import discord import logging import os 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()]) # 인텐트 설정 intents = discord.Intents.default() intents.message_content = True intents.messages = True intents.guilds = True intents.guild_messages = True # 데이터셋 로드 data_files = ['train_0.csv', 'train_1.csv', 'train_2.csv', 'train_3.csv', 'train_4.csv', 'train_5.csv'] law_dataset = load_dataset('csv', data_files=data_files) # 추론 API 클라이언트 설정 hf_client = InferenceClient("CohereForAI/c4ai-command-r-plus", token=os.getenv("HF_TOKEN")) # 특정 채널 ID SPECIFIC_CHANNEL_ID = int(os.getenv("DISCORD_CHANNEL_ID")) # 대화 히스토리를 저장할 전역 변수 conversation_history = [] class MyClient(discord.Client): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.is_processing = False async def on_ready(self): logging.info(f'{self.user}로 로그인되었습니다!') subprocess.Popen(["python", "web.py"]) logging.info("Web.py server has been started.") async def on_message(self, message): if message.author == self.user: return if not self.is_message_in_specific_channel(message): return if self.is_processing: return self.is_processing = True try: response = await generate_response(message) await message.channel.send(response) finally: self.is_processing = False def is_message_in_specific_channel(self, message): return message.channel.id == SPECIFIC_CHANNEL_ID or ( isinstance(message.channel, discord.Thread) and message.channel.parent_id == SPECIFIC_CHANNEL_ID ) async def generate_response(message): global conversation_history user_input = message.content user_mention = message.author.mention system_message = f"{user_mention}, DISCORD에서 사용자들의 질문에 답하는 어시스턴트입니다." answer = search_in_dataset(user_input, law_dataset) full_response_text = system_message + "\n\n" + answer await message.channel.send(full_response_text) conversation_history.append({"role": "assistant", "content": full_response_text}) def search_in_dataset(query, dataset): response = [] for record in dataset['train']: if record['사건명'] and query in record['사건명']: # 사건명으로 검색 시 사건번호만 반환 response.append(f"사건번호: {record['사건번호']}") elif record['사건번호'] == query: # 사건번호로 검색 시 전체 내용 반환 detail = ( f"판례정보일련번호: {record['판례정보일련번호']}\n" f"사건명: {record['사건명']}\n" f"사건번호: {record['사건번호']}\n" f"선고일자: {record['선고일자']}\n" f"선고: {record['선고']}\n" f"법원명: {record['법원명']}\n" f"사건종류명: {record['사건종류명']}\n" f"판결유형: {record['판결유형']}\n" f"판시사항: {record['판시사항']}\n" f"판결요지: {record['판결요지']}\n" f"참조조문: {record['참조조문']}\n" f"참조판례: {record['참조판례']}\n" f"전문: {record['전문']}\n" ) response.append(detail) return "\n".join(response) if response else "관련 법률 정보를 찾을 수 없습니다." if __name__ == "__main__": discord_client = MyClient(intents=intents) discord_client.run(os.getenv('DISCORD_TOKEN'))