add down message

app.py

@@ -100,6 +100,11 @@ class RAGChatbot:
         "meta-llama/Llama-3.1-8B-Instruct",
     ]
 
+    # Shown to users when the HF account behind HF_TOKEN has run out of Inference Providers credits (HTTP 402)
+    OUT_OF_CREDITS_MESSAGE = (
+        "The Organic Certification Assistant is temporarily unavailable. Please try again later."
+    )
+
     def __init__(
         self,
         model_name: str = None,
@@ -155,6 +160,15 @@ class RAGChatbot:
         
         self.chat_history = []
     
+    @staticmethod
+    def _is_out_of_credits(error: Exception) -> bool:
+        """Return True if the error is a 402 Payment Required (no Inference Providers credits left)."""
+        status = getattr(getattr(error, "response", None), "status_code", None)
+        if status == 402:
+            return True
+        err_str = str(error).lower()
+        return "402 payment required" in err_str or "no remaining credits" in err_str
+
     def _generate_with_chat(self, user_content: str, max_new_tokens: int = 512) -> str:
         """Call the Inference API using chat_completion; try fallback models if current is not supported."""
         last_error = None
@@ -302,6 +316,10 @@ Answer:"""
                         raise ValueError("Empty response from model")
                     except Exception as api_error:
                         print(f"[DEBUG] RAG generation failed: {type(api_error).__name__}: {api_error}")
+                        if self._is_out_of_credits(api_error):
+                            print("[DEBUG] Inference Providers credits exhausted for the account behind HF_TOKEN. "
+                                  "Add credits at https://huggingface.co/settings/billing")
+                            return self.OUT_OF_CREDITS_MESSAGE
                         err_str = str(api_error).lower()
                         if "model_not_supported" in err_str or "not supported by any provider" in err_str:
                             return (
@@ -341,6 +359,10 @@ Answer:"""
             return "I couldn't generate a response. Please try again."
         except Exception as e:
             print(f"Error generating response: {e}")
+            if self._is_out_of_credits(e):
+                print("[DEBUG] Inference Providers credits exhausted for the account behind HF_TOKEN. "
+                      "Add credits at https://huggingface.co/settings/billing")
+                return self.OUT_OF_CREDITS_MESSAGE
             return f"I encountered an error while generating a response: {str(e)}. Please check your HF_TOKEN configuration."
     
     def chat(self, message: str, history):