Allow local providers to fetch model list

[khasinski]

What this does

This prevents model capabilities inaccurate warning from happening while keeping local model registry check in place.

# This outputs a warning, which can be silenced, but silencing it disables local model registry check
RubyLLM.chat(provider: :ollama, model: "llama3.2:latest").ask("hello?")
# W, [2025-07-23T21:22:09.067904 #50236]  WARN -- RubyLLM: Assuming model 'llama3.2:latest' exists for provider 'RubyLLM::Providers::Ollama'. Capabilities may not be accurately reflected.

# Fetching model list puts them into a local in-memory registry
RubyLLM.models.refresh!
RubyLLM.chat(provider: :ollama, model: "llama3.2:latest").ask("hello?")
# No warning

Basically an alternative to silencing log_assume_model_exists, which still validates the model's presence.

Includes model list fetching for Ollama. Since both GPUStack and Ollama can now fetch model lists running RubyLLM.models.refresh! is all that's needed to verify's model availability.

Type of change

• Bug fix
• New feature
• Breaking change
• Documentation
• Performance improvement

Scope check

• I read the Contributing Guide
• This aligns with RubyLLM's focus on LLM communication
• This isn't application-specific logic that belongs in user code
• This benefits most users, not just my specific use case (mostly for local model users)

Quality check

• I ran overcommit --install and all hooks pass
• I tested my changes thoroughly
• I updated documentation if needed (should I update Ollama and GPUStack manuals?)
• I didn't modify auto-generated files manually (models.json, aliases.json)

API changes

• Breaking change
• New public methods/classes
• Changed method signatures
• No API changes

Related issues