Update app.py

app.py

@@ -151,6 +151,7 @@ def deepseek_respond(
     message,
     chat_history,
     system_message,
+    deepseek_model_choice,
     max_tokens,
     temperature,
     top_p,
@@ -158,11 +159,14 @@ def deepseek_respond(
     """
     DeepSeek 모델 응답 함수.
     DeepSeek 토큰은 함수 내부에서 os.environ을 통해 불러온다.
+    deepseek_model_choice에 따라 deepseek-chat 또는 deepseek-reasoner를 선택하며,
+    스트리밍 방식으로 응답을 받아옵니다.
     """
     deepseek_token = os.environ.get("DEEPSEEK_TOKEN")
     if not deepseek_token:
         chat_history.append((message, "DeepSeek API 토큰이 필요합니다. (환경변수 DEEPSEEK_TOKEN 미설정)"))
-        return chat_history
+        yield chat_history
+        return
 
     openai.api_key = deepseek_token
     openai.api_base = "https://api.deepseek.com/v1"
@@ -173,21 +177,36 @@ def deepseek_respond(
         messages.append({"role": "assistant", "content": assistant})
     messages.append({"role": "user", "content": message})
 
+    # 모델 선택: 기본은 deepseek-chat
+    if deepseek_model_choice == "R1(deepseek-reasoner)":
+        model = "deepseek-reasoner"
+    else:
+        model = "deepseek-chat"
+
     try:
         response = openai.ChatCompletion.create(
-            model="deepseek-chat",
+            model=model,
             messages=messages,
             max_tokens=max_tokens,
             temperature=temperature,
             top_p=top_p,
+            stream=True
         )
-        assistant_message = response.choices[0].message['content']
+        assistant_message = ""
+        # 새로운 대화 항목을 추가하고 초기값을 스트리밍하면서 갱신
         chat_history.append((message, assistant_message))
-        return chat_history
+        yield chat_history
+        for chunk in response:
+            delta = chunk.choices[0].delta.get("content", "")
+            assistant_message += delta
+            chat_history[-1] = (message, assistant_message)
+            yield chat_history
+        return
     except Exception as e:
         error_message = f"오류가 발생했습니다: {str(e)}"
         chat_history.append((message, error_message))
-        return chat_history
+        yield chat_history
+        return
 
 def clear_conversation():
     return []
@@ -331,6 +350,11 @@ with gr.Blocks() as demo:
                 label="System Message",
                 lines=3
             )
+            deepseek_model_choice = gr.Radio(
+                choices=["V3(deepseek-chat)", "R1(deepseek-reasoner)"],
+                value="V3(deepseek-chat)",
+                label="모델 선택"
+            )
             deepseek_max_tokens = gr.Slider(minimum=100, maximum=8000, value=2000, step=100, label="Max Tokens")
             deepseek_temperature = gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.05, label="Temperature")
             deepseek_top_p = gr.Slider(
@@ -351,12 +375,13 @@ with gr.Blocks() as demo:
             deepseek_msg,
             deepseek_chatbot,
             deepseek_system_message,
+            deepseek_model_choice,
             deepseek_max_tokens,
             deepseek_temperature,
             deepseek_top_p
         ]
-        deepseek_msg.submit(deepseek_respond, inputs_for_deepseek, deepseek_chatbot)
-        deepseek_submit_button.click(deepseek_respond, inputs_for_deepseek, deepseek_chatbot)
+        deepseek_msg.submit(deepseek_respond, inputs_for_deepseek, deepseek_chatbot, stream=True)
+        deepseek_submit_button.click(deepseek_respond, inputs_for_deepseek, deepseek_chatbot, stream=True)
         deepseek_clear_button.click(clear_conversation, outputs=deepseek_chatbot, queue=False)
 
 if __name__ == "__main__":