FIX add HF token (from env variable)

rag_system.py

@@ -132,7 +132,9 @@ class RAGSystem:
         """Set or change the LLM model"""
         if self.llm_model_name != model_name:
             self.llm_model_name = model_name
-            self.llm_client = InferenceClient(model_name)
+            # Use HF_TOKEN from environment if available
+            hf_token = os.environ.get("HF_TOKEN", None)
+            self.llm_client = InferenceClient(model_name, token=hf_token)
 
     @spaces.GPU
     def retrieve(