ru-chat-bot

English | Русский

🔍 About

This is a modular deep-learning-based chatbot designed for natural language understanding and day-to-day task automation. Unlike simple script-based bots, "Neuro" (Нейро) analyzes user intent to provide contextually relevant responses.

🚀 Key Technologies

The project is built on a modern Natural Language Processing (NLP) stack:

• Transformers (RuBERT): Text vectorization is powered by the rubert-tiny2 model, optimized via ONNX Runtime for lightning-fast performance even on standard CPUs.
• NLU & Intent Classification: A custom neural network intent classifier with a ResNet-inspired architecture that determines exactly what the user wants.
• Semantic Search (DSSM): The "Small Talk" module utilizes FAISS vector search to retrieve the most semantically appropriate responses for casual conversation.
• NER (Named Entity Recognition): Entity extraction is used for accurate weather forecasting and Wikipedia information retrieval.

🛠 Core Features

1. Weather: Real-time forecasts for any city in Russia, supporting date-specific queries.
2. News: Latest news headlines fetched in real-time.
3. Wikipedia: Instant retrieval of definitions and biographies directly within the chat.
4. Smart Dialogue: Ability to maintain a simple, natural conversation.
5. Safety: Integrated toxicity detector to filter out unwanted or harmful content.

📚 Usage

1. Environment Setup

To run the bot, create a .env file in the project root and add your API keys:

TELEGRAM_BOT_TOKEN=YOUR_TELEGRAM_TOKEN
WEATHER_API_KEY=YOUR_OPENWEATHERMAP_KEY
# You can delete the line below to disable
PROXY_URL=http://127.0.0.1:12345

2. Launching the Bot

Example of a Telegram bot implementation using aiogram 3.x:

import asyncio
import sys
import os
import logging
import aiohttp
from aiogram import Bot, Dispatcher, types, F
from aiogram.filters import Command
from aiogram.client.default import DefaultBotProperties
from aiogram.client.session.aiohttp import AiohttpSession
from dotenv import load_dotenv
from src.chat_bot import ChatBot

if sys.platform == 'win32':
    asyncio.set_event_loop_policy(asyncio.WindowsSelectorEventLoopPolicy())

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s | %(levelname)-8s | %(name)-25s | %(message)s",
    handlers=[
        logging.StreamHandler(sys.stdout),
        logging.FileHandler("chat-bot.log", encoding="utf-8")
    ]
)
logger = logging.getLogger(__name__)

load_dotenv()

dp = Dispatcher()


@dp.message(Command("start"))
async def cmd_start(message: types.Message):
    logger.info(f"User {message.from_user.id} started the bot")
    await message.answer("Напишите что-нибудь, чтобы начать беседу...")


@dp.message(F.text)
async def handle_text(message: types.Message, chat_bot: ChatBot):
    waiting_msg = await message.answer("Пожалуйста, подождите ⏳")
    response = await chat_bot.handle_message(message.from_user.id, message.text)
    try:
        await waiting_msg.edit_text(response)
    except Exception as e:
        logger.error(f"Ошибка отправки: {e}")


async def on_startup():
    logger.info("Бот успешно запущен и готов к работе.")


async def main():
    dp.startup.register(on_startup)
    proxy_url = os.getenv("PROXY_URL")
    async with aiohttp.ClientSession() as global_http_session:
        aiogram_session = AiohttpSession(proxy=proxy_url)
        bot = Bot(
            token=os.getenv("TELEGRAM_BOT_TOKEN"),
            session=aiogram_session,
            default=DefaultBotProperties(parse_mode="HTML", link_preview_is_disabled=True)
        )
        chat_bot = ChatBot(http_session=global_http_session)
        dp["chat_bot"] = chat_bot
        await bot.delete_webhook(drop_pending_updates=True)
        try:
            await dp.start_polling(bot)
        except Exception as e:
            logger.critical(f"Ошибка: {e}", exc_info=True)
        finally:
            await bot.session.close()
            logger.info("Сессия бота закрыта.")


if __name__ == "__main__":
    try:
        asyncio.run(main())
    except (KeyboardInterrupt, SystemExit):
        logger.warning("Бот остановлен вручную.")

📥 Installation

pip install git+https://github.com/KvaytG/ru-chat-bot.git

📝 License

Licensed under the PolyForm Noncommercial license.

This project uses open-source components. For license details see pyproject.toml and dependencies' official websites.