""" 测试 01: 基础功能测试 测试 WebSocket 连接、会话建立、闲聊和飞控指令识别 """ import asyncio import json import sys import time from pathlib import Path import websockets import numpy as np import sounddevice as sd # 添加测试目录到路径 sys.path.insert(0, str(Path(__file__).parent)) from test_config import SERVER_URL, AUTH_TOKEN, DEVICE_ID, TEST_CASES from test_utils import play_audio, print_json, calculate_metrics, print_metrics, TestResult async def test_basic(): """基础功能测试""" result = TestResult("基础功能测试") print(f"\n🔗 连接到: {SERVER_URL}") try: async with websockets.connect(SERVER_URL) as ws: print("✅ WebSocket 连接成功\n") # ========== 1. 建立会话 ========== print("=" * 60) print(" 测试 1: 建立会话") print("=" * 60) session_start = { "type": "session.start", "proto_version": "1.0", "transport_profile": "text_uplink", "session_id": "test-session-001", "auth_token": AUTH_TOKEN, "client": { "device_id": DEVICE_ID, "locale": "zh-CN", "capabilities": { "playback_sample_rate_hz": 24000, "prefer_tts_codec": "pcm_s16le" }, "protocol": { "dialog_result": "cloud_voice_dialog_v1" }, } } await ws.send(json.dumps(session_start, ensure_ascii=False)) print("→ session.start") # 接收 session.ready ready_msg = await ws.recv() ready_data = json.loads(ready_msg) if ready_data.get("type") != "session.ready": result.add_error(f"期望 session.ready，收到: {ready_data.get('type')}") result.print_report() return result print("← session.ready") print(f" 服务端能力: {json.dumps(ready_data.get('server_caps', {}), ensure_ascii=False)}") print("✅ 会话建立成功\n") # ========== 2. 闲聊测试 ========== print("=" * 60) print(" 测试 2: 闲聊对话") print("=" * 60) test_text = TEST_CASES["chitchat"][0] turn_result = await test_turn(ws, test_text, expect_routing="chitchat") if not turn_result["success"]: result.add_error(f"闲聊测试失败: {turn_result['error']}") else: print("✅ 闲聊测试通过\n") # ========== 3. 飞控指令测试 ========== print("=" * 60) print(" 测试 3: 飞控指令识别") print("=" * 60) test_text = "起飞然后在前方十米悬停" turn_result = await test_turn(ws, test_text, expect_routing="flight_intent") if not turn_result["success"]: result.add_error(f"飞控指令测试失败: {turn_result['error']}") else: print("✅ 飞控指令测试通过\n") # 验证飞控意图结构 flight_intent = turn_result.get("flight_intent") if flight_intent: print(" 📋 飞控意图详情:") print(f" Summary: {flight_intent.get('summary')}") print(f" Actions: {json.dumps(flight_intent.get('actions', []), ensure_ascii=False, indent=6)}") # ========== 4. 返航测试 ========== print("\n" + "=" * 60) print(" 测试 4: 返航指令") print("=" * 60) turn_result = await test_turn(ws, "返航", expect_routing="flight_intent") if turn_result["success"]: actions = turn_result.get("flight_intent", {}).get("actions", []) has_return_home = any(a.get("type") == "return_home" for a in actions) if has_return_home: print("✅ 返航指令识别正确\n") else: result.add_error("返航指令未识别到 return_home 动作") # ========== 5. 结束会话 ========== print("=" * 60) print(" 测试 5: 结束会话") print("=" * 60) session_end = { "type": "session.end", "proto_version": "1.0", "session_id": "test-session-001" } await ws.send(json.dumps(session_end)) print("→ session.end") print("✅ 会话已结束\n") result.mark_success() result.add_detail("test_cases", "闲聊 + 飞控 + 返航") except Exception as e: result.add_error(f"测试异常: {e}") import traceback traceback.print_exc() return result.finalize().print_report() async def test_turn(ws, text: str, expect_routing: str) -> dict: """ 测试单轮对话 Returns: dict: { "success": bool, "routing": str, "flight_intent": dict or None, "chat_reply": str or None, "audio_data": np.ndarray or None, "metrics": dict, "error": str or None } """ print(f"\n 输入: {text}") print(f" 期望路由: {expect_routing}") turn_id = f"turn-{int(time.time())}" audio_chunks = [] dialog_result = None metrics = {} t_start = time.time() try: # 发送 turn.text turn_text = { "type": "turn.text", "proto_version": "1.0", "transport_profile": "text_uplink", "turn_id": turn_id, "text": text, "is_final": True, "source": "device_stt" } await ws.send(json.dumps(turn_text, ensure_ascii=False)) print(f" → turn.text (turn_id={turn_id[:8]}...)") # 接收响应 while True: msg = await asyncio.wait_for(ws.recv(), timeout=60) if isinstance(msg, bytes): # 音频数据 audio_chunks.append(msg) else: data = json.loads(msg) msg_type = data.get("type") if msg_type == "dialog_result": dialog_result = data routing = data.get("routing") print(f" ← dialog_result (routing={routing})") if routing == "flight_intent": intent = data.get("flight_intent") if intent: print(f" Summary: {intent.get('summary')}") elif routing == "chitchat": reply = data.get("chat_reply") print(f" Reply: {reply}") elif msg_type == "llm.text_delta": # 流式 LLM 增量；旧测试仅忽略，仍以 dialog_result 为准 if data.get("done"): print(f" ← llm.text_delta (done)") elif msg_type == "tts_audio_chunk": seq = data.get("seq") is_final = data.get("is_final") if seq == 0: print(f" ← TTS 音频流开始") if is_final: print(f" ← TTS 音频流结束 (seq={seq})") elif msg_type == "turn.complete": metrics = data.get("metrics", {}) print(f" ← turn.complete") print(f" LLM: {metrics.get('llm_ms')}ms") print(f" TTS: {metrics.get('tts_first_byte_ms')}ms") break elif msg_type == "error": print(f" ❌ 错误: {data.get('code')} - {data.get('message')}") return { "success": False, "error": f"{data.get('code')}: {data.get('message')}" } t_end = time.time() total_time = t_end - t_start # 拼接音频 audio_data = None if audio_chunks: full_pcm = b"".join(audio_chunks) audio_data = np.frombuffer(full_pcm, dtype=np.int16) audio_length = len(audio_data) / 24000 print(f"\n 📊 生成音频: {audio_length:.2f}s ({len(audio_data)} samples)") # 验证路由类型 actual_routing = dialog_result.get("routing") if dialog_result else None if actual_routing != expect_routing: return { "success": False, "error": f"路由不匹配: 期望 {expect_routing}, 实际 {actual_routing}" } # 播放音频 if audio_data is not None and len(audio_data) > 0: print(f"\n 🔊 播放 TTS 音频...") play_audio(audio_data, sample_rate=24000) # 计算指标 perf_metrics = calculate_metrics( llm_ms=metrics.get("llm_ms"), tts_ms=metrics.get("tts_first_byte_ms"), audio_length_s=len(audio_data) / 24000 if audio_data is not None else 0, total_time_s=total_time ) print_metrics(perf_metrics) return { "success": True, "routing": actual_routing, "flight_intent": dialog_result.get("flight_intent") if dialog_result else None, "chat_reply": dialog_result.get("chat_reply") if dialog_result else None, "audio_data": audio_data, "metrics": perf_metrics } except asyncio.TimeoutError: return { "success": False, "error": "响应超时 (60s)" } except Exception as e: return { "success": False, "error": str(e) } async def main(): print("=" * 60) print(" 测试 01: 基础功能测试") print("=" * 60) success = await test_basic() print("\n" + "=" * 60) if success: print(" ✅ 测试通过！") else: print(" ❌ 测试失败！") print("=" * 60) sys.exit(0 if success else 1) if __name__ == "__main__": asyncio.run(main())