Can't run Safety example

[tnsai]

I can run the examples (with some small fixes for the latest Llama Stack). However, the safety example errors at the point:

response = client.safety.run_shield

with:

fastapi.exceptions.RequestValidationError: [{'type': 'missing', 'loc': ('body', 'identifier'), 'msg': 'Field required', 'input': None}]

The preceding code that lists the available shields outputs:

{
    "identifier": "llama_guard",
    "params": {},
    "provider_id": "meta-reference",
    "type": "shield",
    "shield_type": "llama_guard"
}

llama-guard is running fine, llama-stack is running on localhost port 5000 and the safety examples is called with:

python safety.py 127.0.0.1 5000

What to do?

[ashwinb]

@tnsai what version of the llama stack server / llama stack client package are you running?

The error message indicates you are using a newer version of the stack-client than the server. We deprecated the identifier field and renamed it shield_id.

cc @dineshyv in any case

[tnsai]

@dineshyv I'm running pip installed version from last week, with some PR amends for remote::vllm committed a few days ago and the latest available pip installed client from yesterday.

When was identifier renamed (so that I can feed back to vllm provider contributors)?