GeezerVison

This commit introduces a dedicated Ollama provider to the application. Previously, local models were handled by a generic `local` provider. This change adds a new `ollama` provider, making the integration more explicit and easier to maintain.

The following changes are included:

- A new `ollama.go` file in `internal/llm/provider` to define the `OllamaClient`.
- A new `ollama.go` file in `internal/llm/models` to define Ollama models.
- Updates to `provider.go` and `models.go` to register the new provider and its models.
- A new test file `internal/llm/provider/ollama_test.go` to verify the integration.

This commit introduces several new features and improvements to the codebase.

- Reordered the provider priority to prioritize Ollama, OpenRouter, and Gemini.
- Added a new `config` command to the CLI that allows you to set your API keys and other settings.
- Added support for Hugging Face, Replicate, and Cohere as new LLM providers.
- Added a testing framework for the providers, including a mock provider.
- Improved the error handling in the application by creating a new `errors.go` file with custom error types.