use gpt-4o-mini-tts insteas of n8n

self_hosted_env/docker-compose.yml

@@ -61,16 +61,18 @@ services:
       - minio
 
   voice-agent:
-    build: ./always   # path to your Dockerfile folder
+    build: ./voice_agent    # path to your Dockerfile folder
     container_name: voice-agent
     ports:
       - "8000:8000"  # Expose the FastAPI server
-    restart: on-failure
+    restart: always
     environment:
       MINIO_ENDPOINT: "http://minio:9000"
       MINIO_ACCESS_KEY: "${MINIO_ROOT_USER}"
       MINIO_SECRET_KEY: "${MINIO_ROOT_PASSWORD}"
       N8N_WEBHOOK_URL: "${N8N_WEBHOOK_URL}"
+      OPENAI_API_KEY: "${OPENAI_API_KEY}"
+      MINIO_BUCKET: "${MINIO_BUCKET}"
     volumes:
       - ./uploads:/app/uploads
     depends_on:

self_hosted_env/voice_agent/core/config.py

 import os
 from dataclasses import dataclass
+from dotenv import load_dotenv
+
+# Load environment variables from .env file
+load_dotenv()
 
 # Configuration Management
 
@@ -10,6 +14,7 @@ class AppConfig:
     minio_secret_key: str
     minio_bucket: str
     n8n_webhook_url: str
+    openai_api_key: str
 
     @classmethod
     def from_env(cls) -> 'AppConfig':
@@ -17,6 +22,7 @@ class AppConfig:
             minio_endpoint=os.getenv("MINIO_ENDPOINT", "http://minio:9000"),
             minio_access_key=os.getenv("MINIO_ACCESS_KEY"),
             minio_secret_key=os.getenv("MINIO_SECRET_KEY"),
-            minio_bucket="coversation",
-            n8n_webhook_url=os.getenv("N8N_WEBHOOK_URL")
+            minio_bucket=os.getenv("MINIO_BUCKET"),
+            n8n_webhook_url=os.getenv("N8N_WEBHOOK_URL"),
+            openai_api_key=os.getenv("OPENAI_API_KEY")
         )
\ No newline at end of file

self_hosted_env/voice_agent/main.py

@@ -33,11 +33,10 @@ class DIContainer:
         self.response_manager = ResponseManager()
         self.audio_service = AudioService(self.storage_repo, self.config.minio_bucket)
         self.chat_service = ChatService(self.storage_repo, self.webhook_client, self.response_manager, self.config)
-        self.webhook_service = WebhookService(self.response_manager, self.audio_service)
+        self.webhook_service = WebhookService(self.response_manager, self.storage_repo, self.config.minio_bucket)
         self.response_service = ResponseService(self.response_manager, self.audio_service)
         self.health_service = HealthService(self.storage_repo, self.config)
 
-
 # FastAPI App Factory
 def create_app() -> FastAPI:
     app = FastAPI(title="Unified Chat API")

self_hosted_env/voice_agent/requirements.txt

@@ -5,3 +5,4 @@ soundfile
 fastapi
 uvicorn[standard]
 python-multipart
+openai

self_hosted_env/voice_agent/services/audio_service.py

@@ -17,14 +17,21 @@ class AudioService:
 
     def get_audio_file(self, filename: str) -> str:
         try:
+            # Add the audio/ prefix to match the MinIO structure
+            minio_file_path = f"audio/{filename}"
+            print(f"Attempting to download from MinIO: bucket={self.bucket}, path={minio_file_path}")
+            
             temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
-            self.storage_repo.download_file(self.bucket, filename, temp_file.name)
+            self.storage_repo.download_file(self.bucket, minio_file_path, temp_file.name)
             temp_file.close()
+            
+            print(f"Successfully downloaded audio file: {minio_file_path}")
             return temp_file.name
+            
         except ClientError as e:
             print(f"MinIO ClientError: {e}")
+            print(f"Failed to find: bucket={self.bucket}, path=audio/{filename}")
             raise HTTPException(status_code=404, detail=f"Audio file '{filename}' not found.")
         except Exception as e:
             print(f"An error occurred: {e}")
-            raise HTTPException(status_code=500, detail=f"An error occurred: {e}")
-
+            raise HTTPException(status_code=500, detail=f"An error occurred: {e}")
\ No newline at end of file

self_hosted_env/voice_agent/services/response_manager.py

@@ -3,12 +3,13 @@ from typing import Optional
 
 class ResponseManager:
     def __init__(self):
-        self._latest_response = {"text": None, "audio_file_path": None, "timestamp": 0}
+        self._latest_response = {"text": None, "audio_filename": None, "timestamp": 0}
 
-    def store_response(self, text: str, audio_file_path: Optional[str] = None) -> None:
+    def store_response(self, text: str, audio_filename: Optional[str] = None) -> None:
+        """Store response with audio filename instead of file path"""
         self._latest_response = {
             "text": text,
-            "audio_file_path": audio_file_path,
+            "audio_filename": audio_filename,
             "timestamp": time.time()
         }
 
@@ -16,8 +17,8 @@ class ResponseManager:
         return self._latest_response.copy()
 
     def clear_response(self) -> None:
-        self._latest_response = {"text": None, "audio_file_path": None, "timestamp": 0}
+        self._latest_response = {"text": None, "audio_filename": None, "timestamp": 0}
 
     def is_response_fresh(self, max_age_seconds: int = 300) -> bool:
         return (self._latest_response["text"] and 
-                (time.time() - self._latest_response["timestamp"] < max_age_seconds))
+                (time.time() - self._latest_response["timestamp"] < max_age_seconds))
\ No newline at end of file

self_hosted_env/voice_agent/services/response_service.py

@@ -21,8 +21,9 @@ class ResponseService:
         
         response_data = self.response_manager.get_response()
         
