feat: Ollama Assistant Support

[MadMaximusJB]

It would be great to have local Ollama support to keep financial data on the system instead of using OpnenAI.

[jjmata]

See #17 ... being discussed there!

[jjmata]

FWIW, it's not as simple as "add ollama" since all that gives you is the ability to switch the model that current codepaths that use the LLM hit (all OpenAI/gpt-4.1* currently) and one would have to fine-tune the prompts and see how well they work against the models.

Definitely possible but involves a few steps/full thought given to it.

[jjmata]

Related: #116