Show data agent review summaries in sessions

2026-04-30 15:11:46 +08:00
parent 5bdc0ffb01
commit 97bf5b9a54
2 changed files with 168 additions and 10 deletions
@@ -1118,17 +1118,25 @@ class LocalCodingAgent:
                if history_entry is not None:
                    file_history.append(history_entry)
                if tool_result.metadata.get('requires_user_review') is True:
+                    review_output = self._build_user_review_required_output(tool_result)
+                    review_message_id = f'assistant_{len(session.messages)}'
+                    session.append_assistant(
+                        review_output,
+                        message_id=review_message_id,
+                        stop_reason='user_review_required',
+                    )
                    stream_events.append(
                        {
                            'type': 'user_review_required',
                            'tool_name': tool_call.name,
                            'tool_call_id': tool_call.id,
                            'message_id': session.messages[tool_message_index].message_id,
+                            'review_message_id': review_message_id,
                            'metadata': dict(tool_result.metadata),
                        }
                    )
                    result = AgentRunResult(
-                        final_output=self._build_user_review_required_output(tool_result),
+                        final_output=review_output,
                        turns=turn_index + 1,
                        tool_calls=tool_calls,
                        transcript=session.transcript(),
@@ -1904,10 +1912,60 @@ class LocalCodingAgent:
            payload = json.loads(tool_result.content)
        except json.JSONDecodeError:
            return fallback
-        plan = payload.get('plan') if isinstance(payload, dict) else None
-        if not isinstance(plan, dict):
+        if not isinstance(payload, dict):
            return fallback
+        goal = payload.get('goal')
+        if isinstance(goal, dict):
+            return self._format_generation_goal_review(goal)
+        plan = payload.get('plan')
+        if isinstance(plan, dict):
+            return self._format_generation_plan_review(plan)
+        return fallback

+    def _format_generation_goal_review(self, goal: dict[str, object]) -> str:
+        lines = [
+            '已生成待 review 的数据生成目标，本轮已暂停。',
+            '',
+            f"- goal_id: `{goal.get('goal_id', '')}`",
+            f"- revision: `{goal.get('revision', '')}`",
+            f"- dataset_label: {goal.get('dataset_label', '')}",
+            f"- 目标摘要: {goal.get('goal_summary', '')}",
+        ]
+        targets = goal.get('target_definitions')
+        if isinstance(targets, list) and targets:
+            lines.append('- target_definitions:')
+            for item in targets:
+                if isinstance(item, dict):
+                    lines.append(
+                        f"  - {item.get('name', '')}: `{item.get('target', '')}`；规则：{item.get('rule', '')}"
+                    )
+        elif goal.get('target'):
+            lines.append(f"- target: `{goal.get('target')}`")
+        if goal.get('plan_hint'):
+            lines.append(f"- 计划提示: {goal.get('plan_hint')}")
+        open_questions = goal.get('open_questions')
+        if isinstance(open_questions, list) and open_questions:
+            lines.append('- 待确认问题:')
+            for item in open_questions:
+                if isinstance(item, str) and item:
+                    lines.append(f'  - {item}')
+        source_refs = goal.get('source_refs')
+        if isinstance(source_refs, list) and source_refs:
+            lines.append('- 来源:')
+            for item in source_refs:
+                if isinstance(item, str) and item:
+                    lines.append(f'  - {item}')
+        lines.extend(
+            [
+                f"- 覆盖范围: {goal.get('coverage', '')}",
+                f"- 排除项: {goal.get('exclusions', '')}",
+                '',
+                '请 review 这个生成目标：需要修改就直接回复修改意见；认可的话回复“确认目标”。',
+            ]
+        )
+        return '\n'.join(lines)
+
+    def _format_generation_plan_review(self, plan: dict[str, object]) -> str:
        lines = [
            '已生成待 review 的数据生成计划，本轮已暂停。',
            '',
@@ -274,29 +274,29 @@ class AgentRuntimeTests(unittest.TestCase):
        self.assertGreaterEqual(len(result.transcript), 5)
        self.assertGreaterEqual(len(result.file_history), 0)

-    def test_agent_stops_after_data_agent_plan_requires_review(self) -> None:
+    def test_agent_stops_after_data_agent_goal_requires_review(self) -> None:
        responses = [
            {
                'choices': [
                    {
                        'message': {
                            'role': 'assistant',
-                            'content': 'I will prepare a reviewable plan.',
+                            'content': 'I will prepare a reviewable generation goal.',
                            'tool_calls': [
                                {
                                    'id': 'call_1',
                                    'type': 'function',
                                    'function': {
-                                        'name': 'data_agent_prepare_generation_plan',
+                                        'name': 'data_agent_prepare_generation_goal',
                                        'arguments': json.dumps(
                                            {
                                                'dataset_label': '地图和生活边界数据',
+                                                'goal_summary': '生成附近生活服务边界数据',
                                                'target': 'Agent(tag="life_service")',
-                                                'total_count': 2,
-                                                'turn_mix': '1 条单轮，1 条多轮',
+                                                'plan_hint': '建议先生成 10 条单轮，输出到 tasks/demo/records.jsonl；具体数量、轮次和路径在 generation plan 中确认。',
                                                'coverage': '附近吃喝玩乐',
                                                'exclusions': '不要生成导航路线类 query',
-                                                'output_path': 'tasks/demo/artifacts',
+                                                'source_refs': ['manual:user'],
                                            }
                                        ),
                                    },
@@ -318,14 +318,114 @@ class AgentRuntimeTests(unittest.TestCase):
                    ),
                    runtime_config=AgentRuntimeConfig(cwd=workspace),
                )
-                result = agent.run('Prepare data generation plan')
+                result = agent.run('Prepare data generation goal')