-        if response_data["audio_file_path"]:
-            file_path = response_data["audio_file_path"]
+        if response_data["audio_filename"]:
+            # Download audio file from MinIO using filename
+            file_path = self.audio_service.get_audio_file(response_data["audio_filename"])
             response_text = response_data["text"]
             self.response_manager.clear_response()
             
@@ -44,4 +45,4 @@ class ResponseService:
                 "status": ResponseStatus.SUCCESS, 
                 "message": "Text response available.", 
                 "text": response_text
-            }
+            }
\ No newline at end of file

self_hosted_env/voice_agent/services/webhook_service.py

@@ -4,28 +4,102 @@ from typing import Optional
 from enum import Enum
 import sys
 import os
+import time
+import tempfile
+from openai import OpenAI
 sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
 from core import ResponseStatus
 from schemas import WebhookResponse
 from services.response_manager import ResponseManager
-from services.audio_service import AudioService
+from repositories import StorageRepository
 
 class WebhookService:
-    def __init__(self, response_manager: ResponseManager, audio_service: AudioService):
+    def __init__(self, response_manager: ResponseManager, storage_repo: StorageRepository, bucket: str):
         self.response_manager = response_manager
-        self.audio_service = audio_service
+        self.storage_repo = storage_repo
+        self.bucket = bucket
+        
+        # Initialize OpenAI client
+        openai_api_key = os.getenv("OPENAI_API_KEY")
+        if not openai_api_key:
+            print("Warning: OPENAI_API_KEY not found. TTS functionality will be disabled.")
+            self.openai_client = None
+        else:
+            self.openai_client = OpenAI(api_key=openai_api_key)
+
+    def generate_audio_from_text(self, text: str) -> str:
+        """Generate audio from text using OpenAI TTS and upload to MinIO"""
+        if not self.openai_client:
+            raise HTTPException(status_code=500, detail="OpenAI API key not configured")
+            
+        try:
+            # Create temporary file for audio
+            temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+            temp_file_path = temp_file.name
+            temp_file.close()
+            
+            print(f"Generating TTS audio for text: {text[:50]}...")
+            
+            # Generate audio using OpenAI TTS
+            with self.openai_client.audio.speech.with_streaming_response.create(
+                model="gpt-4o-mini-tts",  
+                voice="alloy",  # Available voices: alloy, echo, fable, onyx, nova, shimmer
+                input=text,
+                response_format="mp3"
+            ) as response:
+                response.stream_to_file(temp_file_path)
+            
+            # Generate unique filename for MinIO 
+            timestamp = int(time.time())
+            filename = f"tts_response_{timestamp}.mp3"
+            # The full path in MinIO will be audio/filename
+            minio_file_path = f"audio/{filename}"
+            
+            print(f"Uploading generated audio to MinIO: {minio_file_path}")
+            
+            # Upload to MinIO
+            with open(temp_file_path, 'rb') as audio_file:
+                self.storage_repo.upload_file(audio_file, self.bucket, minio_file_path)
+            
+            # Clean up temporary file
+            if os.path.exists(temp_file_path):
+                os.unlink(temp_file_path)
+                print(f"Cleaned up temporary file: {temp_file_path}")
+            
+            print(f"Successfully generated and uploaded TTS audio: {filename}")
+            # Return just the filename, the audio service will add the audio/ prefix
+            return filename
+            
+        except Exception as e:
+            # Clean up temporary file in case of error
+            if 'temp_file_path' in locals() and os.path.exists(temp_file_path):
+                os.unlink(temp_file_path)
+            
+            print(f"Error generating TTS audio: {e}")
+            raise HTTPException(status_code=500, detail=f"Failed to generate audio: {str(e)}")
 
     def process_webhook_response(self, response: WebhookResponse) -> dict:
+        """Process webhook response from n8n and generate TTS audio"""
         print("Received webhook data from n8n:", response.dict())
         
         if response.agent_responded == 'yes' and response.agent_response:
-            if response.filename:
-                audio_file_path = self.audio_service.get_audio_file(response.filename)
-                self.response_manager.store_response(response.agent_response, audio_file_path)
-                print("Agent response stored successfully.")
-            else:
+            try:
+                # Generate audio from the text response
+                audio_filename = self.generate_audio_from_text(response.agent_response)
+                
+                # Store response with generated audio filename (just the filename, not the path)
+                self.response_manager.store_response(response.agent_response, audio_filename)
+                print("Agent response with generated TTS audio stored successfully.")
+                
+            except HTTPException:
+                # Re-raise HTTP exceptions
+                raise
+            except Exception as e:
+                print(f"Error generating TTS audio, storing text-only response: {e}")
+                # Fallback to text-only response if audio generation fails
                 self.response_manager.store_response(response.agent_response)
-                print("Agent text response stored successfully.")
+                print("Stored text-only response as fallback.")
+        else:
+            print("No valid agent response received from webhook.")
         
-        return {"status": ResponseStatus.SUCCESS, "message": "Webhook received and processed successfully."}
-
+        return {"status": ResponseStatus.SUCCESS, "message": "Webhook received and processed successfully."}
\ No newline at end of file