Ollama container deploy - #57

- Add .azure/container-group.yml: ACI multi-container group YAML template
  with Bottomly and Ollama co-located (share localhost network namespace).
  Bottomly connects to Ollama at http://localhost:11434 with no public
  Ollama port exposed.
- Replace azure/aci-deploy@v1 with az container create --file in
  push_to_live.yml, using envsubst to inject secrets at deploy time.
- Enable LLM in production via EnableLlm=true and
  ConnectionStrings__bottomlymodel env var.
- Resource allocation: 0.5 CPU / 0.5 GB RAM per container (1 CPU / 1 GB total).

Requires OLLAMA_API_KEY secret in the live GitHub Actions environment.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>