feat(report): annotate connectivity cases and add judged pass rate (ticket 02)

报告层推导连通用例标记（无判定结果 + 每轮有回复 + 无用例级错误），
summary 新增 connectivity_cases 与 judged_pass_rate（无判定型用例时为 null）。
对比报告同步标注且连通用例按引擎口径计通过；总通过率口径不变（ADR-0002）。

.scratch/v0.5/issues/02-connectivity-case-annotation.md

@@ -4,10 +4,10 @@
 
 **Blocked by:** None — can start immediately.
 
-**Status:** ready-for-agent
+**Status:** done
 
-- [ ] 单次报告中连通用例带明确标记，前端可见
+- [x] 单次报告中连通用例带明确标记，前端可见
-- [ ] 对比报告中连通用例同样标注
+- [x] 对比报告中连通用例同样标注
-- [ ] summary 含连通用例数与判定型通过率；全为连通用例时判定型通过率不除零
+- [x] summary 含连通用例数与判定型通过率；全为连通用例时判定型通过率不除零
-- [ ] 总通过率数值与升级前一致（口径未变）
+- [x] 总通过率数值与升级前一致（口径未变）
-- [ ] 报告单元测试覆盖标注与判定型通过率计算（先例：现有报告测试）
+- [x] 报告单元测试覆盖标注与判定型通过率计算（先例：现有报告测试）

backend/agenteval/evaluation/report.py

@@ -111,9 +111,11 @@ def generate_report(run_id: str, session=None) -> dict[str, Any]:
     # Group by case
     case_map: dict[str, dict[str, Any]] = {}
     for turn in turns:
-        case_map.setdefault(turn.case_id, {"turns": [], "results": []})
+        case_map.setdefault(turn.case_id, {"turns": [], "results": [], "all_replied": True})
         sent = turn.get_sent_message()
         reply = turn.get_reply()
+        if reply is None:
+            case_map[turn.case_id]["all_replied"] = False
         case_map[turn.case_id]["turns"].append(
             {
                 "round": turn.round_index,
@@ -125,7 +127,7 @@ def generate_report(run_id: str, session=None) -> dict[str, Any]:
         )
 
     for result in results:
-        case_map.setdefault(result.case_id, {"turns": [], "results": []})
+        case_map.setdefault(result.case_id, {"turns": [], "results": [], "all_replied": True})
         case_map[result.case_id]["results"].append(
             {
                 "rule_type": result.rule_type,
@@ -135,18 +137,35 @@ def generate_report(run_id: str, session=None) -> dict[str, Any]:
             }
         )
 
+    summary = run.summary or {}
+    errored_case_ids = {e.get("case_id") for e in summary.get("case_errors", [])}
+
     cases = []
     for case_id in sorted(case_map.keys()):
         item = case_map[case_id]
+        # 连通用例：无任何判定结果，且每轮都收到回复、无用例级错误（CONTEXT.md）
+        connectivity = (
+            not item["results"]
+            and bool(item["turns"])
+            and item["all_replied"]
+            and case_id not in errored_case_ids
+        )
         cases.append(
             {
                 "case_id": case_id,
+                "connectivity": connectivity,
                 "turns": sorted(item["turns"], key=lambda x: x["round"]),
                 "results": item["results"],
             }
         )
 
