Update app.py

app.py

@@ -1632,8 +1632,10 @@ print("Transformer Output:", transformer_output)
 
 
 
-import gradio as gr  
-from openai import OpenAI  
+import gradio as gr
+import os
+import time
+from openai import OpenAI
 
 # Load system prompt from environment variable
 SYSTEM_PROMPT = os.getenv("SYSTEM_PROMPT")
@@ -1646,23 +1648,43 @@ client = OpenAI(
     api_key="hyperbolic"  
 )  
 
-def predict(message, history):  
+# Retry settings
+MAX_RETRIES = 5  # Number of retries
+RETRY_DELAY = 1  # Delay in seconds between retries
+
+def predict(message, history):
     # If history is empty, insert the system prompt
     if not any(msg["role"] == "system" for msg in history):
         history.insert(0, {"role": "system", "content": SYSTEM_PROMPT})
 
-    history.append({"role": "user", "content": message})  
-
-    stream = client.chat.completions.create(  
-        messages=history,  
-        model=os.getenv("ACCEMULECTPLUS"),  
-        stream=True  
-    )  
-
-    chunks = []  
-    for chunk in stream:  
-        chunks.append(chunk.choices[0].delta.content or "")  
-        yield "".join(chunks)  
+    history.append({"role": "user", "content": message})
+
+    retries = 0
+    while retries < MAX_RETRIES:
+        try:
+            # Attempt to get the completion stream
+            stream = client.chat.completions.create(  
+                messages=history,  
+                model=os.getenv("ACCEMULECTPLUS"),  
+                stream=True  
+            )
+
+            # Collect the response
+            chunks = []  
+            for chunk in stream:  
+                chunks.append(chunk.choices[0].delta.content or "")  
+                yield "".join(chunks)
+            
+            # If no exception occurs, break the retry loop
+            break
+        except Exception as e:
+            print(f"Error occurred: {e}. Retrying ({retries + 1}/{MAX_RETRIES})...")
+            retries += 1
+            if retries < MAX_RETRIES:
+                time.sleep(RETRY_DELAY)  # Delay before retrying
+            else:
+                print("Max retries reached. Could not complete the request.")
+                yield "Sorry, there was an error processing your request. Please try again later."
 
 demo = gr.ChatInterface(
     fn=predict,