Update app.py

app.py

@@ -1,10 +1,10 @@
 import discord
 import logging
 import os
+from datasets import load_dataset
 from huggingface_hub import InferenceClient
 import asyncio
 import subprocess
-from datasets import load_dataset
 
 # 로깅 설정
 logging.basicConfig(level=logging.DEBUG, format='%(asctime)s:%(levelname)s:%(name)s: %(message)s', handlers=[logging.StreamHandler()])
@@ -26,7 +26,16 @@ SPECIFIC_CHANNEL_ID = int(os.getenv("DISCORD_CHANNEL_ID"))
 conversation_history = []
 
 # 법률 데이터셋 로드
-law_dataset = load_dataset("aiqtech/kolaw")
+law_dataset = load_dataset('csv', data_files={
+    'train': [
+        '/path/to/train_0.csv',
+        '/path/to/train_1.csv',
+        '/path/to/train_2.csv',
+        '/path/to/train_3.csv',
+        '/path/to/train_4.csv',
+        '/path/to/train_5.csv'
+    ]
+})
 
 class MyClient(discord.Client):
     def __init__(self, *args, **kwargs):
@@ -39,35 +48,25 @@ class MyClient(discord.Client):
         logging.info("Web.py server has been started.")
 
     async def on_message(self, message):
-        logging.debug(f"Message received: {message.content} from {message.author}")
         if message.author == self.user:
-            logging.debug("Message from bot itself; ignoring.")
             return
         if not self.is_message_in_specific_channel(message):
-            logging.debug(f"Message not in the specific channel or its thread: {message.channel}")
             return
         if self.is_processing:
-            logging.debug("Already processing another message.")
             return
-        logging.debug("Processing message...")
         self.is_processing = True
         try:
-            response = await self.generate_response(message)  # Note the use of self here
+            response = await self.generate_response(message)
             await message.channel.send(response)
-        except Exception as e:
-            logging.error(f"Error during message processing: {e}", exc_info=True)
         finally:
             self.is_processing = False
 
     def is_message_in_specific_channel(self, message):
-        result = message.channel.id == SPECIFIC_CHANNEL_ID or (
+        return message.channel.id == SPECIFIC_CHANNEL_ID or (
             isinstance(message.channel, discord.Thread) and message.channel.parent_id == SPECIFIC_CHANNEL_ID
         )
-        logging.debug(f"Checking if message is in specific channel: {result}")
-        return result
 
     async def generate_response(self, message):
-        global conversation_history
         user_input = message.content
         user_mention = message.author.mention
         system_message = f"{user_mention}, 법률에 대한 답변을 제공하는 어시스턴트입니다."
@@ -85,17 +84,15 @@ class MyClient(discord.Client):
 
         answer = self.search_in_dataset(user_input, law_dataset)
 
-        full_response_text = system_message + "\n\n" + answer
+        full_response_text = system_prefix + "\n\n" + answer
         logging.debug(f'Full model response: {full_response_text}')
 
         conversation_history.append({"role": "assistant", "content": full_response_text})
         return f"{user_mention}, {full_response_text}"
 
     def search_in_dataset(self, query, dataset):
-        # 여기에서 query에 따라 데이터셋에서 적절한 정보를 검색하는 로직을 구현합니다.
-        # 예제는 단순화를 위해 첫 번째 항목을 반환합니다.
-        return dataset['train'][0]['document']  # 'text' 대신 'document'로 변경
-
+        # 첫 번째 항목 반환 (필드 이름 확인 필요)
+        return dataset['train'][0]['판결요지']  # 'text' 대신 실제 필드 이름 '판결요지' 사용
 
 if __name__ == "__main__":
     discord_client = MyClient(intents=intents)