-    summary = run.summary or {}
+    total_cases = summary.get("total_cases", 0)
+    passed_cases = summary.get("passed_cases", 0)
+    connectivity_count = sum(1 for c in cases if c["connectivity"])
+    judged_total = total_cases - connectivity_count
+    # 连通用例按引擎口径计通过，判定型通过数 = 总通过数 - 连通用例数
+    judged_pass_rate = round((passed_cases - connectivity_count) / judged_total, 4) if judged_total > 0 else None
+
     return {
         "run_id": run.id,
         "target_id": run.target_id,
@@ -157,12 +176,14 @@ def generate_report(run_id: str, session=None) -> dict[str, Any]:
         "started_at": iso_utc(run.started_at),
         "completed_at": iso_utc(run.completed_at),
         "summary": {
-            "total_cases": summary.get("total_cases", 0),
+            "total_cases": total_cases,
-            "passed_cases": summary.get("passed_cases", 0),
+            "passed_cases": passed_cases,
             "failed_cases": summary.get("failed_cases", 0),
             "total_rules": summary.get("total_rules", 0),
             "passed_rules": summary.get("passed_rules", 0),
             "pass_rate": summary.get("pass_rate", 0.0),
+            "connectivity_cases": connectivity_count,
+            "judged_pass_rate": judged_pass_rate,
         },
         "cases": cases,
     }
@@ -194,6 +215,9 @@ def generate_compare_report(run_id_1: str, run_id_2: str, session=None) -> dict[
         def _case_passed(c):
             if not c:
                 return None
+            if c.get("connectivity"):
+                # 连通用例收到回复即通过（引擎口径）
+                return True
             results = c.get("results", [])
             if not results:
                 # No rule results (e.g. errored case) must not count as passed.
@@ -203,6 +227,7 @@ def generate_compare_report(run_id_1: str, run_id_2: str, session=None) -> dict[
         case_diffs.append(
             {
                 "case_id": cid,
+                "connectivity": bool((ca and ca.get("connectivity")) or (cb and cb.get("connectivity"))),
                 "run_a_passed": _case_passed(ca),
                 "run_b_passed": _case_passed(cb),
                 "changed": _case_passed(ca) != _case_passed(cb),
@@ -242,6 +267,8 @@ def render_markdown_report(run_id: str, session=None) -> str:
     """Render a report as Markdown string."""
     report = generate_report(run_id, session)
     s = report["summary"]
+    judged_rate = s.get("judged_pass_rate")
+    judged_rate_text = "—" if judged_rate is None else f"{judged_rate * 100:.1f}%"
     lines: list[str] = [
         f"# 评测报告 — {report.get('scenario_name', run_id)}",
         "",
@@ -261,15 +288,23 @@ def render_markdown_report(run_id: str, session=None) -> str:
         f"| 总规则数 | {s['total_rules']} |",
         f"| 通过规则 | {s['passed_rules']} |",
         f"| 通过率 | {s['pass_rate'] * 100:.1f}% |",
+        f"| 连通用例 | {s.get('connectivity_cases', 0)} |",
+        f"| 判定型通过率 | {judged_rate_text} |",
         "",
         "## 用例明细",
         "",
     ]
 
     for case in report.get("cases", []):
-        case_passed = all(r["passed"] for r in case.get("results", []))
+        if case.get("connectivity"):
-        badge = "✅" if case_passed else "❌"
+            badge = "🔗"
-        lines.append(f"### {badge} 用例 `{case['case_id']}`")
+        else:
+            case_passed = all(r["passed"] for r in case.get("results", []))
+            badge = "✅" if case_passed else "❌"
+        title = f"### {badge} 用例 `{case['case_id']}`"
+        if case.get("connectivity"):
+            title += "（连通用例，未配置判定标准）"
+        lines.append(title)
         lines.append("")
 
         for turn in case.get("turns", []):

frontend/web/src/pages/Reports.tsx

@@ -1,7 +1,7 @@
 import { useEffect, useMemo, useState } from 'react'
 import { useSearchParams } from 'react-router-dom'
 import {
-  Button, Card, Col, Collapse, Descriptions, Empty, Row, Segmented,
+  Alert, Button, Card, Col, Collapse, Descriptions, Empty, Row, Segmented,
   Select, Space, Spin, Statistic, Table, Tag, Tooltip, Badge, message,
 } from 'antd'
 import {
@@ -29,6 +29,7 @@ interface RuleResultData {
 
 interface CaseReport {
   case_id: string
+  connectivity: boolean
   turns: TurnData[]
   results: RuleResultData[]
 }
@@ -47,6 +48,8 @@ interface Report {
     total_rules: number
     passed_rules: number
     pass_rate: number
+    connectivity_cases: number
+    judged_pass_rate: number | null
   }
   cases: CaseReport[]
 }
@@ -57,6 +60,7 @@ interface CompareResult {
   delta: { pass_rate: number; passed_cases: number; passed_rules: number }
   cases: Array<{
     case_id: string
+    connectivity: boolean
     run_a_passed: boolean | null
     run_b_passed: boolean | null
     changed: boolean
@@ -364,6 +368,14 @@ function SingleReportView({ report }: { report: Report | null }) {
         </Col>
       </Row>
 
