feat: Updating Llama Stack ENV var names

[vladimirivic]

What does this PR do?

We are updating ENV var names to follow shorter naming pattern, the env vars will be prefixed by LLAMA_STACK instead of LLAMA_STACK_CLIENT which was previously auto-generated by Stainless by reading the project name field in the spec, but Stainless added a new feature where we can override this prefix.

Test Plan

export LLAMA_STACK_BASE_URL=<base-url>
export LLAMA_STACK_API_KEY=<api-key>

from llama_stack_client import LlamaStackClient
client = LlamaStackClient()

# will create a client pointing to <base-url>, with <api-key>.

[raghotham]

we should probably allow for both env vars and deprecate the older one - this is not a backward compatible change

[ashwinb]

this code is auto-generated @raghotham

[ashwinb]

unless we do more work, you aren't going to get back-compat. I don't think this one is worth it at all.

[vladimirivic]

I could extend the sync script and add a manual modification of lines like os.environ.get("LLAMA_STACK_API_KEY") to try the new then the old env var and we keep that for a while. I evaluated the risk of breaking things was small but we can be extra cautious if we think there is a chance something breaks.

[raghotham]

ok

[raghotham]

Not needed. Overwritten by #157