Yuriy Yuriev
264 lines
8.5 KiB
Python
264 lines
8.5 KiB
Python
"""
|
|
GraphQL WebSocket сервис для мониторинга Twitch чата.
|
|
Замена IRC — не требует OAuth токена.
|
|
"""
|
|
|
|
import asyncio
|
|
import json
|
|
import logging
|
|
import re
|
|
from typing import Optional, Callable, Awaitable, List
|
|
from datetime import datetime
|
|
from urllib.parse import urlparse
|
|
|
|
import aiohttp
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
_GQL_WS_URL = "wss://gql.twitch.tv/connection/websocket"
|
|
_CLIENT_ID = "kimne78kx3ncx6brgo4mv6wki5h1ko"
|
|
|
|
_CHAT_SUBSCRIPTION = """\
|
|
subscription ChatMessages($channelLogin: String!) {
|
|
chatEvents(input: { channelLogin: $channelLogin }) {
|
|
... on ChatMessageEvent {
|
|
message {
|
|
sender {
|
|
displayName
|
|
login
|
|
}
|
|
content {
|
|
text
|
|
fragments {
|
|
... on TextFragment {
|
|
text
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}"""
|
|
|
|
|
|
class TwitchGQLChatClient:
|
|
"""
|
|
Twitch чат клиент через GQL WebSocket.
|
|
Интерфейс совместим с TwitchIRCClient.
|
|
"""
|
|
|
|
def __init__(self, channel: str, target_username: str):
|
|
self.channel = channel.lower()
|
|
self.target_username = target_username.lower()
|
|
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(
|
|
_GQL_WS_URL,
|
|
protocols=["graphql-ws"],
|
|
headers={"Client-ID": _CLIENT_ID},
|
|
timeout=aiohttp.ClientTimeout(total=15),
|
|
)
|
|
|
|
await self._ws.send_str(json.dumps({
|
|
"type": "connection_init",
|
|
"payload": {
|
|
"Authorization": "undefined",
|
|
"Client-ID": _CLIENT_ID,
|
|
}
|
|
}))
|
|
|
|
ack = await asyncio.wait_for(self._ws.receive(), timeout=10)
|
|
msg = json.loads(ack.data)
|
|
if msg.get("type") != "connection_ack":
|
|
logger.error(f"GQL WS: unexpected init response: {msg}")
|
|
await self._cleanup()
|
|
return False
|
|
|
|
await self._ws.send_str(json.dumps({
|
|
"type": "start",
|
|
"id": "chat",
|
|
"payload": {
|
|
"query": _CHAT_SUBSCRIPTION,
|
|
"variables": {"channelLogin": self.channel},
|
|
"operationName": "ChatMessages",
|
|
}
|
|
}))
|
|
|
|
self._connected = True
|
|
logger.info(f"✅ GQL подключен к #{self.channel}")
|
|
return True
|
|
|
|
except Exception as e:
|
|
logger.error(f"❌ GQL ошибка подключения: {e}")
|
|
await self._cleanup()
|
|
return False
|
|
|
|
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 _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}
|
|
start_time = datetime.now()
|
|
reconnect_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"GQL connect failed for #{self.channel}")
|
|
|
|
while duration == 0 or (datetime.now() - start_time).seconds < duration:
|
|
if is_active and not is_active():
|
|
logger.info(f"⏸️ #{self.channel} пауза...")
|
|
await self._cleanup()
|
|
while is_active and not is_active():
|
|
await asyncio.sleep(2)
|
|
logger.info(f"▶️ #{self.channel} resume, переподключение...")
|
|
if not await self.connect():
|
|
reconnect_errors += 1
|
|
if reconnect_errors >= 3 and on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
reconnect_errors = 0
|
|
stats["reconnects"] += 1
|
|
continue
|
|
|
|
if not self._connected or self._ws is None or self._ws.closed:
|
|
await asyncio.sleep(5)
|
|
if not await self.connect():
|
|
reconnect_errors += 1
|
|
if reconnect_errors >= 3 and on_drain_fail:
|
|
await on_drain_fail()
|
|
return stats
|
|
else:
|
|
reconnect_errors = 0
|
|
stats["reconnects"] += 1
|
|
continue
|
|
|
|
try:
|
|
raw = await asyncio.wait_for(self._ws.receive(), timeout=30.0)
|
|
except asyncio.CancelledError:
|
|
raise
|
|
except asyncio.TimeoutError:
|
|
continue
|
|
except Exception as e:
|
|
logger.warning(f"GQL 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
|
|
|
|
try:
|
|
msg = json.loads(raw.data)
|
|
except Exception:
|
|
continue
|
|
|
|
msg_type = msg.get("type")
|
|
|
|
if msg_type == "ka": # keep-alive от Twitch
|
|
continue
|
|
|
|
if msg_type == "connection_error":
|
|
logger.error(f"GQL connection error: {msg}")
|
|
self._connected = False
|
|
continue
|
|
|
|
if msg_type != "data":
|
|
continue
|
|
|
|
payload = msg.get("payload", {})
|
|
data = payload.get("data", {})
|
|
chat_event = data.get("chatEvents")
|
|
if not chat_event:
|
|
continue
|
|
|
|
message_obj = chat_event.get("message", {})
|
|
if not message_obj:
|
|
continue
|
|
|
|
sender = message_obj.get("sender", {})
|
|
display_name = sender.get("displayName", "unknown")
|
|
content = message_obj.get("content", {})
|
|
text = content.get("text", "")
|
|
|
|
if not text.strip():
|
|
continue
|
|
|
|
stats["messages"] += 1
|
|
logger.debug(f"GQL message from {display_name}: {text[:50]}...")
|
|
|
|
for url in self._extract_urls(text):
|
|
if self._is_domain_allowed(url, allowed_domains):
|
|
stats["links_found"] += 1
|
|
try:
|
|
await on_url_found(url, display_name)
|
|
except asyncio.CancelledError:
|
|
raise
|
|
except Exception:
|
|
pass
|
|
|
|
return stats
|
|
|
|
async def disconnect(self) -> None:
|
|
await self._cleanup()
|