services(cartesia): load voices on startup

CHANGELOG.md

@@ -11,6 +11,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Fixed an issue with Deepgram TTS that was introduced in the previous release.
 
+### Performance
+
+- Load Cartesia voice on startup.
+
 ## [0.0.25] - 2024-05-31
 
 ### Added

src/pipecat/services/cartesia.py

@@ -6,10 +6,9 @@
 
 from cartesia.tts import AsyncCartesiaTTS
 
-import time
 from typing import AsyncGenerator
 
-from pipecat.frames.frames import AudioRawFrame, ErrorFrame, Frame
+from pipecat.frames.frames import AudioRawFrame, Frame
 from pipecat.services.ai_services import TTSService
 
 from loguru import logger
@@ -28,18 +27,18 @@ class CartesiaTTSService(TTSService):
         self._api_key = api_key
         self._voice_name = voice_name
 
-        self._client = None
+        try:
+            self._client = AsyncCartesiaTTS(api_key=self._api_key)
+            voices = self._client.get_voices()
+            voice_id = voices[self._voice_name]["id"]
+            self._voice = self._client.get_voice_embedding(voice_id=voice_id)
+        except Exception as e:
+            logger.error(f"Cartesia initialization error: {e}")
 
     async def run_tts(self, text: str) -> AsyncGenerator[Frame, None]:
         logger.debug(f"Transcribing text: [{text}]")
 
         try:
-            if self._client is None:
-                self._client = AsyncCartesiaTTS(api_key=self._api_key)
-                voices = self._client.get_voices()
-                self._voice_id = voices[self._voice_name]["id"]
-                self._voice = self._client.get_voice_embedding(voice_id=self._voice_id)
-
             chunk_generator = await self._client.generate(
                 transcript=text, voice=self._voice, stream=True,
                 model_id="upbeat-moon", data_rtype='array', output_format='pcm_16000',
@@ -49,8 +48,6 @@ class CartesiaTTSService(TTSService):
             )
 
             async for chunk in chunk_generator:
-                # print(f"")
-                frame = AudioRawFrame(chunk['audio'], 16000, 1)
-                yield frame
+                yield AudioRawFrame(chunk['audio'], 16000, 1)
         except Exception as e:
-            logger.error(f"Exception {e}")
+            logger.error(f"Cartesia error: {e}")