Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Python agent-contextagent-memoryagent-pluginsagentic-ragcontext-databasedsh-pluginknowledge-basememorymemory-managementself-evolvingvector-database
feat: memory v2 test suite with ov-install plugin update & split test cases (#1354)
kaisonglif1a89549430f
diff --git a/.github/workflows/api_test.yml b/.github/workflows/api_test.ymlindex 2611cd003..4b7bcfe38 100644--- a/.github/workflows/api_test.yml+++ b/.github/workflows/api_test.yml@@ -42,7 +42,7 @@ jobs:api-tests:name: API Integration Tests (${{ matrix.os }})runs-on: ${{ matrix.os }}- timeout-minutes: 50+ timeout-minutes: 60strategy:fail-fast: falsemax-parallel: 1diff --git a/.github/workflows/oc2ov_test.yml b/.github/workflows/oc2ov_test.ymlindex 01e024876..c127ab34d 100644--- a/.github/workflows/oc2ov_test.yml+++ b/.github/workflows/oc2ov_test.yml@@ -30,7 +30,7 @@ jobs:p0-tests:name: P0 Memory Testsruns-on: [self-hosted, linux, x64]- timeout-minutes: 50+ timeout-minutes: 70if: inputs.skip_tests != truesteps:diff --git a/tests/oc2ov_test/tests/p0/test_memory_v2_full_suite.py b/tests/oc2ov_test/tests/p0/test_memory_v2_full_suite.pynew file mode 100644index 000000000..3fdc198eb--- /dev/null+++ b/tests/oc2ov_test/tests/p0/test_memory_v2_full_suite.py@@ -0,0 +1,708 @@+"""+Memory V2 全面端到端测试套件+覆盖所有记忆文件类型:+- preferences (偏好设置) - User scope+- entities (实体信息) - User scope+- events (事件记录) - User scope+- profile (用户画像) - User scope, 单文件+- skills (技能) - Agent scope+- tools (工具) - Agent scope++测试方式:通过 OpenClaw agent 进行对话,然后通过 OV API commit session 触发记忆提取。+OpenClaw Gateway 会自动分配 OV session ID,需要通过对比 sessions 列表来定位。+"""++import json+import os+import subprocess+import time+from pathlib import Path+from typing import Any, Dict, List, Optional, Set, Tuple++import requests+from utils.test_utils import SessionIdManager++SERVER_URL = os.environ.get("SERVER_URL", "http://127.0.0.1:1933")+OPENVIKING_API_KEY = os.environ.get("OPENVIKING_API_KEY", "test-root-api-key")+TASK_POLL_INTERVAL = 5+TASK_POLL_MAX_WAIT = 120+++def _get_api_headers() -> Dict[str, str]:+ return {+ "X-OpenViking-API-Key": OPENVIKING_API_KEY,+ "Content-Type": "application/json",+ }+++def get_viking_data_dir() -> Path:+ """动态获取 viking 数据目录路径(user scope memories)"""+ if os.environ.get("VIKING_DATA_DIR"):+ return (+ Path(os.environ["VIKING_DATA_DIR"])+ / "viking"+ / "default"+ / "user"+ / "default"+ / "memories"+ )++ project_root = Path(__file__).parent.parent.parent.parent.parent+ project_data_dir = (+ project_root / ".openviking_data" / "viking" / "default" / "user" / "default" / "memories"+ )+ if project_data_dir.exists():+ return project_data_dir++ config_candidates = [+ project_root / "ov.conf.temp",+ project_root / "ov.conf",+ Path.home() / ".openviking" / "ov.conf",+ ]+ for config_path in config_candidates:+ if config_path.exists():+ try:+ with open(config_path, "r") as f:+ config = json.load(f)+ workspace = config.get("storage", {}).get("workspace") or config.get(+ "agfs", {}+ ).get("path")+ if workspace:+ return Path(workspace) / "viking" / "default" / "user" / "default" / "memories"+ except Exception:+ pass++ cwd_data_dir = (+ Path.cwd() / ".openviking_data" / "viking" / "default" / "user" / "default" / "memories"+ )+ if cwd_data_dir.exists():+ return cwd_data_dir++ return Path.home() / ".openviking_data" / "viking" / "default" / "user" / "default" / "memories"+++class OpenVikingAPIClient:+ """OV API 客户端"""++ def __init__(self, server_url: str = SERVER_URL, api_key: str = OPENVIKING_API_KEY):+ self.server_url = server_url+ self.api_key = api_key+ self.headers = _get_api_headers()++ def list_session_ids(self) -> Set[str]:+ """获取当前所有 OV session ID 集合"""+ try:+ resp = requests.get(+ f"{self.server_url}/api/v1/sessions",+ headers=self.headers,+ timeout=10,+ )+ if resp.status_code != 200:+ return set()+ sessions = resp.json().get("result", [])+ return {s["session_id"] for s in sessions}+ except Exception:+ return set()++ def find_new_session_id(self, before_ids: Set[str]) -> Optional[str]:+ """对比前后 session 列表,找到新创建的 session ID"""+ after_ids = self.list_session_ids()+ new_ids = after_ids - before_ids+ if new_ids:+ return new_ids.pop()+ return None++ def find_session_with_most_messages(+ self, candidate_ids: Optional[Set[str]] = None+ ) -> Optional[str]:+ """在候选 session 中找到消息数最多的 session"""+ try:+ resp = requests.get(+ f"{self.server_url}/api/v1/sessions",+ headers=self.headers,+ timeout=10,+ )+ if resp.status_code != 200:+ return None+ sessions = resp.json().get("result", [])+ best_id = None+ best_pending = -1+ for s in sessions:+ sid = s["session_id"]+ if candidate_ids and sid not in candidate_ids:+ continue+ try:+ detail_resp = requests.get(+ f"{self.server_url}/api/v1/sessions/{sid}",+ headers=self.headers,+ timeout=5,+ )+ if detail_resp.status_code == 200:+ pending = detail_resp.json().get("result", {}).get("pending_tokens", 0)+ if pending > best_pending:+ best_pending = pending+ best_id = sid+ except Exception:+ continue+ return best_id+ except Exception:+ return None++ def commit_session(self, session_id: str) -> Dict[str, Any]:+ resp = requests.post(+ f"{self.server_url}/api/v1/sessions/{session_id}/commit",+ headers=self.headers,+ timeout=30,+ )+ return {"status_code": resp.status_code, "data": resp.json() if resp.text else {}}++ def get_task(self, task_id: str) -> Dict[str, Any]:+ resp = requests.get(+ f"{self.server_url}/api/v1/tasks/{task_id}",+ headers=self.headers,+ timeout=10,+ )+ return {"status_code": resp.status_code, "data": resp.json() if resp.text else {}}++ def poll_task_until_done(+ self, task_id: str, max_wait: int = TASK_POLL_MAX_WAIT+ ) -> Dict[str, Any]:+ start = time.time()+ while time.time() - start < max_wait:+ task_resp = self.get_task(task_id)+ if task_resp["status_code"] != 200:+ time.sleep(TASK_POLL_INTERVAL)+ continue+ task_data = task_resp["data"].get("result", {})+ status = task_data.get("status", "unknown")+ if status in ("completed", "failed"):+ return task_data+ time.sleep(TASK_POLL_INTERVAL)+ return {"status": "timeout", "task_id": task_id}+++class MemoryV2TestSuite:+ """Memory V2 全面测试套件 - OpenClaw 对话 + OV API commit"""++ def __init__(self):+ self.test_scenarios = self._create_test_scenarios()+ self.viking_data_dir = get_viking_data_dir()+ self.api = OpenVikingAPIClient()++ def _create_test_scenarios(self) -> List[Dict[str, Any]]:+ return [+ {+ "name": "preferences",+ "description": "测试偏好设置记忆",+ "test_message": "我喜欢用Python编程,偏好使用VS Code编辑器,喜欢喝咖啡,特别是美式咖啡",+ "memory_type": "preferences",+ },+ {+ "name": "entities",+ "description": "测试实体信息记忆",+ "test_message": "我叫李明,今年28岁,是一名软件工程师,在字节跳动工作,住在北京海淀区",+ "memory_type": "entities",+ },+ {+ "name": "events",+ "description": "测试事件记录记忆",+ "test_message": "昨天我参加了公司的技术分享会,主题是微服务架构,会议持续了2小时,有50多人参加",+ "memory_type": "events",+ },+ {+ "name": "profile",+ "description": "测试用户画像记忆",+ "test_message": "我是一名技术负责人,有10年开发经验,专注于后端架构设计,喜欢用Python和Go语言",+ "memory_type": "profile",+ },+ {+ "name": "skills",+ "description": "测试技能记忆",+ "test_message": "我擅长使用Docker进行容器化部署,熟练掌握Kubernetes集群管理,有丰富的CI/CD流水线搭建经验",+ "memory_type": "skills",+ },+ {+ "name": "tools",+ "description": "测试工具记忆",+ "test_message": "我经常使用Git进行版本控制,用Jenkins做持续集成,用Prometheus和Grafana监控服务",+ "memory_type": "tools",+ },+ ]++ def run_openclaw_command(self, message: str, session_id: str) -> Dict[str, Any]:+ """执行 openclaw agent 命令"""+ cmd = [+ "openclaw",+ "agent",+ "--session-id",+ session_id,+ "--message",+ message,+ "--json",+ ]+ result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)++ if result.returncode != 0:+ raise Exception(f"OpenClaw command failed: {result.stderr}")++ try:+ return json.loads(result.stdout)+ except json.JSONDecodeError:+ return {"raw_output": result.stdout}++ def _snapshot_all_memory_files(self) -> Dict[str, float]:+ """记录所有记忆目录的文件快照(路径 → mtime)"""+ files: Dict[str, float] = {}++ # user scope: preferences, entities, events, profile+ for md_file in self.viking_data_dir.rglob("*.md"):+ files[str(md_file)] = md_file.stat().st_mtime++ # agent scope: skills, tools+ agent_base_dir = self.viking_data_dir.parent.parent.parent / "agent"+ if agent_base_dir.exists():+ for md_file in agent_base_dir.rglob("*.md"):+ files[str(md_file)] = md_file.stat().st_mtime++ return files++ def check_memory_files(+ self, memory_type: str, before_files: Dict[str, float]+ ) -> Dict[str, Any]:+ """检查 commit 后是否有新增或修改的记忆文件(全目录扫描)"""+ result = {+ "memory_type": memory_type,+ "found": False,+ "new_files": [],+ "modified_files": [],+ "all_files": [],+ }++ # 先检查目标类型目录+ target_result = self._check_target_type(memory_type, before_files)+ result["new_files"].extend(target_result["new_files"])+ result["modified_files"].extend(target_result["modified_files"])+ result["all_files"].extend(target_result["all_files"])++ if target_result["found"]:+ result["found"] = True+ return result++ # 目标目录无变化,扫描全目录查找新增/修改的文件+ print(f" ⚠ {memory_type} 目录无直接变化,扫描全目录...")+ after_files = self._snapshot_all_memory_files()++ for file_str, new_mtime in after_files.items():+ if file_str not in before_files:+ result["found"] = True+ result["new_files"].append(file_str)+ rel_path = self._relative_display(file_str)+ print(f" ✓ 其他目录新增文件: {rel_path}")+ else:+ old_mtime = before_files[file_str]+ if new_mtime > old_mtime:+ result["found"] = True+ result["modified_files"].append(file_str)+ rel_path = self._relative_display(file_str)+ print(f" ✓ 其他目录文件已更新: {rel_path}")++ if not result["found"]:+ print(" ✗ 全目录扫描均无新增或修改的记忆文件")+ return result++ def _relative_display(self, file_str: str) -> str:+ """生成相对路径用于显示"""+ try:+ base = self.viking_data_dir.parent.parent.parent+ return str(Path(file_str).relative_to(base))+ except ValueError:+ return Path(file_str).name++ def _check_target_type(+ self, memory_type: str, before_files: Dict[str, float]+ ) -> Dict[str, Any]:+ """检查目标 memory_type 目录的文件变化"""+ result: Dict[str, Any] = {+ "found": False,+ "new_files": [],+ "modified_files": [],+ "all_files": [],+ }++ if memory_type == "profile":+ profile_file = self.viking_data_dir / "profile.md"+ if profile_file.exists():+ file_str = str(profile_file)+ result["all_files"].append(file_str)+ if file_str not in before_files:+ result["found"] = True+ result["new_files"].append(file_str)+ print(f" ✓ 新增 profile 文件: {profile_file}")+ else:+ old_mtime = before_files[file_str]+ new_mtime = profile_file.stat().st_mtime+ if new_mtime > old_mtime:+ result["found"] = True+ result["modified_files"].append(file_str)+ print(" ✓ profile 文件已更新 (mtime 变化)")+ else:+ print(f" ⚠ profile 文件未变化: {profile_file}")+ else:+ print(f" ✗ profile 文件不存在: {profile_file}")+ return result++ if memory_type in ["skills", "tools"]:+ agent_base_dir = self.viking_data_dir.parent.parent.parent / "agent"+ if not agent_base_dir.exists():+ print(f" ✗ agent 目录不存在: {agent_base_dir}")+ return result++ search_types = [memory_type]+ if memory_type == "tools":+ search_types = ["tools", "skills"]++ for agent_space_dir in sorted(agent_base_dir.iterdir()):+ if not agent_space_dir.is_dir() or agent_space_dir.name.startswith("."):+ continue+ for search_type in search_types:+ memory_dir = agent_space_dir / "memories" / search_type+ if not memory_dir.exists():+ continue++ for md_file in sorted(memory_dir.glob("*.md")):+ file_str = str(md_file)+ result["all_files"].append(file_str)+ if file_str not in before_files:+ result["found"] = True+ result["new_files"].append(file_str)+ print(f" ✓ 新增记忆文件: {md_file.name} (在 {search_type}/ 目录下)")+ else:+ old_mtime = before_files[file_str]+ new_mtime = md_file.stat().st_mtime+ if new_mtime > old_mtime:+ result["found"] = True+ result["modified_files"].append(file_str)+ print(+ f" ✓ 记忆文件已更新: {md_file.name} (在 {search_type}/ 目录下)"+ )++ if not result["found"]:+ if result["all_files"]:+ print(f" ⚠ {memory_type} 目录有文件但均无变化")+ else:+ print(f" ✗ 未找到 {memory_type} 相关记忆文件")+ return result++ # user scope: preferences, entities, events+ memory_dir = self.viking_data_dir / memory_type+ if not memory_dir.exists():+ print(f" ✗ 记忆目录不存在: {memory_dir}")+ return result++ for md_file in sorted(memory_dir.rglob("*.md")):+ file_str = str(md_file)+ result["all_files"].append(file_str)+ if file_str not in before_files:+ result["found"] = True+ result["new_files"].append(file_str)+ rel = md_file.relative_to(memory_dir)+ print(f" ✓ 新增记忆文件: {rel}")+ else:+ old_mtime = before_files[file_str]+ new_mtime = md_file.stat().st_mtime+ if new_mtime > old_mtime:+ result["found"] = True+ result["modified_files"].append(file_str)+ rel = md_file.relative_to(memory_dir)+ print(f" ✓ 记忆文件已更新: {rel}")++ if not result["found"]:+ if result["all_files"]:+ print(f" ⚠ {memory_type} 目录有文件但均无变化")+ else:+ print(f" ✗ 未找到 {memory_type} 记忆文件")+ return result++ def test_single_scenario(self, scenario: Dict[str, Any]) -> Tuple[bool, Dict[str, Any]]:+ """测试单个场景:OpenClaw 对话 → OV API commit → 轮询任务 → 检查记忆文件"""+ print(f"\n{'=' * 60}")+ print(f"测试场景: {scenario['name']} - {scenario['description']}")+ print(f"{'=' * 60}")++ result = {+ "scenario": scenario["name"],+ "description": scenario["description"],+ "memory_type": scenario["memory_type"],+ "steps": {},+ }++ try:+ # 每个场景使用独立的 session ID,确保 OpenClaw 创建新的 OV session+ scenario_session_id = SessionIdManager.generate_session_id(+ prefix=f"memv2_{scenario['name']}"+ )++ # 步骤 1: 记录当前 OV sessions 快照和全目录记忆文件快照+ print("\n[步骤 1/5] 记录快照")+ before_session_ids = self.api.list_session_ids()+ before_memory_files = self._snapshot_all_memory_files()+ print(f" 当前 session 数量: {len(before_session_ids)}")+ print(f" 当前全目录记忆文件数量: {len(before_memory_files)}")+ print(f" 场景 session ID: {scenario_session_id}")+ result["steps"]["snapshot"] = "success"++ # 步骤 2: 通过 OpenClaw 发送消息+ print("\n[步骤 2/5] 通过 OpenClaw 发送消息")+ print(f" 消息: {scenario['test_message'][:50]}...")+ self.run_openclaw_command(scenario["test_message"], scenario_session_id)+ print("✓ OpenClaw 对话完成")+ result["steps"]["openclaw_chat"] = "success"+ time.sleep(2)++ # 步骤 3: 找到 OV session 并 commit+ print("\n[步骤 3/5] 查找 OV session 并 commit")+ ov_session_id = self.api.find_new_session_id(before_session_ids)+ if not ov_session_id:+ print(" ⚠ 未找到新 session,在已有 session 中查找消息最多的")+ ov_session_id = self.api.find_session_with_most_messages()++ if not ov_session_id:+ print(" ✗ 无可用 session")+ result["steps"]["commit"] = "failed"+ result["status"] = "error"+ result["error"] = "未找到 OV session"+ return False, result++ print(f" OV session ID: {ov_session_id}")+ commit_resp = self.api.commit_session(ov_session_id)+ commit_data = commit_resp.get("data", {})+ commit_result = commit_data.get("result", {})+ task_id = commit_result.get("task_id")++ if commit_resp["status_code"] == 200 and task_id:+ print(f"✓ Commit 成功 (task_id: {task_id})")+ result["steps"]["commit"] = "success"+ result["ov_session_id"] = ov_session_id+ result["task_id"] = task_id+ else:+ print(f"✗ Commit 失败: {commit_resp}")+ result["steps"]["commit"] = "failed"+ result["status"] = "failed"+ result["error"] = f"Commit 失败: {commit_resp}"+ return False, result++ # 步骤 4: 轮询任务直到完成+ print(f"\n[步骤 4/5] 等待记忆提取完成 (轮询 task_id: {task_id})")+ task_result = self.api.poll_task_until_done(task_id)+ task_status = task_result.get("status", "unknown")+ memories_extracted = task_result.get("result", {}).get("memories_extracted", {})+ token_usage = task_result.get("result", {}).get("token_usage", {})++ print(f" 任务状态: {task_status}")+ print(f" 提取的记忆: {memories_extracted}")+ if token_usage:+ llm_tokens = token_usage.get("llm", {}).get("total_tokens", 0)+ print(f" LLM token 用量: {llm_tokens}")++ if task_status == "completed":+ total_extracted = (+ sum(len(v) if isinstance(v, list) else v for v in memories_extracted.values())+ if memories_extracted+ else 0+ )+ if total_extracted > 0:+ print(f"✓ 记忆提取成功,共提取 {total_extracted} 条记忆")+ result["steps"]["memory_extraction"] = "success"+ else:+ print("⚠ 记忆提取完成但结果为空 (VLM 可能返回了无效 JSON)")+ result["steps"]["memory_extraction"] = "empty"+ elif task_status == "failed":+ error_msg = task_result.get("error", "unknown error")+ print(f"✗ 记忆提取失败: {error_msg}")+ result["steps"]["memory_extraction"] = "failed"+ else:+ print("✗ 记忆提取超时")+ result["steps"]["memory_extraction"] = "timeout"++ result["task_result"] = task_result++ # 步骤 5: 验证记忆文件变化(硬性断言)+ print("\n[步骤 5/5] 验证记忆文件变化")+ memory_files_result = self.check_memory_files(+ scenario["memory_type"], before_memory_files+ )+ result["memory_files"] = memory_files_result++ if memory_files_result["found"]:+ new_count = len(memory_files_result["new_files"])+ mod_count = len(memory_files_result["modified_files"])+ print(f"✓ 记忆文件有变化,新增 {new_count} 个,更新 {mod_count} 个")+ result["steps"]["memory_files"] = "success"+ else:+ print("✗ 记忆文件验证失败(全目录无新增或修改的文件)")+ result["steps"]["memory_files"] = "failed"++ # 综合判断:记忆提取成功 + 文件有变化 = 通过+ extraction_ok = result["steps"].get("memory_extraction") == "success"+ files_ok = result["steps"].get("memory_files") == "success"++ if extraction_ok and files_ok:+ result["status"] = "passed"+ return True, result+ else:+ reasons = []+ if not extraction_ok:+ reasons.append("记忆提取失败或为空")+ if not files_ok:+ reasons.append("无新增或修改的记忆文件")+ result["status"] = "failed"+ result["error"] = f"{scenario['memory_type']}: {', '.join(reasons)}"+ return False, result++ except Exception as e:+ print(f"\n✗ 场景测试执行失败: {e}")+ result["status"] = "error"+ result["error"] = str(e)+ return False, result++ def run_full_test_suite(self) -> Dict[str, Any]:+ """运行完整的测试套件"""+ print("\n" + "=" * 60)+ print("Memory V2 全面端到端测试套件")+ print(" 对话方式: OpenClaw agent")+ print(" Commit 方式: OV API")+ print("=" * 60)+ print(f"\n测试场景数量: {len(self.test_scenarios)}")+ print(f"记忆类型覆盖: {', '.join([s['memory_type'] for s in self.test_scenarios])}")+ print(f"数据目录: {self.viking_data_dir}")+ print(f"OV Server: {SERVER_URL}")++ results = {+ "total_scenarios": len(self.test_scenarios),+ "scenarios": [],+ "summary": {"passed": 0, "failed": 0, "error": 0},+ }++ for scenario in self.test_scenarios:+ passed, result = self.test_single_scenario(scenario)+ results["scenarios"].append(result)++ if result["status"] == "passed":+ results["summary"]["passed"] += 1+ elif result["status"] == "failed":+ results["summary"]["failed"] += 1+ else:+ results["summary"]["error"] += 1++ print("\n" + "=" * 60)+ print("测试总结")+ print("=" * 60)+ print(f"总场景数: {results['total_scenarios']}")+ print(f"通过: {results['summary']['passed']}")+ print(f"失败: {results['summary']['failed']}")+ print(f"错误: {results['summary']['error']}")++ pass_rate = results["summary"]["passed"] / results["total_scenarios"]+ print(f"\n通过率: {pass_rate * 100:.1f}%")++ if pass_rate >= 0.7:+ print("\n✓ Memory V2 测试套件通过!")