fix + client

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+    "python.pythonPath": "/Users/michael/.pyenv/versions/3.12.3/bin/python3.12"
+}

app.py

@@ -3,7 +3,6 @@ from fastapi.responses import StreamingResponse
 from huggingface_hub import InferenceClient
 from pydantic import BaseModel
 from typing import List, Tuple
-import asyncio
 
 # Initialisation du client Hugging Face
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
@@ -12,6 +11,8 @@ client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 app = FastAPI()
 
 # Modèle pour les données d'entrée
+
+
 class PredictionRequest(BaseModel):
     message: str
     history: List[Tuple[str, str]] = []
@@ -21,20 +22,21 @@ class PredictionRequest(BaseModel):
     top_p: float = 0.95
 
 
-async def generate_stream(request: PredictionRequest):
+def generate_stream(request: PredictionRequest):
     """
-    Générateur asynchrone pour le streaming de réponse.
+    Générateur synchrone pour le streaming de réponse.
     """
     messages = [{"role": "system", "content": request.system_message}]
     for user_input, assistant_response in request.history:
         if user_input:
             messages.append({"role": "user", "content": user_input})
         if assistant_response:
-            messages.append({"role": "assistant", "content": assistant_response})
+            messages.append(
+                {"role": "assistant", "content": assistant_response})
     messages.append({"role": "user", "content": request.message})
-    
+
     try:
-        async for message in client.chat_completion(
+        for message in client.chat_completion(
             messages,
             max_tokens=request.max_tokens,
             stream=True,
@@ -44,7 +46,7 @@ async def generate_stream(request: PredictionRequest):
             token = message.choices[0].delta.content
             yield token
     except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
+        yield f"Error: {str(e)}"
 
 
 @app.post("/predict")

client.py

@@ -0,0 +1,26 @@
+import requests
+
+# URL de ton endpoint
+url = "https://micksoftware-laria-startup.hf.space/predict"
+
+# Données pour la requête
+payload = {
+    "message": "salut",
+    "history": [],
+    "system_message": "You are a friendly Chatbot.",
+    "max_tokens": 512,
+    "temperature": 0.7,
+    "top_p": 0.95,
+}
+
+# Faire la requête avec streaming activé
+response = requests.post(url, json=payload, stream=True)
+
+# Lire la réponse en streaming
+if response.status_code == 200:
+    print("Streaming response:")
+    for chunk in response.iter_lines(decode_unicode=True):
+        if chunk:
+            print(chunk, end="")
+else:
+    print(f"Erreur : {response.status_code} - {response.text}")

doc.md

@@ -0,0 +1,17 @@
+# HG URL
+https://huggingface.co/spaces/MiCkSoftware/laria-startup
+
+# Sample call
+curl -X 'POST' \
+  'https://micksoftware-laria-startup.hf.space/predict' \
+  -H 'accept: application/json' \
+  -H 'Content-Type: application/json' \
+  -d '{
+  "message": "salut",
+  "history": [],
+  "system_message": "You are a friendly Chatbot.",
+  "max_tokens": 512,
+  "temperature": 0.7,
+  "top_p": 0.95
+}'
+