feat: Update TTS functionality to use gTTS for text-to-speech conversion, supporting only MP3 format. Adjust requirements and model fields for compatibility.

models/tts.py

@@ -2,14 +2,16 @@ from pydantic import BaseModel, Field
 
 class TTSRequest(BaseModel):
     text: str = Field(..., min_length=1, max_length=5000)
-    voice: str = Field(default="Aaliyah-PlayAI")
-    format: str = Field(default="wav", pattern="^(wav|mp3)$")
-    
+    # gTTS uses language codes; keep voice field for compatibility.
+    voice: str = Field(default="en")
+    # Free backend supports only mp3.
+    format: str = Field(default="mp3", pattern="^(mp3)$")
+
     class Config:
         json_schema_extra = {
             "example": {
                 "text": "Hello, this is a test of text-to-speech.",
-                "voice": "Aaliyah-PlayAI",
-                "format": "wav"
+                "voice": "en",
+                "format": "mp3",
             }
         }

requirements.txt

@@ -19,6 +19,7 @@ langsmith>=0.1.0
 # Fix urllib3 compatibility issues
 urllib3>=1.26.0,<3.0.0
 requests-toolbelt>=1.0.0
+gTTS>=2.5.3
 
 # Audio processing (optionnel si vous avez besoin de traitement local)
 soundfile>=0.12.1

routes/tts_routes.py

@@ -9,32 +9,28 @@ router = APIRouter(prefix="/tts", tags=["Text To Speech"])
 @router.post("/")
 async def generate_tts(request: TTSRequest):
     """
-    Convert text to speech (English only)
+    Convert text to speech using the free gTTS backend (MP3 only).
     """
     try:
-        # Generate audio
         audio_path = text_to_speech(
             text=request.text,
             voice=request.voice,
-            fmt=request.format
+            fmt=request.format,
         )
-        
-        # Verify file exists
+
         if not Path(audio_path).exists():
             raise HTTPException(status_code=500, detail="Audio file generation failed")
-        
-        # Determine MIME type
-        media_type = "audio/wav" if request.format == "wav" else "audio/mpeg"
-        
-        # Return audio file
+
+        media_type = "audio/mpeg"
+
         return FileResponse(
             path=audio_path,
             filename=f"speech.{request.format}",
             media_type=media_type,
             headers={
                 "Content-Disposition": f"attachment; filename=speech.{request.format}"
-            }
+            },
         )
-        
+
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))

services/tts_service.py

@@ -1,17 +1,21 @@
-import requests
 import uuid
-import os
 from pathlib import Path
 from config import GROQ_TTS_API_KEY, GROQ_TTS_MODEL
+<<<<<<< Updated upstream
+=======
+from gtts import gTTS
+>>>>>>> Stashed changes
 
 def text_to_speech(
-    text: str, 
-    voice: str = "Aaliyah-PlayAI",
-    fmt: str = "wav"
+    text: str,
+    voice: str = "en",
+    fmt: str = "mp3",
 ) -> str:
     """
-    Convert text to speech using Groq's TTS API (English only)
+    Convert text to speech using gTTS (Google Translate, free).
+    Only MP3 is supported.
     """
+<<<<<<< Updated upstream
     if not GROQ_TTS_API_KEY:
         raise RuntimeError("GROQ_TTS_API_KEY is not set in config")
     
@@ -32,26 +36,26 @@ def text_to_speech(
         "response_format": fmt
     }
     
+=======
+    if not text or not text.strip():
+        raise ValueError("Text cannot be empty")
+
+    if fmt != "mp3":
+        raise ValueError("Only MP3 format is supported by the free TTS backend")
+
+>>>>>>> Stashed changes
     try:
-        # Create temp directory for audio files
         temp_dir = Path("temp_audio")
         temp_dir.mkdir(exist_ok=True)
-        
-        # Unique filename
+
         output_filename = f"tts_{uuid.uuid4().hex[:8]}.{fmt}"
         output_path = temp_dir / output_filename
-        
-        # Call Groq API
-        response = requests.post(url, headers=headers, json=payload, timeout=30)
-        response.raise_for_status()
-        
-        # Save audio file
-        with open(output_path, "wb") as f:
-            f.write(response.content)
-        
+
+        # gTTS uses language codes; voice kept for compatibility.
+        tts = gTTS(text=text.strip(), lang=voice or "en")
+        tts.save(str(output_path))
+
         return str(output_path)
-        
-    except requests.exceptions.RequestException as e:
-        raise Exception(f"Groq TTS API error: {str(e)}")
+
     except Exception as e:
         raise Exception(f"Unexpected error in text_to_speech: {str(e)}")