+ else:+ print("\n✗ Memory V2 测试套件未通过")++ return results+++SCENARIO_MAP = {+ "preferences": {+ "name": "preferences",+ "description": "测试偏好设置记忆",+ "test_message": "我喜欢用Python编程,偏好使用VS Code编辑器,喜欢喝咖啡,特别是美式咖啡",+ "memory_type": "preferences",+ },+ "entities": {+ "name": "entities",+ "description": "测试实体信息记忆",+ "test_message": "我叫李明,今年28岁,是一名软件工程师,在字节跳动工作,住在北京海淀区",+ "memory_type": "entities",+ },+ "events": {+ "name": "events",+ "description": "测试事件记录记忆",+ "test_message": "昨天我参加了公司的技术分享会,主题是微服务架构,会议持续了2小时,有50多人参加",+ "memory_type": "events",+ },+ "profile": {+ "name": "profile",+ "description": "测试用户画像记忆",+ "test_message": "我是一名技术负责人,有10年开发经验,专注于后端架构设计,喜欢用Python和Go语言",+ "memory_type": "profile",+ },+ "skills": {+ "name": "skills",+ "description": "测试技能记忆",+ "test_message": "我擅长使用Docker进行容器化部署,熟练掌握Kubernetes集群管理,有丰富的CI/CD流水线搭建经验",+ "memory_type": "skills",+ },+ "tools": {+ "name": "tools",+ "description": "测试工具记忆",+ "test_message": "我经常使用Git进行版本控制,用Jenkins做持续集成,用Prometheus和Grafana监控服务",+ "memory_type": "tools",+ },+}+++def _run_single_memory_test(scenario_key: str):+ tester = MemoryV2TestSuite()+ scenario = SCENARIO_MAP[scenario_key]+ passed, result = tester.test_single_scenario(scenario)+ assert passed, f"{scenario['name']} 测试失败: {result.get('error', '未知错误')}"+++def test_memory_v2_preferences():+ _run_single_memory_test("preferences")+++def test_memory_v2_entities():+ _run_single_memory_test("entities")+++def test_memory_v2_events():+ _run_single_memory_test("events")+++def test_memory_v2_profile():+ _run_single_memory_test("profile")+++def test_memory_v2_skills():+ _run_single_memory_test("skills")+++def test_memory_v2_tools():+ _run_single_memory_test("tools")+++if __name__ == "__main__":+ """直接运行测试"""+ tester = MemoryV2TestSuite()+ results = tester.run_full_test_suite()++ print("\n" + "=" * 60)+ print("详细测试报告")+ print("=" * 60)++ for scenario in results["scenarios"]:+ status_icon = "✓" if scenario["status"] == "passed" else "✗"+ print(f"\n{status_icon} {scenario['scenario']}: {scenario['status']}")+ steps = scenario.get("steps", {})+ for step_name, step_status in steps.items():+ step_icon = (+ "✓" if step_status in ("success",) else "⚠" if step_status == "empty" else "✗"+ )+ print(f" {step_icon} {step_name}: {step_status}")+ if scenario.get("error"):+ print(f" 错误: {scenario['error']}")++ exit(0 if results["summary"]["passed"] >= results["total_scenarios"] * 0.7 else 1)diff --git a/tests/oc2ov_test/upgrade_openviking.sh b/tests/oc2ov_test/upgrade_openviking.shindex c1880408e..8a93d117d 100644--- a/tests/oc2ov_test/upgrade_openviking.sh+++ b/tests/oc2ov_test/upgrade_openviking.sh@@ -253,6 +253,26 @@ if [ "$BUILD_SUCCESS" = false ]; thenexit 1fi+log "[7.5/8] Installing OpenClaw openviking plugin dependencies..."+PLUGIN_DIR="$PROJECT_DIR/examples/openclaw-plugin"++if [ -d "$PLUGIN_DIR" ]; then+ log "Plugin directory: $PLUGIN_DIR"+ if command -v npm &> /dev/null; then+ cd "$PLUGIN_DIR"+ if npm install --omit=dev 2>&1 | tee -a "$LOG_FILE"; then+ log "✅ Plugin npm dependencies installed"+ else+ log "⚠️ WARNING: npm install --omit=dev failed, continuing anyway"+ fi+ cd "$PROJECT_DIR"+ else+ log "⚠️ npm command not found, skipping plugin dependency install"+ fi+else+ log "⚠️ Plugin directory not found: $PLUGIN_DIR, skipping"+fi+log "[8/8] Restarting OpenClaw service to load latest OpenViking..."