Add retry logic to download_and_extract_meshes

[gbocchio-bdai]

Description

This PR adds a _request_with_retry function to judo/utils/assets.py and uses it to download the meshes; essentially making download_and_extract_meshes more resilient in CI.

It's not infrequent for one given HTTP request to fail in CI due to network instability. Adding a few retries solves most transient failures.

[dta-bdai]

is it in seconds or minutes? Consider specify the unit in the variable name?

[gbocchio-bdai]

Done!

[dta-bdai]

consider removing this if.. raise: Other raise statements below won't be reached if we raise here.

[gbocchio-bdai]

I removed the if ... raise. I added a similar if to prevent sleeping on the last retry if it fails.

[dta-bdai]

Thanks for the update! One more request change to simplify the code. Can you please take a look?

[dta-bdai]

this line won't never been reached.

[dta-bdai]

Suggested change

	last_exception: Exception | None = None
	for attempt in range(retries):
	for attempt in range(retries + 1):
	try:
	response = requests.request(method, url, timeout=timeout, **request_kwargs)
	response = requests.request(method, url, timeout=timeout_s, **request_kwargs)
	response.raise_for_status()
	return response
	except (requests.ConnectionError, requests.Timeout, requests.HTTPError) as exc:
	if attempt >= retries - 1:
	raise
	last_exception = exc
	time.sleep(backoff_time)
	except (requests.ConnectionError, requests.Timeout, requests.HTTPError) as request_exception:
	last_exception = request_exception
	if attempt < retries:
	time.sleep(backoff_time_s)
	if last_exception is not None:
	raise last_exception
	raise RuntimeError("Unexpected retry loop termination in _request_with_retry")
	for attempt in range(retries):
	try:
	response = requests.request(method, url, timeout=timeout, **request_kwargs)
	response.raise_for_status()
	return response
	except (requests.ConnectionError, requests.Timeout, requests.HTTPError) as request_exception:
	if attempt < retries:
	time.sleep(backoff_time_s)
	continue
	raise request_exception

[dta-bdai]

can you try the suggestion?