[FEATURE] Expose a transport-level cancel/close handle on async streaming responses

[adtyavrdhn]

Confirmation of Request Source

• I confirm that this feature request is for the xAI SDK Python library (e.g., client features, SDK methods, or documentation) and not for the xAI API itself (e.g., model capabilities or API changes).

Describe the feature you'd like

While working on pydantic/pydantic-ai#1524 I hit: Chat.stream() (in xai_sdk.aio.chat) only exposes the async generator, so cancelling it from another task while we're iterating raises RuntimeError: aclose(): asynchronous generator is already running.

A transport handle we could close/cancel directly — exposing the underlying UnaryStreamCall (or a wrapper with cancel() / close() on it), like the OpenAI (AsyncStream.close()) and
Anthropic (AsyncStream.close()) SDKs do. Would be nice to be able to do the same here.

Parallel issue on google-genai: googleapis/python-genai#2425

Additional Context

No response