docs: add proxy_client to init_llm (harmonized model initialization)

[yamaceay]

References

• Harmonized Model Initialization.

What this PR does and why it is needed

Since 7.2.0+, some breaking changes have been introduced to the API. First of all, init_llm and init_embedding_model now require the model name / id to be passed (previously it was optional). Secondly, proxy_client is essential to perform further provider calls.

Currently, in docs, no proxy_client is passed through. This is why, in the underlying subclass call, proxy client is forwarded as None, so that we run into the error "NoneType has no such XXX attribute / property". For this, we should always instantiate proxy client using gen_ai_hub.proxy.core.get_proxy_client(...).

So the correct fix to the docs is to add:

from gen_ai_hub.proxy.core import get_proxy_client
proxy_client=get_proxy_client('gen-ai-hub')
llm = init_llm(..., proxy_client=proxy_client, ...)

Same applies for embeddings.

Definition of Done

• Code is tested (Unit, Integration, E2E)
• Error handling created / updated & covered by the tests above
• Documentation updated
  • Only Public APIs are allowed to be used in documentation/tutorials/sample code
• (Optional) Aligned changes with the JS/TS and Java SDK
• (Optional) Release notes updated -->

[yamaceay]

In my opinion, it would be better to mark this as a bug rather than another breaking change, because this is not intended to operate that way. So, we need to revert the change of breaking changes and add a new release note saying that the bug is now fixed.

[yamaceay]

Closing this PR because it should work as init_llm("model_id") without any proxy_client according to the spec. The problem is not in the docs, the problem is in the major release.