Yuriy Yuriev
316 lines
12 KiB
Python
316 lines
12 KiB
Python
"""
|
|
IRC сервис для мониторинга Twitch чата.
|
|
"""
|
|
|
|
import asyncio
|
|
import re
|
|
import logging
|
|
from typing import Optional, Callable, Awaitable, List
|
|
from datetime import datetime
|
|
from urllib.parse import urlparse
|
|
|
|
from config.settings import settings
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
class TwitchIRCClient:
|
|
"""
|
|
Twitch IRC клиент для мониторинга чата.
|
|
"""
|
|
|
|
def __init__(
|
|
self,
|
|
channel: str,
|
|
target_username: str,
|
|
irc_username: str = None,
|
|
irc_oauth: str = None
|
|
):
|
|
self.channel = channel.lower()
|
|
self.target_username = target_username.lower()
|
|
self.irc_username = irc_username or settings.IRC_USERNAME
|
|
self.irc_oauth = irc_oauth or settings.IRC_OAUTH
|
|
|
|
self._reader: Optional[asyncio.StreamReader] = None
|
|
self._writer: Optional[asyncio.StreamWriter] = None
|
|
self._connected = False
|
|
self._url_pattern = re.compile(r'https?://[^\s<>"]+')
|
|
|
|
async def connect(self) -> bool:
|
|
"""Подключение к IRC серверу."""
|
|
try:
|
|
import ssl
|
|
ssl_context = ssl.create_default_context()
|
|
self._reader, self._writer = await asyncio.wait_for(
|
|
asyncio.open_connection(
|
|
settings.IRC_SERVER, settings.IRC_PORT,
|
|
ssl=ssl_context
|
|
),
|
|
timeout=15.0
|
|
)
|
|
|
|
await self._send(f"PASS {self.irc_oauth}")
|
|
await self._send(f"NICK {self.irc_username}")
|
|
await self._send("CAP REQ :twitch.tv/tags")
|
|
await self._send("CAP REQ :twitch.tv/commands")
|
|
await self._send(f"JOIN #{self.channel}")
|
|
|
|
await asyncio.sleep(2)
|
|
self._connected = True
|
|
logger.info(f"✅ Подключен к #{self.channel}")
|
|
return True
|
|
|
|
except Exception as e:
|
|
logger.error(f"❌ Ошибка подключения: {e}")
|
|
return False
|
|
|
|
async def _send(self, message: str) -> None:
|
|
"""Отправка сообщения."""
|
|
if self._writer:
|
|
self._writer.write(f"{message}\r\n".encode())
|
|
await self._writer.drain()
|
|
|
|
async def _read_line(self) -> Optional[str]:
|
|
"""Чтение строки с обработкой PING."""
|
|
if not self._reader:
|
|
return None
|
|
|
|
try:
|
|
line = await self._reader.readline()
|
|
|
|
if not line:
|
|
logger.warning("Empty line - connection closed")
|
|
self._connected = False
|
|
return None
|
|
|
|
decoded = line.decode('utf-8', errors='ignore').strip()
|
|
|
|
# PING - отвечаем СРАЗУ
|
|
if decoded.startswith('PING'):
|
|
pong_msg = f"PONG {decoded.split()[1]}"
|
|
self._writer.write(f"{pong_msg}\r\n".encode())
|
|
try:
|
|
await self._writer.drain()
|
|
except (ConnectionResetError, BrokenPipeError, OSError) as e:
|
|
logger.warning(f"Failed to send PONG: {e}")
|
|
logger.debug(f"🏓 {pong_msg}")
|
|
return f"PONG {decoded.split()[1]}" # Возвращаем PONG для отслеживания
|
|
|
|
return decoded
|
|
|
|
except Exception as e:
|
|
logger.error(f"Read error: {e}")
|
|
self._connected = False
|
|
return None
|
|
|
|
|
|
def _parse_message(self, line: str) -> Optional[dict]:
|
|
if not line or 'PRIVMSG' not in line:
|
|
return None
|
|
|
|
try:
|
|