        self.assertEqual(result.stop_reason, 'user_review_required')
        self.assertEqual(result.tool_calls, 1)
        self.assertIn('本轮已暂停', result.final_output)
+        self.assertIn('goal_id', result.final_output)
+        self.assertIn('计划提示', result.final_output)
+        self.assertIn('建议先生成 10 条单轮', result.final_output)
+        review_messages = [
+            entry
+            for entry in result.transcript
+            if entry.get('role') == 'assistant'
+            and entry.get('stop_reason') == 'user_review_required'
+        ]
+        self.assertEqual(len(review_messages), 1)
+        self.assertIn('数据生成目标', str(review_messages[0].get('content', '')))
        self.assertTrue(
            any(event.get('type') == 'user_review_required' for event in result.events)
        )
+        self.assertTrue(
+            any(
+                event.get('type') == 'user_review_required'
+                and isinstance(event.get('review_message_id'), str)
+                for event in result.events
+            )
+        )
+
+    def test_agent_appends_review_message_after_data_agent_plan_requires_review(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            workspace = Path(tmp_dir)
+            context = build_tool_context(AgentRuntimeConfig(cwd=workspace))
+            goal_result = execute_tool(
+                default_tool_registry(),
+                'data_agent_prepare_generation_goal',
+                {
+                    'dataset_label': '地图和生活边界数据',
+                    'goal_summary': '生成附近生活服务边界数据',
+                    'target': 'Agent(tag="life_service")',
+                    'coverage': '附近吃喝玩乐',
+                    'exclusions': '不要生成导航路线类 query',
+                    'source_refs': ['manual:user'],
+                },
+                context,
+            )
+            self.assertTrue(goal_result.ok, goal_result.content)
+            goal_id = json.loads(goal_result.content)['goal']['goal_id']
+            confirm_result = execute_tool(
+                default_tool_registry(),
+                'data_agent_confirm_generation_goal',
+                {'goal_id': goal_id, 'confirmation': '确认目标', 'reviewed_revision': 1},
+                context,
+            )
+            self.assertTrue(confirm_result.ok, confirm_result.content)
+            responses = [
+                {
+                    'choices': [
+                        {
+                            'message': {
+                                'role': 'assistant',
+                                'content': 'I will prepare a reviewable generation plan.',
+                                'tool_calls': [
+                                    {
+                                        'id': 'call_1',
+                                        'type': 'function',
+                                        'function': {
+                                            'name': 'data_agent_prepare_generation_plan',
+                                            'arguments': json.dumps(
+                                                {
+                                                    'confirmed_goal_id': goal_id,
+                                                    'dataset_label': '地图和生活边界数据',
+                                                    'target': 'Agent(tag="life_service")',
+                                                    'total_count': 10,
+                                                    'turn_mix': '10 条单轮',
+                                                    'coverage': '附近吃喝玩乐',
+                                                    'exclusions': '不要生成导航路线类 query',
+                                                    'output_path': 'tasks/demo/records.jsonl',
+                                                }
+                                            ),
+                                        },
+                                    }
+                                ],
+                            },
+                            'finish_reason': 'tool_calls',
+                        }
+                    ]
+                }
+            ]
+            with patch('src.openai_compat.request.urlopen', side_effect=make_urlopen_side_effect(responses)):
+                agent = LocalCodingAgent(
+                    model_config=ModelConfig(
+                        model='Qwen/Qwen3-Coder-30B-A3B-Instruct',
+                        base_url='http://127.0.0.1:8000/v1',
+                    ),
+                    runtime_config=AgentRuntimeConfig(cwd=workspace),
+                )
+                result = agent.run('Prepare data generation plan')
+
+        self.assertEqual(result.stop_reason, 'user_review_required')
+        self.assertIn('数据生成计划', result.final_output)
+        self.assertIn('plan_id', result.final_output)
+        review_messages = [
+            entry
+            for entry in result.transcript
+            if entry.get('role') == 'assistant'
+            and entry.get('stop_reason') == 'user_review_required'
+        ]
+        self.assertEqual(len(review_messages), 1)
+        self.assertIn('数据生成计划', str(review_messages[0].get('content', '')))

    def test_write_tool_is_blocked_without_permission(self) -> None:
        with tempfile.TemporaryDirectory() as tmp_dir: