Update app.py

app.py

@@ -1,10 +1,51 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
-from gradio_client import Client
 
+# --- OpenAI (SDK officiel) ---
+from openai import OpenAI
+openai_client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
 
-# Se connecter au Space qui expose le MCP
-mcp_client = Client("HackathonCRA/mcp")
+# --- Mistral (SDK officiel) ---
+from mistralai import Mistral
+mistral_client = Mistral(api_key=os.getenv("MISTRAL_API_KEY"))
+
+# Choisis à l’exécution quel LLM utiliser
+DEFAULT_PROVIDER = "openai"  # ou "mistral"
+
+def llm_chat(messages, provider=DEFAULT_PROVIDER, max_tokens, temperature, top_p):
+    if provider == "openai":
+        # Chat Completions (ex: gpt-4o-mini / gpt-4.1 / … selon ton compte)
+        stream = openai_client.chat.completions.create(
+            model="gpt-4o-mini",
+            messages=messages,
+            temperature=temperature,
+            top_p=top_p,
+            max_tokens=max_tokens,
+            stream=True,
+        )
+        for chunk in stream:
+            delta = chunk.choices[0].delta
+            if delta and delta.content:
+                yield delta.content
+
+    elif provider == "mistral":
+        # Models: "mistral-large-latest", "open-mistral-nemo", etc.
+        stream = mistral_client.chat.stream(
+            model="mistral-large-latest",
+            messages=messages,
+            temperature=temperature,
+            top_p=top_p,
+            max_tokens=max_tokens,
+        )
+        for event in stream:
+            if event.type == "chat.completion.chunk":
+                piece = event.data.delta or ""
+                if piece:
+                    yield piece
+        stream.close()
+
+    else:
+        yield "[Erreur] Provider inconnu."
 
 
 def respond(
@@ -14,57 +55,32 @@ def respond(
     max_tokens,
     temperature,
     top_p,
-    hf_token: gr.OAuthToken,
+    provider  # "openai" ou "mistral"
 ):
-    """
-    Chat client qui envoie d'abord au modèle HF, et qui peut ensuite
-    appeler le MCP via gradio_client si besoin.
-    """
-    llm = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
-
+    # Construit le format messages attendu par les SDKs
     messages = [{"role": "system", "content": system_message}]
     messages.extend(history)
     messages.append({"role": "user", "content": message})
 
     response = ""
-
-    for msg in llm.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        choices = msg.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
-
+    for token in llm_chat(messages, provider, max_tokens, temperature, top_p):
         response += token
         yield response
 
-    # 👉 Exemple : appel au MCP (ici on appelle une fn "search_docs" du serveur)
-    # tu adaptes en fonction de ce que ton serveur expose comme endpoints
-    result = mcp_client.predict("ma requête", api_name="/search_docs")
-    yield f"\n\n[MCP result] {result}"
-
-
 chatbot = gr.ChatInterface(
     respond,
     type="messages",
     additional_inputs=[
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
+        gr.Slider(minimum=1, maximum=4096, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.0, maximum=2.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.0, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
+        gr.Dropdown(choices=["openai", "mistral"], value=DEFAULT_PROVIDER, label="Provider"),
     ],
 )
 
 with gr.Blocks() as demo:
-    with gr.Sidebar():
-        gr.LoginButton()
     chatbot.render()
 
-
 if __name__ == "__main__":
     demo.launch()