Add OllamaInterface

[standsleeping]

This branch adds support for Ollama via the Ollama Python Library.

I mostly followed the aimodels.ipynb as an implementation guide, but with a few adjustments:

1. Allow the user to specify the host URL for their Ollama server, and fall back to localhost:11434 (Ollama's default).
2. Catch ConnectError from the underlying httpx library used by the Ollama client, and raise a RuntimeError with a helpful error message that guides the user to start their Ollama server.

I also renamed the test_anthropic.ipynb notebook in the examples folder to a more general multi_fm_client.ipynb and added an Ollama example in a new cell beneath the Anthropic example. My assumption here is that we won't want to be adding a new notebook for each and every provider as we expand.