Update app.py

app.py

@@ -43,22 +43,36 @@ async def respond(
         "stream": True
     }
 
-    async with aiohttp.ClientSession() as session:
-        async with session.post("https://api-inference.huggingface.co/v1/chat/completions", headers=headers, json=payload) as response:
-            try:
+    try:
+        async with aiohttp.ClientSession() as session:
+            async with session.post("https://api-inference.huggingface.co/v1/chat/completions", headers=headers, json=payload) as response:
+                response_text = ""
                 async for chunk in response.content:
                     if chunk:
-                        chunk_data = chunk.decode('utf-8')
-                        response_json = json.loads(chunk_data)
-                        if "choices" in response_json:
-                            content = response_json["choices"][0]["message"]["content"]
-                            yield content
-            except json.JSONDecodeError:
-                pass
-            except StopAsyncIteration:
-                pass
-            finally:
-                pass  # 스트림 종료 시 아무것도 하지 않음
+                        try:
+                            chunk_data = chunk.decode('utf-8')
+                            response_json = json.loads(chunk_data)
+                            if "choices" in response_json:
+                                content = response_json["choices"][0]["message"]["content"]
+                                response_text += content
+                                yield response_text
+                        except json.JSONDecodeError:
+                            continue
+                
+                if not response_text:
+                    yield "I apologize, but I couldn't generate a response. Please try again."
+    except Exception as e:
+        yield f"An error occurred: {str(e)}"
+
+    memory[-1] = (message, response_text)
+
+async def chat(message, history, system_message, max_tokens, temperature, top_p):
+    response = ""
+    async for chunk in respond(message, history, system_message, max_tokens, temperature, top_p):
+        response = chunk
+        yield response
+
+
                 
 theme = "Nymbo/Nymbo_Theme"
 
@@ -70,7 +84,7 @@ footer {
 
 demo = gr.ChatInterface(
     css=css,
-    fn=respond,
+    fn=chat,
     theme=theme,
     additional_inputs=[
         gr.Textbox(value="AI Assistant Role", label="System message"),