# Load OpenClaw environment variables@@ -299,6 +319,67 @@ elselog "Actual path: $OV_PATH"fi+# Step 3.5: Ensure OpenViking server is running on port 1933+log "Step 3.5: Ensuring OpenViking server is running..."+OV_SERVER_RUNNING=false++if command -v ss &> /dev/null; then+ if ss -tuln 2>/dev/null | grep -q ":1933 "; then+ OV_SERVER_RUNNING=true+ log "✅ OpenViking server is already listening on port 1933"+ fi+elif command -v netstat &> /dev/null; then+ if netstat -tuln 2>/dev/null | grep -q ":1933 "; then+ OV_SERVER_RUNNING=true+ log "✅ OpenViking server is already listening on port 1933"+ fi+fi++if [ "$OV_SERVER_RUNNING" = false ]; then+ log "OpenViking server not running, starting it..."++ pkill -f "openviking.server.bootstrap" 2>/dev/null || true+ sleep 1++ OV_CONF=""+ for conf_candidate in "$PROJECT_DIR/ov.conf.temp" "$PROJECT_DIR/ov.conf" "$HOME/.openviking/ov.conf"; do+ if [ -f "$conf_candidate" ]; then+ OV_CONF="$conf_candidate"+ break+ fi+ done++ if [ -n "$OV_CONF" ]; then+ nohup python -m openviking.server.bootstrap --config "$OV_CONF" > /tmp/openviking.log 2>&1 &+ log "Started OpenViking server with config: $OV_CONF (PID: $!)"+ else+ nohup python -m openviking.server.bootstrap > /tmp/openviking.log 2>&1 &+ log "Started OpenViking server without explicit config (PID: $!)"+ fi++ for i in $(seq 1 15); do+ sleep 2+ if command -v ss &> /dev/null; then+ if ss -tuln 2>/dev/null | grep -q ":1933 "; then+ OV_SERVER_RUNNING=true+ log "✅ OpenViking server is listening on port 1933 (after ${i}x2s)"+ break+ fi+ elif command -v netstat &> /dev/null; then+ if netstat -tuln 2>/dev/null | grep -q ":1933 "; then+ OV_SERVER_RUNNING=true+ log "✅ OpenViking server is listening on port 1933 (after ${i}x2s)"+ break+ fi+ fi+ done++ if [ "$OV_SERVER_RUNNING" = false ]; then+ log "⚠️ WARNING: OpenViking server failed to start on port 1933"+ log " Check logs: tail -50 /tmp/openviking.log"+ fi+fi+# Step 4: Start OpenClaw gatewayRESTART_SUCCESS=falsefor i in $(seq 1 $MAX_RETRIES); dodiff --git a/tests/oc2ov_test/utils/openclaw_cli_client.py b/tests/oc2ov_test/utils/openclaw_cli_client.pyindex 228383e8f..d23818f80 100644--- a/tests/oc2ov_test/utils/openclaw_cli_client.py+++ b/tests/oc2ov_test/utils/openclaw_cli_client.py@@ -20,7 +20,7 @@ class OpenClawCLIClient:初始化客户端"""self.session_id = session_id or "test_session_default"- self.timeout = 300+ self.timeout = 180def send_message(self, message: str, session_id: Optional[str] = None, agent_id: Optional[str] = Nonediff --git a/tests/oc2ov_test/utils/test_utils.py b/tests/oc2ov_test/utils/test_utils.pyindex f3facd4ba..2c1bacb1f 100644--- a/tests/oc2ov_test/utils/test_utils.py+++ b/tests/oc2ov_test/utils/test_utils.py@@ -3,23 +3,41 @@提供 Session ID 管理、智能等待、重试机制、测试数据管理等功能"""-import time-import uuidimport functoolsimport logging+import os+import time+import uuidfrom dataclasses import dataclass, field-from typing import Any, Callable, Dict, List, Optional, TypeVar, Unionfrom datetime import datetime+from typing import Any, Callable, Dict, List, Optional, TypeVar, Unionlogger = logging.getLogger(__name__)T = TypeVar("T")+# 固定的 Session ID,用于 CI 环境避免 session 爆满+# 使用 UUID 格式,确保 OpenClaw 直接使用,不做 SHA256 转换+FIXED_SESSION_ID = "00000000-0000-0000-0000-000000000001"+++def _is_ci_environment() -> bool:+ """检测是否在 CI 环境中运行"""+ return bool(+ os.environ.get("CI")+ or os.environ.get("GITHUB_ACTIONS")+ or os.environ.get("GITLAB_CI")+ or os.environ.get("TRAVIS")+ or os.environ.get("CIRCLECI")+ or os.environ.get("JENKINS_URL")+ )+class SessionIdManager:"""Session ID 管理器自动生成唯一的 session_id,支持前缀和后缀+ 在 CI 环境中使用固定的 session ID,避免 session 爆满"""_instance = None@@ -37,25 +55,28 @@ class SessionIdManager:include_uuid: bool = True,) -> str:"""- 生成唯一的 session_id+ 生成唯一的 session_id (UUID 格式)++ OpenClaw 会将非 UUID 格式的 session ID 转换为 SHA256 哈希,+ 使用 UUID 格式可以确保 OpenClaw 直接使用,不做转换。