feat: implement VoiceService with multi-model speech-to-text and text-to-speech capabilities

ai-patient-onboarding/app/services/voice_service.py

@@ -121,8 +121,8 @@ class VoiceService:
             and lang_prefix in ["ta", "hi", "te", "kn", "bn", "mr", "gu", "ml", "pa", "od", "en"]
         ):
             try:
-                speaker = "anushka" if lang_prefix in ["ta", "hi", "te", "kn"] else "karun"
-                async with httpx2.AsyncClient(timeout=3.0) as client:
+                speaker = "kavya" if lang_prefix in ["ta", "hi", "te", "kn"] else "gokul"
+                async with httpx2.AsyncClient(timeout=5.0) as client:
                     headers = {
                         "api-subscription-key": self.sarvam_key,
                         "Content-Type": "application/json",
@@ -131,7 +131,7 @@ class VoiceService:
                         "inputs": [cleaned_text],
                         "target_language_code": language_code,
                         "speaker": speaker,
-                        "model": "bulbul:v2",
+                        "model": "bulbul:v3",
                     }
                     res = await client.post(
                         "https://api.sarvam.ai/text-to-speech",
@@ -153,7 +153,7 @@ class VoiceService:
 
         # 2. Use OpenAI TTS
         if self.openai_key and self.openai_key.strip() not in ["your_openai_api_key", "your_openai_key", ""]:
-            client = AsyncOpenAI(api_key=self.openai_key, timeout=1.5)
+            client = AsyncOpenAI(api_key=self.openai_key, timeout=5.0)
             try:
                 response = await client.audio.speech.create(
                     model="tts-1",