update services

routes/tts_routes.py

@@ -8,11 +8,12 @@ router = APIRouter(prefix="/tts", tags=["Text To Speech"])
 
 @router.post("/")
 async def generate_tts(request: TTSRequest):
-    output_path: Path = text_to_speech(
-        text=request.text,
-        voice=request.voice,
-        fmt=request.format
-    )
+    output_path = text_to_speech(
+    text=request.text,
+    voice=request.voice,
+    fmt=request.format
+)
+
 
     return FileResponse(
         output_path,

services/stt_service.py

@@ -1,26 +1,26 @@
-# services/stt_service.py
 import requests
 from config import GROQ_API_KEY, GROQ_STT_MODEL
 
 def speech_to_text(audio_file: str) -> str:
-    """
-    Convertit un fichier audio en texte via l'API Groq STT.
-    audio_file : str -> chemin du fichier audio
-    """
     if not GROQ_API_KEY:
         raise RuntimeError("GROQ_API_KEY is not set in config")
 
+    url = "https://api.groq.ai/openai/v1/audio/transcriptions"
+
     headers = {
         "Authorization": f"Bearer {GROQ_API_KEY}"
     }
 
-    files = {"file": open(audio_file, "rb")}
-
-    url = f"https://api.groq.ai/v1/models/{GROQ_STT_MODEL}/predict"
+    with open(audio_file, "rb") as f:
+        files = {
+            "file": (audio_file, f, "audio/wav")
+        }
+        data = {
+            "model": GROQ_STT_MODEL
+        }
 
-    response = requests.post(url, headers=headers, files=files)
-    response.raise_for_status()
+        response = requests.post(url, headers=headers, data=data, files=files)
+        response.raise_for_status()
 
-    result = response.json()
-    # Supposons que l'API retourne le texte sous 'text'
-    return result.get("text", "")
+        result = response.json()
+        return result.get("text", "")

services/tts_service.py

@@ -1,34 +1,31 @@
-# services/tts_service.py
 import requests
-from config import GROQ_API_KEY, GROQ_TTS_MODEL, GROQ_TTS_VOICE, GROQ_TTS_FORMAT
-
-def text_to_speech(text: str, output_file: str):
-    """
-    Convertit du texte en audio via l'API Groq TTS.
-    text : str -> texte à convertir
-    output_file : str -> chemin du fichier de sortie (ex: 'output.wav')
-    """
+import uuid
+from config import GROQ_API_KEY, GROQ_TTS_MODEL
+
+def text_to_speech(text: str, voice: str, fmt: str):
     if not GROQ_API_KEY:
         raise RuntimeError("GROQ_API_KEY is not set in config")
 
+    url = "https://api.groq.ai/openai/v1/audio/speech"
+
     headers = {
         "Authorization": f"Bearer {GROQ_API_KEY}",
         "Content-Type": "application/json"
     }
 
     payload = {
-        "text": text,
-        "voice": GROQ_TTS_VOICE,
-        "format": GROQ_TTS_FORMAT
+        "model": GROQ_TTS_MODEL,
+        "voice": voice,
+        "format": fmt,
+        "input": text
     }
 
-    url = f"https://api.groq.ai/v1/models/{GROQ_TTS_MODEL}/predict"
+    output_file = f"audio_{uuid.uuid4()}.{fmt}"
 
     response = requests.post(url, headers=headers, json=payload)
-    response.raise_for_status()  # Si erreur, lève une exception
+    response.raise_for_status()
 
-    audio_content = response.content
     with open(output_file, "wb") as f:
-        f.write(audio_content)
+        f.write(response.content)
 
     return output_file