Args:- prefix: session_id 前缀- include_timestamp: 是否包含时间戳- include_uuid: 是否包含 UUID+ prefix: session_id 前缀 (仅用于日志,不影响 UUID 格式)+ include_timestamp: 是否包含时间戳 (已忽略,保持接口兼容)+ include_uuid: 是否包含 UUID (已忽略,始终使用 UUID)Returns:- str: 唯一的 session_id+ str: UUID 格式的 session_id"""- parts = [prefix]-- if include_timestamp:- parts.append(datetime.now().strftime("%Y%m%d_%H%M%S"))+ # 在 CI 环境中使用固定的 session ID+ if _is_ci_environment():+ logger.info(f"CI 环境检测到,使用固定 Session ID: {FIXED_SESSION_ID}")+ return FIXED_SESSION_ID- if include_uuid:- parts.append(uuid.uuid4().hex[:8])-- return "_".join(parts)+ # 生成 UUID 格式的 session ID+ session_id = str(uuid.uuid4())+ logger.info(f"生成 Session ID: {session_id} (prefix: {prefix})")+ return session_id@staticmethoddef generate_test_class_session_id(test_class_name: str) -> str:@@ -68,6 +89,11 @@ class SessionIdManager:Returns:str: 唯一的 session_id"""+ # 在 CI 环境中使用固定的 session ID+ if _is_ci_environment():+ logger.info(f"CI 环境检测到,使用固定 Session ID: {FIXED_SESSION_ID}")+ return FIXED_SESSION_ID+return f"test_{test_class_name}_{uuid.uuid4().hex[:8]}"@staticmethod@@ -82,6 +108,11 @@ class SessionIdManager:Returns:str: 唯一的 session_id"""+ # 在 CI 环境中使用固定的 session ID+ if _is_ci_environment():+ logger.info(f"CI 环境检测到,使用固定 Session ID: {FIXED_SESSION_ID}")+ return FIXED_SESSION_ID+return f"test_{test_class_name}_{test_method_name}_{uuid.uuid4().hex[:8]}"def register_session(@@ -199,7 +230,9 @@ class SmartWaiter:try:if condition():elapsed = time.time() - start_time- logger.info(f"✅ 条件满足: {message} (耗时: {elapsed:.2f}秒, 尝试次数: {attempt})")+ logger.info(+ f"✅ 条件满足: {message} (耗时: {elapsed:.2f}秒, 尝试次数: {attempt})"+ )return Trueexcept Exception as e:logger.warning(f"条件检查异常 (尝试 {attempt}): {e}")@@ -326,6 +359,7 @@ class RetryManager:Returns:Callable: 装饰器函数"""+def decorator(func: Callable[..., T]) -> Callable[..., T]:@functools.wraps(func)def wrapper(*args, **kwargs) -> T:@@ -350,14 +384,13 @@ class RetryManager:if self.exponential_backoff:delay = min(delay * self.backoff_factor, self.max_delay)else:- logger.error(- f"重试次数耗尽: {func.__name__} - {e}"- )+ logger.error(f"重试次数耗尽: {func.__name__} - {e}")raiseraise last_exceptionreturn wrapper+return decoratordef retry_on_result(@@ -399,13 +432,12 @@ class RetryManager:if self.exponential_backoff:delay = min(delay * self.backoff_factor, self.max_delay)else:- logger.warning(- f"重试次数耗尽,返回最后结果: {func.__name__}"- )+ logger.warning(f"重试次数耗尽,返回最后结果: {func.__name__}")return last_resultreturn wrapper+return decoratordef execute_with_retry(@@ -427,6 +459,7 @@ class RetryManager:Returns:T: 函数返回值"""+@self.retry_on_exception(exceptions)def wrapped():return func(*args, **kwargs)@@ -440,6 +473,7 @@ class TestData:测试数据类用于管理测试数据"""+name: strdescription: str = ""input_data: Dict[str, Any] = field(default_factory=dict)@@ -500,10 +534,7 @@ class TestDataManager:Returns:List[TestData]: 匹配的测试数据列表"""- return [- data for data in self._data_registry.values()- if tag in data.tags- ]+ return [data for data in self._data_registry.values() if tag in data.tags]def validate_data(self, data: TestData) -> bool:"""