Arcanum.EnsureModel (arcanum v0.1.0)

Ensures a model is loaded with the correct configuration on local providers (LM Studio, Ollama, vLLM) before inference begins.

LM Studio: checks GET /api/v1/models for loaded_instances first,

       only calls POST /api/v1/models/load when the model
       isn't already loaded with the required context_length.

Ollama: POST /api/generate with keep_alive (or model is auto-loaded)

For cloud providers this is a no-op.

Summary

Ensures the model is loaded on the provider with the configured context length. Returns :ok or {:error, reason}.

Ensures the model is loaded on the provider with the configured context length. Returns :ok or {:error, reason}.

No-op for cloud providers or when context_length is nil.