add model service check

[MichaelClifford]

This PR adds a check to the application to make sure the model service is running before proceeding. Once this PR is approved, we can roll out this same approach on the rest of the AI applications.

[jeffmaury]

Not sure this is related to this PR but the model service does not start:

[jeffmaury]

And seems the client application has opened its HTTP port:

[MichaelClifford]

Not sure this is related to this PR but the model service does not start:

This PR does not make any changes to the model service. However, it looks like you may have added the environment variables incorrectly. This model service is the same as the playground, so it should be deployed the same way.

podman run -it -p 8001:8001 -v <LOCAL_PATH>locallm/models:/locallm/models:Z -e MODEL_PATH=models/mistral-7b-instruct-v0.1.Q4_K_M.gguf -e HOST=0.0.0.0 -e PORT=8001 llamacppserver

[MichaelClifford]

And seems the client application has opened its HTTP port:

yes, but it should not load the chat interface until the model service is ready. Do we need to have a stricter check in place?

[MichaelClifford]

Added an additional commit that incudes the same check in chatbot/ai_applications/chat_ui.py.

[sallyom]

LGTM

[jeffmaury]

@MichaelClifford can you rebase it so that it includes my latest patch

[MichaelClifford]

@jeffmaury rebased 😄

[jeffmaury]

LGTM, but the client application opens port 8501 before the model is loaded and if you use browser, you have this:

which is ok but I was wondering is there would be an option to replace checking_model_service by a more user friendly sentence

[jeffmaury]

Another suggestion: would also be good to have the time spent to get the model service available at the end of loop

[MichaelClifford]

@jeffmaury How does this look?

[lstocchi]

Works fine for me

[jeffmaury]

LGTM

[lstocchi]

merging as it has been approved by everyone. Thanks @MichaelClifford!!