+      {report.summary.connectivity_cases > 0 && (
+        <Alert type="info" showIcon style={{ marginBottom: 16 }}
+          message={`本次运行含 ${report.summary.connectivity_cases} 个连通用例（未配置判定标准，收到回复即通过）` +
+            (report.summary.judged_pass_rate != null
+              ? `，判定型通过率 ${(report.summary.judged_pass_rate * 100).toFixed(1)}%`
+              : '，无判定型用例')} />
+      )}
+
       <Card style={{ marginBottom: 16 }}>
         <Descriptions size="small" column={2}>
           <Descriptions.Item label="评测对象">{report.target_name}</Descriptions.Item>
@@ -379,9 +391,11 @@ function SingleReportView({ report }: { report: Report | null }) {
           label: (
             <Space>
               <span style={{ fontWeight: 500 }}>{c.case_id}</span>
-              {c.results.every((r) => r.passed)
+              {c.connectivity
-                ? <Tag color="success">全部通过</Tag>
+                ? <Tag color="blue">连通用例</Tag>
-                : <Tag color="error">存在失败</Tag>}
+                : c.results.every((r) => r.passed)
+                  ? <Tag color="success">全部通过</Tag>
+                  : <Tag color="error">存在失败</Tag>}
               <Tag>{c.turns.length} 轮对话</Tag>
             </Space>
           ),
@@ -501,6 +515,7 @@ function CompareView({ result }: { result: CompareResult | null }) {
                 <Space>
                   {r.changed && <Badge dot color="orange" />}
                   <span style={{ fontWeight: r.changed ? 600 : 400 }}>{id}</span>
+                  {r.connectivity && <Tag color="blue">连通</Tag>}
                 </Space>
               ),
             },

tests/unit/test_report.py

@@ -40,8 +40,14 @@ def _seed_run(
     pass_rate: float = 1.0,
     n_cases: int = 1,
     scenario_id: str | None = None,
+    connectivity_cases: int = 0,
+    errored_cases: int = 0,
 ) -> str:
-    """Create a minimal completed run with real data in the DB and return run_id."""
+    """Create a minimal completed run with real data in the DB and return run_id.
+
+    connectivity_cases: extra cases with replied turns but no rule results.
+    errored_cases: extra cases with a missing-reply turn and no rule results.
+    """
     target = EvalTarget(
         name="测试对象",
         platform=PlatformType.AI_DIGITAL_EMPLOYEE,
@@ -54,7 +60,11 @@ def _seed_run(
     if scenario_id is None:
         scenario = Scenario(
             name="测试场景",
-            cases=[Case(id=f"c{i}", type=CaseType.SINGLE, messages=["hi"]) for i in range(n_cases)],
+            cases=(
+                [Case(id=f"c{i}", type=CaseType.SINGLE, messages=["hi"]) for i in range(n_cases)]
+                + [Case(id=f"conn{i}", type=CaseType.SINGLE, messages=["ping"]) for i in range(connectivity_cases)]
+                + [Case(id=f"err{i}", type=CaseType.SINGLE, messages=["ping"]) for i in range(errored_cases)]
+            ),
         )
         scenario = ScenarioRepository(session).create(scenario)
         scenario_id = scenario.id
@@ -95,13 +105,35 @@ def _seed_run(
         )
         result_repo.save_result(eval_result)
 