# Парсим display-name из тегов
|
|
tags = {}
|
|
if line.startswith('@'):
|
|
tags_part = line.split(' ', 1)[0]
|
|
for tag in tags_part.lstrip('@').split(';'):
|
|
if '=' in tag:
|
|
k, v = tag.split('=', 1)
|
|
tags[k] = v
|
|
|
|
display_name = tags.get('display-name', '')
|
|
|
|
# Извлекаем сообщение
|
|
if ' :' in line:
|
|
message_text = line.split(' :', 1)[1]
|
|
else:
|
|
message_text = ''
|
|
|
|
# Пропускаем пустые сообщения
|
|
if not message_text.strip():
|
|
return None
|
|
|
|
return {
|
|
'display_name': display_name or 'unknown',
|
|
'message': message_text,
|
|
'tags': tags
|
|
}
|
|
except Exception as e:
|
|
logger.error(f"Parse error: {e}")
|
|
return None
|
|
|
|
|
|
def _extract_urls(self, text: str) -> List[str]:
|
|
"""Извлечение URL из текста."""
|
|
urls = []
|
|
for raw_url in self._url_pattern.findall(text):
|
|
try:
|
|
parsed = urlparse(raw_url)
|
|
clean = f"{parsed.scheme}://{parsed.netloc}{parsed.path}"
|
|
if parsed.query:
|
|
clean += f"?{parsed.query}"
|
|
urls.append(clean)
|
|
except Exception:
|
|
urls.append(raw_url.strip())
|
|
return urls
|
|
|
|
def _is_domain_allowed(self, url: str, allowed_domains: List[str]) -> bool:
|
|
"""Проверка домена."""
|
|
if not allowed_domains:
|
|
return True
|
|
try:
|
|
domain = urlparse(url).netloc.lower()
|
|
return any(d.lower() in domain for d in allowed_domains)
|
|
except Exception:
|
|
return False
|
|
|
|
async def listen_for_messages(
|
|
self,
|
|
on_url_found: Callable[[str, str], Awaitable[None]],
|
|
duration: int,
|
|
allowed_domains: Optional[List[str]] = None,
|
|
is_active: Callable[[], bool] = None,
|
|
on_drain_fail: Callable[[], Awaitable[None]] = None,
|
|
) -> dict:
|
|
stats = {"links_found": 0, "messages": 0, "reconnects": 0}
|
|
|
|
start_time = datetime.now()
|
|
last_message_time = datetime.now()
|
|
drain_errors = 0
|
|
|
|
# Если задача стартует на паузе — ждём resume до подключения к IRC
|
|
if is_active and not is_active():
|
|
logger.info(f"⏸️ #{self.channel} waiting for resume before connecting...")
|
|
while is_active and not is_active():
|
|
await asyncio.sleep(2)
|
|
if not is_active and not is_active():
|
|
return stats
|
|
logger.info(f"▶️ #{self.channel} resumed, connecting to IRC")
|
|
|
|
if not await self.connect():
|
|
raise ConnectionError(f"IRC connect failed for #{self.channel}")
|
|
|
|
while duration == 0 or (datetime.now() - start_time).seconds < duration:
|
|
# === ПРОВЕРКА АКТИВНОСТИ ===
|
|
if is_active and not is_active():
|
|
logger.info(f"⏸️ #{self.channel} paused, waiting for resume...")
|
|
await self.disconnect()
|
|
while is_active and not is_active():
|
|
await asyncio.sleep(2)
|
|
logger.info(f"▶️ #{self.channel} resumed, reconnecting")
|
|
await self.connect()
|
|
last_message_time = datetime.now()
|
|
continue
|
|
|
|
# Нет соединения — ждём перед повтором
|
|
if not self._connected:
|
|
await asyncio.sleep(5)
|
|
await self.connect()
|
|
continue
|
|
|
|
# === ПЕРЕПОДКЛЮЧЕНИЕ КАЖДЫЕ 30 СЕКУНД ===
|
|
if (datetime.now() - last_message_time).seconds > 30:
|
|
logger.info("🔄 Reconnecting (30s)...")
