Support local Ollama model created from Modelfile

[denverdino]

As a developer, I can build some local Ollama model for testing. But it cannot be used by Genai Stack directly.

Currently pull_model.Dockerfile will invoke ollama pull command to pull a model from a registry.

        (process/shell {:env {"OLLAMA_HOST" url} :out :inherit :err :inherit} (format "./bin/ollama pull %s" llm))

I think the proper way is to pull model if the model does not exist locally. The change should be simple, e.g.

        (process/shell {:env {"OLLAMA_HOST" url} :out :inherit :err :inherit} (format "bash -c './bin/ollama show %s --modelfile > /dev/null || ./bin/ollama pull %s'" llm llm))

[slimslenderslacks]

Cool, I think this will work for most people. It wouldn't pull new versions of the model as part of compose up but maybe that should be outside of compose up anyway.

[jexp]

@mchiang0610 what do you think? There is also a PR for this