feat: Test code suggested by Gemini

Asked Gemini (formerly Bard) LLM model to generate unittests for this:

https://github.com/redhat-performance/opl/blob/6cbe66432321394202ff0c8e26ee20fe08698ea1/opl/cluster_read.py

Related chats are here:

https://g.co/gemini/share/ee61acadbabe
https://g.co/gemini/share/f0a796512339

They are bit messy as looks like Gemini have issues with long prompt and
somehow truncates response (probably also when it is too long).

So later I have been giving smaller questions per class/function.

This is raw copy&paste of the code it generated and needs cleanup,
testing and edits.