ci: make Ollama testing workflow more robust

[anakin87]

Related Issues

• Ollama tests are failing often: https://github.com/deepset-ai/haystack-core-integrations/actions/runs/13147246857/job/36687949059
• this happens for several reasons, including network errors and subtle bugs on some Ollama versions

Proposed Changes:

After some attempts, I decide to make this testing workflow more structured and robust:

• separate Ollama installation and model pulling
• add a retry mechanism for these different steps: if they fail, it's not possible that python tests will pass
• (python tests already have a retry mechanism)

How did you test it?

CI
if you look at latest CI runs, you'll notice that the pulling step sometimes failed and has been retried

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test:.

[Amnah199]

From my understanding, it looks good.