Update app.py

app.py

@@ -4,6 +4,7 @@ import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import time
 import uuid
+from typing import Optional, List
 
 # --- Configuration ---
 MODEL_ID = "deepseek-ai/deepseek-coder-1.3b-instruct"
@@ -28,9 +29,10 @@ class ChatMessage(BaseModel):
     content: str
 
 class ChatCompletionRequest(BaseModel):
-    model: str
-    messages: list[ChatMessage]
-    max_tokens: int = 250
+    # CHANGEMENT 1: Rendre le champ 'model' optionnel pour éviter l'erreur 422
+    model: Optional[str] = None
+    messages: List[ChatMessage]
+    max_tokens: Optional[int] = 250
 
 class ChatCompletionResponseChoice(BaseModel):
     index: int = 0
@@ -42,15 +44,29 @@ class ChatCompletionResponse(BaseModel):
     object: str = "chat.completion"
     created: int
     model: str
-    choices: list[ChatCompletionResponseChoice]
+    choices: List[ChatCompletionResponseChoice]
+
+class ModelData(BaseModel):
+    id: str
+    object: str = "model"
+    created: int = int(time.time())
+    owned_by: str = "user"
+
+class ModelList(BaseModel):
+    object: str = "list"
+    data: List[ModelData]
+
+# --- Définition des API ---
+
+# CHANGEMENT 2: Ajouter un endpoint /models pour satisfaire l'extension
+@app.get("/models", response_model=ModelList)
+async def list_models():
+    """Endpoint pour lister les modèles disponibles."""
+    return ModelList(data=[ModelData(id=MODEL_ID)])
 
-# --- Définition de l'API compatible OpenAI ---
-# LA SEULE LIGNE QUI CHANGE EST CELLE-CI : on a enlevé le "/v1"
-@app.post("/chat/completions")
+@app.post("/chat/completions", response_model=ChatCompletionResponse)
 async def create_chat_completion(request: ChatCompletionRequest):
-    """
-    Endpoint compatible avec l'API OpenAI Chat Completions.
-    """
+    """Endpoint principal pour la génération de texte."""
     user_prompt = ""
     if request.messages and request.messages[-1].role == "user":
         user_prompt = request.messages[-1].content
@@ -70,7 +86,7 @@ async def create_chat_completion(request: ChatCompletionRequest):
     completion_response = ChatCompletionResponse(
         id=f"chatcmpl-{uuid.uuid4()}",
         created=int(time.time()),
-        model=request.model,
+        model=MODEL_ID,
         choices=[choice]
     )