update model

app.py

@@ -86,9 +86,12 @@ class MinimalistTheme(Base):
 class RAGChatbot:
     """Chatbot with RAG capabilities."""
     
+    # Default model: use one supported by HF Inference API (Mistral often requires enabled providers)
+    DEFAULT_CHAT_MODEL = "HuggingFaceH4/zephyr-7b-beta"
+
     def __init__(
         self,
-        model_name: str = "mistralai/Mistral-7B-Instruct-v0.2",
+        model_name: str = None,
         embedding_model: str = "all-mpnet-base-v2",
         vector_store_path: str = "data/vector_store"
     ):
@@ -100,8 +103,8 @@ class RAGChatbot:
             embedding_model: Model for document embeddings
             vector_store_path: Path to saved vector store
         """
-        self.model_name = model_name
-        
+        self.model_name = model_name if model_name else self.DEFAULT_CHAT_MODEL
+
         # Initialize Inference API client
         hf_token = os.environ.get("HF_TOKEN")
         # Debug: report HF_TOKEN status (masked)
@@ -114,10 +117,10 @@ class RAGChatbot:
             print(f"[DEBUG] HF_TOKEN: set (length={len(hf_token)}, masked={masked})")
             print("HF_TOKEN found. Inference API ready.")
         
-        print(f"[DEBUG] Initializing Inference API client for model: {model_name}")
+        print(f"[DEBUG] Initializing Inference API client for model: {self.model_name}")
         try:
             self.inference_client = InferenceClient(
-                model=model_name,
+                model=self.model_name,
                 token=hf_token
             )
             print("[DEBUG] Inference API client initialized successfully (using chat_completion for this model)")
@@ -211,6 +214,13 @@ Answer:"""
                         raise ValueError("Empty response from model")
                     except Exception as api_error:
                         print(f"[DEBUG] RAG generation failed: {type(api_error).__name__}: {api_error}")
+                        err_str = str(api_error).lower()
+                        if "model_not_supported" in err_str or "not supported by any provider" in err_str:
+                            return (
+                                "The current chat model isn't available with your Inference API providers. "
+                                "Try using the default model (Zephyr) or enable a provider for your chosen model at "
+                                "https://huggingface.co/settings/inference-api."
+                            )
                         # Fallback: return formatted chunks with note
                         response_parts = []
                         response_parts.append("I retrieved relevant information, but couldn't generate a synthesized answer. Here are the relevant chunks:\n\n")