(refactor) disable listen for subscriptions for Bitfinex

[0xferit]

Before submitting this PR, please make sure:

• Your code builds clean without any errors or warnings
• You are using approved title ("feat/", "fix/", "docs/", "refactor/")

A description of the changes proposed in the pull request:

The following method (listen_for_subscriptions) is inherited but it's inner dependencies are not implemented in Bitfinex connector, thus it's permanently erroring ("Unexpected error occurred when listening to order book streams. Retrying in 5 seconds...").

hummingbot/hummingbot/core/data_type/order_book_tracker_data_source.py

Lines 70 to 91 in cdfc1f8

	async def listen_for_subscriptions(self):
	"""
	Connects to the trade events and order diffs websocket endpoints and listens to the messages sent by the
	exchange. Each message is stored in its own queue.
	"""
	ws: Optional[WSAssistant] = None
	while True:
	try:
	ws: WSAssistant = await self._connected_websocket_assistant()
	await self._subscribe_channels(ws)
	await self._process_websocket_messages(websocket_assistant=ws)
	except asyncio.CancelledError:
	raise
	except ConnectionError as connection_exception:
	self.logger().warning(f"The websocket connection was closed ({connection_exception})")
	except Exception:
	self.logger().exception(
	"Unexpected error occurred when listening to order book streams. Retrying in 5 seconds...",
	)
	await self._sleep(1.0)
	finally:
	await self._on_order_stream_interruption(websocket_assistant=ws)

Moreover, it's not only error messages that is annoying, but somehow, it causes CPU utilization to max out and break the bot after several hours. I suspect the root cause is the execution logic repeatedly entering listen_for_subscriptions and repeatedly throwing an exception. Seems like a busy-wait issue.

By disabling the execution of listen_for_subscription for Bitfinex, I observed the CPU max-out problem disappear (given that there are no other error message spamming). Moreover, disabling this function seems to cause zero negative effect since the Bitfinex logic does not require listen_for_subscription.

Previously, I tried to implement this function instead of disabling it, but I haven't received adequate support from the team, and without pull request reviews, I was not able to complete it. @yancong001 tipped the idea of disabling the function instead, so I tried it.

Tests performed by the developer:

After disabling the function, the CPU max-out problem disappeared, given that there was no other error spamming. To test, I run PMM on Bitfinex in a Digital Ocean droplet with 4 vCPUs and 8GB of RAM for 3 days. As opposed to v1.23.0, which almost always crashes within the first day, this fork run smoothly for 3 days and counting.

Tips for QA testing:

Run PMM on Bitfinex, compare the latest release versus this fork, and observe CPU utilization.

Fixes #5919.

[0xferit]

UPDATE: Around the 3.5-day mark, CPU shot up to 100% again. Fuck that really, I thought it was solved.