use public token

app.py

@@ -138,20 +138,15 @@ def augment_user_message(user_msg: str, context: str) -> str:
         f"{context}\n\n"
         f"User question: {user_msg}"
     )
-
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    hf_token: gr.OAuthToken,
-):
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    token = os.getenv("HF_TOKEN")  # Space Secret or env var
+    if not token:
+        yield "⚠️ No HF token found on the server. Please set HF_TOKEN in the Space secrets."
+        return
     """
     Augments the conversation with CSV-derived context as SYSTEM messages before calling the model.
     """
-    client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
+    client = InferenceClient(token=token, model="openai/gpt-oss-20b")
 
     # 1) Retrieve CSV context for the latest user message
     retrieved = RAG.retrieve(message, k=10)