feat/ 添加app-expo三种环境切换，待测试 调整tts

api/.env.example

@@ -65,6 +65,16 @@ TENCENT_SECRET_ID=your_tencent_asr_secret_id
 TENCENT_SECRET_KEY=your_tencent_asr_secret_key
 # TENCENT_ASR_APP_ID=
 
+# =============================================================================
+# TTS (openai | tencent)
+# =============================================================================
+TTS_PROVIDER=tencent
+# 仅 TTS_PROVIDER=openai 时需要
+# OPENAI_API_KEY=your_openai_api_key
+# 仅 TTS_PROVIDER=tencent 时生效，与 ASR 共用 TENCENT_SECRET_ID / TENCENT_SECRET_KEY
+# 音色 ID 见 https://cloud.tencent.com/document/product/1073/92668
+TTS_VOICE_TYPE=603004
+
 # =============================================================================
 # WeChat Pay
 # =============================================================================

api/app/adapters/tts/tencent_tts.py

@@ -0,0 +1,79 @@
+"""Tencent Cloud TTS adapter — implements TTSProvider port.
+
+API: https://cloud.tencent.com/document/product/1073/37995
+"""
+
+import asyncio
+import base64
+import uuid
+
+from app.core.logging import get_logger
+
+logger = get_logger(__name__)
+
+
+class TencentTTSProvider:
+    def __init__(
+        self,
+        secret_id: str,
+        secret_key: str,
+        voice_type: int = 603004,
+        codec: str = "mp3",
+        sample_rate: int = 16000,
+    ):
+        self._secret_id = secret_id
+        self._secret_key = secret_key
+        self._voice_type = voice_type
+        self._codec = codec
+        self._sample_rate = sample_rate
+        self._client = None
+
+    def _get_client(self):
+        if self._client is not None:
+            return self._client
+        try:
+            from tencentcloud.common import credential
+            from tencentcloud.common.profile.client_profile import ClientProfile
+            from tencentcloud.common.profile.http_profile import HttpProfile
+            from tencentcloud.tts.v20190823 import tts_client
+
+            cred = credential.Credential(self._secret_id, self._secret_key)
+            http_profile = HttpProfile()
+            http_profile.endpoint = "tts.tencentcloudapi.com"
+            client_profile = ClientProfile()
+            client_profile.httpProfile = http_profile
+            self._client = tts_client.TtsClient(cred, "", client_profile)
+            return self._client
+        except Exception as e:
+            logger.error("Tencent TTS client init failed: %s", e)
+            return None
+
+    def _synthesize_sync(self, text: str) -> bytes:
+        """Sync synthesis (run in executor)."""
+        client = self._get_client()
+        if not client:
+            return b""
+        from tencentcloud.tts.v20190823 import models
+
+        req = models.TextToVoiceRequest()
+        req.Text = text[:500]  # 中文约150字，英文约500字母，保守截断
+        req.SessionId = f"tts-{uuid.uuid4().hex}"
+        req.VoiceType = self._voice_type
+        req.Codec = self._codec
+        req.SampleRate = self._sample_rate
+        req.PrimaryLanguage = 1  # 1=中文
+
+        resp = client.TextToVoice(req)
+        if resp.Audio:
+            return base64.b64decode(resp.Audio)
+        return b""
+
+    async def synthesize(self, text: str, voice: str = "alloy") -> bytes:
+        """Convert text to speech. Returns mp3 bytes."""
+        if not text or not self._secret_id or not self._secret_key:
+            return b""
+        try:
+            return await asyncio.to_thread(self._synthesize_sync, text)
+        except Exception as e:
+            logger.error("Tencent TTS synthesize failed: %s", e)
+            return b""

api/app/core/config.py

@@ -59,8 +59,10 @@ class Settings(BaseSettings):
     tencent_secret_key: str = ""
     tencent_asr_app_id: str = ""
 
-    # ── OpenAI (TTS) ─────────────────────────────────────────
+    # ── TTS (openai | tencent) ───────────────────────────────
+    tts_provider: str = "tencent"
     openai_api_key: str = ""
+    tts_voice_type: int = 603004  # Tencent 音色 ID，见 https://cloud.tencent.com/document/product/1073/92668
 
     # ── WeChat Pay ───────────────────────────────────────────
     wechat_pay_app_id: str = ""

api/app/core/dependencies.py

@@ -60,6 +60,15 @@ def get_llm_provider() -> LLMProvider:
 
 @lru_cache
 def get_tts_provider() -> TTSProvider:
+    if settings.tts_provider == "tencent":
+        from app.adapters.tts.tencent_tts import TencentTTSProvider
+
+        return TencentTTSProvider(
+            secret_id=settings.tencent_secret_id,
+            secret_key=settings.tencent_secret_key,
+            voice_type=settings.tts_voice_type,
+            codec="mp3",
+        )
     from app.adapters.tts.openai_tts import OpenAITTSProvider
 
     return OpenAITTSProvider(api_key=settings.openai_api_key)

api/app/features/conversation/ws/pipeline.py

@@ -26,11 +26,41 @@ from app.features.conversation.ws.profile_collector import (
     get_missing_profile_fields,
 )
 from app.features.user.models import User
-from app.core.dependencies import get_asr_provider
+from app.core.dependencies import get_asr_provider, get_tts_provider
 from app.features.memoir.state_service import get_or_create_state
 
 logger = get_logger(__name__)
 
+
+async def _send_tts_audio(conversation_id: str, text: str) -> None:
+    """Synthesize text to speech and send TTS_AUDIO if successful."""
+    try:
+        tts = get_tts_provider()
+        audio_bytes = await tts.synthesize(text)
+        if not audio_bytes:
+            logger.warning(
+                "TTS skipped: synthesize returned empty. Check TTS config in .env"
+            )
+            return
+        await manager.send_message(conversation_id, {
+            "type": MessageType.TTS_AUDIO,
+            "conversation_id": conversation_id,
+            "data": {
+                "audio_base64": base64.b64encode(audio_bytes).decode("utf-8"),
+                "format": "mp3",
+            },
+            "timestamp": datetime.now(timezone.utc).isoformat(),
+        })
+    except Exception as e:
+        err_str = str(e)
+        if "PkgExhausted" in err_str:
+            logger.warning(
+                "TTS skipped: 腾讯云语音合成资源包已用尽，请在控制台购买或开通后付费: %s",
+                err_str[:100],
+            )
+        else:
+            logger.error("TTS synthesize failed: %s", e)
+
 # ── Agent 实例（从 ConnectionManager 移出） ─────────────────────
 conversation_agent = ConversationAgent()
 memory_agent = MemoryAgent()
@@ -447,6 +477,7 @@ async def process_user_message(
                         "data": {"text": response_text, "index": i, "total": len(responses)},
                         "timestamp": datetime.now(timezone.utc).isoformat(),
                     })
+                    await _send_tts_audio(conversation_id, response_text)
                     if i < len(responses) - 1:
                         await asyncio.sleep(0.5)
                 return
@@ -498,6 +529,7 @@ async def process_user_message(
                 "data": {"text": response_text, "index": i, "total": len(responses)},
                 "timestamp": datetime.now(timezone.utc).isoformat(),
             })
+            await _send_tts_audio(conversation_id, response_text)
             if i < len(responses) - 1:
                 await asyncio.sleep(0.5)