Update app.py

app.py

@@ -4,6 +4,7 @@ import random
 from functools import partial
 import gradio as gr
 from huggingface_hub import InferenceClient
+import threading
 
 css = """
 gradio-app {
@@ -79,16 +80,19 @@ for c in client.chat_completion(messages, max_tokens=200, stream=True):
 """
 
 ip_requests = {}
+ip_requests_lock = threading.Lock()
 
 def allow_ip(request: gr.Request, show_error=True):
     ip = request.client.host
     now = datetime.now()
     window = timedelta(hours=24)
-    if ip in ip_requests:
-        ip_requests[ip] = [timestamp for timestamp in ip_requests[ip] if now - timestamp < window]
-    if len(ip_requests.get(ip, [])) >= 15:
-        raise gr.Error("Rate limit exceeded. Please try again tomorrow or use your Hugging Face Pro token.", visible=show_error)
-    ip_requests.setdefault(ip, []).append(now)
+
+    with ip_requests_lock:
+        if ip in ip_requests:
+            ip_requests[ip] = [timestamp for timestamp in ip_requests[ip] if now - timestamp < window]
+        if len(ip_requests.get(ip, [])) >= 15:
+            raise gr.Error("Rate limit exceeded. Please try again tomorrow or use your Hugging Face Pro token.", visible=show_error)
+        ip_requests.setdefault(ip, []).append(now)
     return True
 
 def inference(prompt, hf_token, model, model_name):
@@ -213,4 +217,6 @@ with gr.Blocks(css=css, theme="NoCrypt/miku", js=js) as demo:
     )
 
 
-demo.launch(show_api=False)
+demo.launch(show_api=False)
+
+