Yuriy Yuriev
265 lines
9.3 KiB
Python
265 lines
9.3 KiB
Python
"""
|
|
Twitch чат через IRC over WebSocket (wss://irc-ws.chat.twitch.tv:443).
|
|
Замена raw TCP IRC — тот же протокол, WebSocket транспорт.
|
|
"""
|
|
|
|
import asyncio
|
|
import logging
|
|
import re
|
|
from typing import Optional, Callable, Awaitable, List
|
|
from datetime import datetime
|
|
from urllib.parse import urlparse
|
|
|
|
import aiohttp
|
|
|
|
from config.settings import settings
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
_IRC_WS_URL = "wss://irc-ws.chat.twitch.tv:443"
|
|
|
|
|
|
class TwitchGQLChatClient:
|
|
"""
|
|
Twitch IRC over WebSocket клиент.
|
|
Интерфейс совместим с TwitchIRCClient.
|
|
"""
|
|
|
|
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._ws: Optional[aiohttp.ClientWebSocketResponse] = None
|
|
self._session: Optional[aiohttp.ClientSession] = None
|
|
self._connected = False
|
|
self._url_pattern = re.compile(r'https?://[^\s<>"]+')
|
|
|
|
async def connect(self) -> bool:
|
|
try:
|
|
self._session = aiohttp.ClientSession()
|
|
self._ws = await self._session.ws_connect(
|
|
_IRC_WS_URL,
|
|
timeout=aiohttp.ClientTimeout(total=15),
|
|
)
|
|
|
|
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"✅ IRC WS подключен к #{self.channel}")
|
|
return True
|
|
|
|
except Exception as e:
|
|
logger.error(f"❌ IRC WS ошибка подключения: {e}")
|
|
await self._cleanup()
|
|
return False
|
|
|
|
async def _send(self, message: str) -> None:
|
|
if self._ws and not self._ws.closed:
|
|
await self._ws.send_str(f"{message}\r\n")
|
|
|
|
async def _cleanup(self) -> None:
|
|
self._connected = False
|
|
try:
|
|
if self._ws and not self._ws.closed:
|
|
await self._ws.close()
|
|
except Exception:
|
|
pass
|
|
try:
|
|
if self._session and not self._session.closed:
|
|
await self._session.close()
|
|
except Exception:
|
|
pass
|
|
self._ws = None
|
|
self._session = None
|
|
|
|
def _parse_message(self, line: str) -> Optional[dict]:
|
|
if not line or "PRIVMSG" not in line:
|
|
return None
|
|
try:
|
|
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", "")
|
|
message_text = line.split(" :", 1)[1] if " :" in line else ""
|
|
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]:
|
|
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:
|
|
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}
|
|
last_message_time = datetime.now()
|
|
drain_errors = 0
|
|
|
|
if is_active and not is_active():
|
|
logger.info(f"⏸️ #{self.channel} ждём resume перед подключением...")
|
|
while is_active and not is_active():
|
|
await asyncio.sleep(2)
|
|
|
|
if not await self.connect():
|
|
raise ConnectionError(f"IRC WS connect failed for #{self.channel}")
|
|
|
|
# start_time после паузы и подключения — пауза не считается в длительность
|
|
start_time = datetime.now()
|
|
|
|
while duration == 0 or (datetime.now() - start_time).total_seconds() < duration:
|
|
if is_active and not is_active():
|
|
logger.info(f"⏸️ #{self.channel} пауза...")
|
|
await self._cleanup()
|
|
pause_begin = datetime.now()
|
|
while is_active and not is_active():
|
|
await asyncio.sleep(2)
|
|
# Исключаем время паузы из отсчёта длительности
|
|
start_time += datetime.now() - pause_begin
|
|
logger.info(f"▶️ #{self.channel} resume, переподключение...")
|
|
if not await self.connect():
|
|
drain_errors += 1
|
|
if drain_errors >= 3 and on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
drain_errors = 0
|
|
stats["reconnects"] += 1
|
|
last_message_time = datetime.now()
|
|
continue
|
|
|
|
if not self._connected or self._ws is None or self._ws.closed:
|
|
await asyncio.sleep(5)
|
|
if not await self.connect():
|
|
drain_errors += 1
|
|
if drain_errors >= 3 and on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
drain_errors = 0
|
|
stats["reconnects"] += 1
|
|
last_message_time = datetime.now()
|
|
continue
|
|
|
|
# Реконнект при отсутствии активности > 6 минут
|
|
if (datetime.now() - last_message_time).total_seconds() > 360:
|
|
logger.info("🔄 Нет активности 6 мин, переподключение...")
|
|
await self._cleanup()
|
|
await asyncio.sleep(1)
|
|
if not await self.connect():
|
|
drain_errors += 1
|
|
if drain_errors >= 3 and on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
drain_errors = 0
|
|
stats["reconnects"] += 1
|
|
last_message_time = datetime.now()
|
|
continue
|
|
|
|
try:
|
|
raw = await asyncio.wait_for(self._ws.receive(), timeout=5.0)
|
|
except asyncio.CancelledError:
|
|
raise
|
|
except asyncio.TimeoutError:
|
|
continue
|
|
except Exception as e:
|
|
logger.warning(f"IRC WS receive error: {e}")
|
|
self._connected = False
|
|
continue
|
|
|
|
if raw.type in (
|
|
aiohttp.WSMsgType.CLOSING,
|
|
aiohttp.WSMsgType.CLOSED,
|
|
aiohttp.WSMsgType.ERROR,
|
|
):
|
|
logger.warning(f"WS закрыт: {raw.type}")
|
|
self._connected = False
|
|
continue
|
|
|
|
if raw.type != aiohttp.WSMsgType.TEXT:
|
|
continue
|
|
|
|
line = raw.data.strip()
|
|
if not line:
|
|
continue
|
|
|
|
if line.startswith("PING"):
|
|
await self._send("PONG :tmi.twitch.tv")
|
|
last_message_time = datetime.now()
|
|
logger.debug("🏓 PONG")
|
|
continue
|
|
|
|
last_message_time = datetime.now()
|
|
|
|
msg = self._parse_message(line)
|
|
if not msg:
|
|
continue
|
|
|
|
stats["messages"] += 1
|
|
logger.debug(f"IRC WS msg 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:
|
|
await self._cleanup()
|