-10 h-10 text-slate-300 mx-auto mb-2" />
+          <p className="text-slate-500">No test history yet</p>
+          <p className="text-xs text-slate-400">Completed test runs will appear here automatically</p>
+        </div>
+      ) : (
+        <div className="space-y-2">
+          {filtered.map((run) => (
+            <div
+              key={run.run_id}
+              className={`flex items-center gap-3 px-4 py-3 border rounded-lg hover:bg-slate-50 cursor-pointer transition ${selected.has(run.run_id) ? "border-blue-400 bg-blue-50 ring-1 ring-blue-400" : ""}`}
+              onClick={() => toggleSelect(run.run_id)}
+            >
+              <input type="checkbox" checked={selected.has(run.run_id)} onChange={() => {}} className="shrink-0" />
+              <div className="flex-1 min-w-0">
+                <div className="font-medium text-sm truncate">{run.target || "Untitled"}</div>
+                <div className="text-xs text-slate-400">{run.date} · {run.duration_s}s</div>
+              </div>
+              <div className="flex items-center gap-3 text-sm shrink-0">
+                <span className="text-green-600">{run.passed}/{run.total} pass</span>
+                {run.failed > 0 && <span className="text-red-600">{run.failed} fail</span>}
+                <a
+                  href={`#/runs/${run.run_id}/report`}
+                  onClick={(e) => e.stopPropagation()}
+                  className="text-blue-600 hover:underline flex items-center gap-1"
+                >
+                  <Eye className="w-3.5 h-3.5" /> View
+                </a>
+              </div>
+            </div>
+          ))}
+        </div>
+      )}
+
+      {filtered.length > 0 && (
+        <p className="text-xs text-slate-400">
+          {selected.size === 0 ? "Select 2 runs to compare and export" : selected.size === 2 ? "2 selected — click Export & Compare" : `Select 1 more to compare (${2 - selected.size()} left)`}
+        </p>
+      )}
+    </div>
+  );
+}