+    for i in range(connectivity_cases):
+        result_repo.save_turn(Turn(
+            run_id=run.id,
+            case_id=f"conn{i}",
+            round_index=1,
+            sent_message={"msgBody": {"content": "ping"}},
+            reply={"msgBody": {"content": "pong"}},
+            latency_ms=100,
+        ))
+
+    for i in range(errored_cases):
+        result_repo.save_turn(Turn(
+            run_id=run.id,
+            case_id=f"err{i}",
+            round_index=1,
+            sent_message={"msgBody": {"content": "ping"}},
+            reply=None,
+            latency_ms=None,
+        ))
+
+    all_total = total + connectivity_cases + errored_cases
+    all_passed = passed + connectivity_cases  # 连通用例收到回复即通过（引擎口径）
     run.summary = {
-        "total_cases": total,
+        "total_cases": all_total,
-        "passed_cases": passed,
+        "passed_cases": all_passed,
-        "failed_cases": total - passed,
+        "failed_cases": all_total - all_passed,
         "total_rules": total,
         "passed_rules": passed,
-        "pass_rate": round(pass_rate, 4),
+        "pass_rate": round(all_passed / all_total, 4) if all_total else 0.0,
     }
     RunRepository(session).update(run)
     return run.id
@@ -210,6 +242,67 @@ def test_compare_report_different_scenarios_rejected(report_session):
         generate_compare_report(run_id_a, run_id_b, report_session)
 
 
+# ── connectivity case annotation (ticket 02) ─────────────────────────────
+
+def test_report_marks_connectivity_case(report_session):
+    run_id = _seed_run(report_session, pass_rate=0.5, n_cases=2, connectivity_cases=1)
+    report = generate_report(run_id, report_session)
+
+    by_id = {c["case_id"]: c for c in report["cases"]}
+    assert by_id["conn0"]["connectivity"] is True
+    assert by_id["c0"]["connectivity"] is False
+    assert by_id["c1"]["connectivity"] is False
+
+    s = report["summary"]
+    assert s["connectivity_cases"] == 1
+    # 判定型通过率 = 判定型通过 1 ÷ 判定型总数 2
+    assert s["judged_pass_rate"] == 0.5
+    # 总通过率口径不变（含连通用例，ADR-0002）
+    assert s["pass_rate"] == round(2 / 3, 4)
+
+
+def test_errored_case_not_marked_connectivity(report_session):
+    """无判定结果但缺回复的用例是执行失败，不是连通用例。"""
+    run_id = _seed_run(report_session, n_cases=1, errored_cases=1)
+    report = generate_report(run_id, report_session)
+
+    by_id = {c["case_id"]: c for c in report["cases"]}
+    assert by_id["err0"]["connectivity"] is False
+    assert report["summary"]["connectivity_cases"] == 0
+
+
+def test_all_connectivity_judged_pass_rate_none(report_session):
+    """全为连通用例时判定型通过率为 None，不除零。"""
+    run_id = _seed_run(report_session, n_cases=0, connectivity_cases=2)
+    report = generate_report(run_id, report_session)
+
+    s = report["summary"]
+    assert s["connectivity_cases"] == 2
+    assert s["judged_pass_rate"] is None
+
+
+def test_compare_report_marks_connectivity(report_session):
+    run_id_a = _seed_run(report_session, n_cases=1, connectivity_cases=1)
+    sid = _scenario_of(report_session, run_id_a)
+    run_id_b = _seed_run(report_session, n_cases=1, connectivity_cases=1, scenario_id=sid)
+    result = generate_compare_report(run_id_a, run_id_b, report_session)
+
+    by_id = {c["case_id"]: c for c in result["cases"]}
+    assert by_id["conn0"]["connectivity"] is True
+    assert by_id["c0"]["connectivity"] is False
+    # 连通用例双方均按引擎口径视为通过，不应标记 changed
+    assert by_id["conn0"]["run_a_passed"] is True
+    assert by_id["conn0"]["run_b_passed"] is True
+    assert by_id["conn0"]["changed"] is False
+
+
+def test_markdown_report_shows_connectivity(report_session):
+    run_id = _seed_run(report_session, n_cases=1, connectivity_cases=1)
+    md = render_markdown_report(run_id, report_session)
+    assert "连通用例" in md
+    assert "判定型通过率" in md
+
+
 # ── render_markdown_report ────────────────────────────────────────────────
 
 def test_render_markdown_contains_header(report_session):