hf_argparser: fix TypeError when resolving Optional of generic type

[Dev-X25874]

What does this PR do?

Fixes a TypeError in HfArgumentParser._parse_dataclass_field when a dataclass field is typed as Optional[X] where X is a subscripted generic (e.g. Optional[List[str]], Optional[list[str]], Optional[Dict[str, int]]).

The root cause is on line 185, where isinstance(None, field.type.__args__[1]) is used to detect whether the second type argument is NoneType. This works for plain classes like int or str, but raises TypeError: Subscripted generics cannot be used with class and instance checks when args[1] is a generic alias — which happens when None appears first in the union, e.g. Union[None, List[str]].

The fix replaces the isinstance call with an identity check: field.type.__args__[1] is type(None). This is the correct, idiomatic way to test for NoneType and mirrors the pattern already used two lines above (type(None) not in field.type.__args__). It is safe for all types including generics, and produces identical results for plain classes where isinstance happened to work.

Fixes # (issue)

Code Agent Policy

• I confirm that this is not a pure code agent PR.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline, Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the documentation guidelines, and here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

@ArthurZucker

[github-actions]

This PR was flagged by our automated quality checks. If you're a genuine
contributor, please reply here and a maintainer will review your PR.

Common reasons for flagging:

• New GitHub account
• Unusually high number of repository forks in a 24-hour window

We appreciate your contribution and apologize if this is a false positive!

[Dev-X25874]

Hi, I understand the automated flag and appreciate the bot catching low-quality submissions — the repo clearly needs it.

To clarify: I found this bug manually by reading through hf_argparser.py while exploring the codebase as part of my open-source learning. The specific issue caught my eye because isinstance(None, X) is a known Python footgun with generics. I verified it by running a local Python snippet before writing a single line of the fix.

I'm happy to answer any questions about the bug or the reasoning behind the fix to demonstrate this. The change is a single expression on line 185 — genuinely as small as it looks.