Update app.py

app.py

@@ -45,15 +45,14 @@ from session_rag import SessionRAG
 from mdsi_analysis import capacity_projection, cost_estimate, outcomes_summary
 
 # ---------- Config ----------
-# Local fallback model (lightweight by default). You can override via env.
-MODEL_ID = os.getenv("MODEL_ID", "microsoft/Phi-3-mini-4k-instruct")
+MODEL_ID = os.getenv("MODEL_ID", "microsoft/Phi-3-mini-4k-instruct")  # local fallback
 HF_TOKEN = os.getenv("HUGGINGFACE_HUB_TOKEN") or os.getenv("HF_TOKEN")
 
 COHERE_API_KEY = os.getenv("COHERE_API_KEY")
 USE_HOSTED_COHERE = bool(COHERE_API_KEY and _HAS_COHERE)
 COHERE_TIMEOUT_SEC = float(os.getenv("COHERE_TIMEOUT_SEC", "30"))
 
-MAX_NEW_TOKENS = int(os.getenv("MAX_NEW_TOKENS", "512"))  # faster defaults; adjust as needed
+MAX_NEW_TOKENS = int(os.getenv("MAX_NEW_TOKENS", "512"))
 
 # ---------- Helpers ----------
 def pick_dtype_and_map():
@@ -376,7 +375,7 @@ with gr.Blocks(theme=theme, css=custom_css) as demo:
 
     uploads.change(fn=_store_uploads, inputs=[uploads, state_uploaded], outputs=state_uploaded)
 
-    # Send / Enter handlers
+    # Send / Enter handlers (set concurrency_limit here per Gradio deprecation notice)
     def _on_send(user_msg, history, tz, up_paths):
         if not user_msg or not user_msg.strip():
             return history, "", history
@@ -387,22 +386,22 @@ with gr.Blocks(theme=theme, css=custom_css) as demo:
         fn=_on_send,
         inputs=[msg, state_history, tz_box, state_uploaded],
         outputs=[chat, msg, state_history],
+        concurrency_limit=2,
         queue=True,
     )
     msg.submit(
         fn=_on_send,
         inputs=[msg, state_history, tz_box, state_uploaded],
         outputs=[chat, msg, state_history],
+        concurrency_limit=2,
         queue=True,
     )
 
     # Clear chat (keep uploads)
     clear.click(lambda: ([], "", []), None, [chat, msg, state_history])
 
-# Enable queue to avoid websocket timeouts on first call / heavy loads
-demo = demo.queue(concurrency_count=2, max_size=32)
-
 if __name__ == "__main__":
     port = int(os.environ.get("PORT", "7860"))
     demo.launch(server_name="0.0.0.0", server_port=port, show_api=False, max_threads=8)
 
+