Spaces:
Sleeping
Sleeping
import discord | |
import logging | |
import os | |
from huggingface_hub import InferenceClient | |
import asyncio | |
import subprocess | |
from datasets import load_dataset | |
# ํ์ฌ ์์ ๋๋ ํ ๋ฆฌ ์ถ๋ ฅ | |
print("Current Working Directory:", os.getcwd()) | |
# ๋ฐ์ดํฐ์ ํ์ผ ์ด๋ฆ | |
data_files = ['train_0.csv', 'train_1.csv', 'train_2.csv', 'train_3.csv', 'train_4.csv', 'train_5.csv'] | |
# ํ์ฌ ์์ ๋๋ ํ ๋ฆฌ์ ๋ชจ๋ ํ์ผ์ด ์๋์ง ํ์ธ | |
missing_files = [file for file in data_files if not os.path.exists(file)] | |
if missing_files: | |
print(f"Missing files: {missing_files}") | |
# ํ์ํ ๊ฒฝ์ฐ ์์ ๋๋ ํ ๋ฆฌ ๋ณ๊ฒฝ | |
os.chdir('/home/user/app') | |
print("Changed directory to:", os.getcwd()) | |
else: | |
print("All files are present in the current directory.") | |
# ๋ฐ์ดํฐ์ ๋ก๋ | |
law_dataset = load_dataset('csv', data_files=data_files) | |
print("Dataset loaded successfully.") | |
# ๋ก๊น ์ค์ | |
logging.basicConfig(level=logging.DEBUG, format='%(asctime)s:%(levelname)s:%(name)s: %(message)s', handlers=[logging.StreamHandler()]) | |
# ์ธํ ํธ ์ค์ | |
intents = discord.Intents.default() | |
intents.message_content = True | |
intents.messages = True | |
intents.guilds = True | |
intents.guild_messages = True | |
# ์ถ๋ก API ํด๋ผ์ด์ธํธ ์ค์ | |
hf_client = InferenceClient("CohereForAI/c4ai-command-r-plus", token=os.getenv("HF_TOKEN")) | |
# ํน์ ์ฑ๋ ID | |
SPECIFIC_CHANNEL_ID = int(os.getenv("DISCORD_CHANNEL_ID")) | |
# ๋ํ ํ์คํ ๋ฆฌ๋ฅผ ์ ์ฅํ ์ ์ญ ๋ณ์ | |
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}๋ก ๋ก๊ทธ์ธ๋์์ต๋๋ค!') | |
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: | |
logging.debug("Currently processing another message, skipping this one.") | |
return | |
self.is_processing = True | |
try: | |
response = await generate_response(message) | |
if response.strip() == "": | |
response = "์ฃ์กํฉ๋๋ค, ์ ๊ณตํ ์ ์๋ ์ ๋ณด๊ฐ ์์ต๋๋ค." | |
await message.channel.send(response) | |
finally: | |
self.is_processing = False | |
logging.debug("Message processing completed, ready for the next one.") | |
def is_message_in_specific_channel(self, message): | |
channel_condition = message.channel.id == SPECIFIC_CHANNEL_ID | |
thread_condition = isinstance(message.channel, discord.Thread) and message.channel.parent_id == SPECIFIC_CHANNEL_ID | |
return channel_condition or thread_condition | |
async def generate_response(message): | |
global conversation_history | |
case_number = message.content.strip() | |
user_mention = message.author.mention | |
system_message = f"{user_mention}, ๋ค์์ ์์ฒญํ์ ์ฌ๊ฑด์ ์ ๋ฌธ์ ๋๋ค:" | |
# ๋ฐ์ดํฐ ๊ฒ์ ๋ฐ ์๋ต ์ค๋น | |
full_text = search_in_dataset_by_case_number(case_number, law_dataset) | |
full_response_text = f"{system_message}\n\n{full_text}" | |
max_length = 2000 | |
if len(full_response_text) > max_length: | |
for i in range(0, len(full_response_text), max_length): | |
part_response = full_response_text[i:i+max_length] | |
await message.channel.send(part_response) | |
logging.debug(f'Full model response sent: {full_response_text}') | |
conversation_history.append({"role": "assistant", "content": full_response_text}) | |
else: | |
await message.channel.send(full_response_text) | |
logging.debug(f'Full model response sent: {full_response_text}') | |
conversation_history.append({"role": "assistant", "content": full_response_text}) | |
def search_in_dataset(query, dataset): | |
for record in dataset['train']: | |
# '์ฌ๊ฑด๋ช '์ด None์ด ์๋์ง ํ์ธํ๊ณ , query๊ฐ ํด๋น ๊ฐ์ ํฌํจ๋์ด ์๋์ง ๊ฒ์ฌ | |
if record['์ฌ๊ฑด๋ช '] and query in record['์ฌ๊ฑด๋ช ']: | |
return record['์ฌ๊ฑด๋ฒํธ'] | |
return "๊ด๋ จ ๋ฒ๋ฅ ์ ๋ณด๋ฅผ ์ฐพ์ ์ ์์ต๋๋ค." | |
def search_in_dataset_by_case_number(case_number, dataset): | |
for record in dataset['train']: | |
if record['์ฌ๊ฑด๋ฒํธ'] == case_number: | |
return record.get('์ ๋ฌธ', "ํด๋น ์ฌ๊ฑด์ ๋ํ ์ ๋ฌธ ์ ๋ณด๊ฐ ์์ต๋๋ค.") | |
return "๊ด๋ จ ๋ฒ๋ฅ ์ ๋ณด๋ฅผ ์ฐพ์ ์ ์์ต๋๋ค." | |
if __name__ == "__main__": | |
discord_client = MyClient(intents=intents) | |
discord_client.run(os.getenv('DISCORD_TOKEN')) | |