[completions and embeddings] Add support for Ollama (local LLMs)

[eolivelli]

Summary:

• add support for https://ollama.ai/
• with this patch you can now connect to local/custom LLMs via the Ollama REST API
• add a sample application that uses Ollama both for Embeddings and completions

Configuration for Ollama

resources:
    - type: "ollama-configuration"
      name: "ollama"
      configuration:
        url: "${secrets.ollama.url}"