|
|
|
|
# Сначала читаем всё что осталось (с retry)
|
|
drain_ok = False
|
|
for attempt in range(3):
|
|
try:
|
|
while True:
|
|
line = await asyncio.wait_for(self._reader.readline(), timeout=1.0)
|
|
if not line:
|
|
break
|
|
decoded = line.decode('utf-8', errors='ignore').strip()
|
|
if 'PRIVMSG' in decoded:
|
|
msg = self._parse_message(decoded)
|
|
if msg:
|
|
stats["messages"] += 1
|
|
for url in self._extract_urls(msg['message']):
|
|
if self._is_domain_allowed(url, allowed_domains):
|
|
stats["links_found"] += 1
|
|
try:
|
|
await on_url_found(url, msg['display_name'])
|
|
except Exception as e:
|
|
logger.warning(f"URL callback error: {e}")
|
|
drain_ok = True
|
|
break
|
|
except asyncio.TimeoutError:
|
|
drain_ok = True
|
|
break
|
|
except Exception as e:
|
|
logger.warning(f"Error draining IRC data (attempt {attempt + 1}/3): {e}")
|
|
await asyncio.sleep(2)
|
|
|
|
if not drain_ok:
|
|
drain_errors += 1
|
|
logger.error(f"IRC drain failed 3 times, total failures: {drain_errors}")
|
|
if drain_errors >= 3:
|
|
logger.error("IRC drain failed 3 consecutive times — pausing task")
|
|
if on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
drain_errors = 0
|
|
|
|
# Переподключаемся
|
|
try:
|
|
await self.disconnect()
|
|
except Exception as e:
|
|
logger.warning(f"Disconnect error: {e}")
|
|
await asyncio.sleep(1)
|
|
await self.connect()
|
|
|
|
stats["reconnects"] += 1
|
|
last_message_time = datetime.now()
|
|
continue
|
|
|
|
# Читаем строку
|
|
try:
|
|
line = await asyncio.wait_for(self._read_line(), timeout=5.0)
|
|
except asyncio.CancelledError:
|
|
raise
|
|
except asyncio.TimeoutError:
|
|
continue
|
|
except Exception as e:
|
|
logger.warning(f"IRC read error: {e}")
|
|
continue
|
|
|
|
if not line:
|
|
await asyncio.sleep(0.5)
|
|
continue
|
|
|
|
if line.startswith('PONG'):
|
|
continue
|
|
|
|
msg = self._parse_message(line)
|
|
if not msg:
|
|
continue
|
|
|
|
last_message_time = datetime.now() # Обновляем таймер
|
|
stats["messages"] += 1
|
|
logger.debug(f"IRC message from {msg['display_name']}: {msg['message'][:50]}...")
|
|
|
|
for url in self._extract_urls(msg['message']):
|
|
if self._is_domain_allowed(url, allowed_domains):
|
|
stats["links_found"] += 1
|
|
try:
|
|
await on_url_found(url, msg['display_name'])
|
|
except asyncio.CancelledError:
|
|
raise
|
|
except Exception:
|
|
pass
|
|
|
|
return stats
|
|
|
|
|
|
async def disconnect(self) -> None:
|
|
"""Закрытие соединения."""
|
|
self._connected = False
|
|
if self._writer:
|
|
try:
|
|
self._writer.close()
|
|
await self._writer.wait_closed()
|
|
except (ConnectionResetError, BrokenPipeError, OSError) as e:
|
|
logger.debug(f"Disconnect warning: {e}")
|
|
self._writer = None
|
|
self._reader = None |