Refactor jointBayesNet

[dellaert]

• Found we computed too much in jointBayesNet
• Refactored it with two helper methods
• exposed more DiscreteBayesTree methods in wrapper

[dellaert]

Thanks for the review. I'll incorporate the doc changes in a follow-up PR to avoid another CI.

[varunagrawal]

Thanks for the review. I'll incorporate the doc changes in a follow-up PR to avoid another CI.

I recommend doing it here so that it's all together if we ever have to come back to this PR. The cost of an extra CI run is not as bad as training LLMs these days. :)

[dellaert]

Thanks for the review. I'll incorporate the doc changes in a follow-up PR to avoid another CI.

I recommend doing it here so that it's all together if we ever have to come back to this PR. The cost of an extra CI run is not as bad as training LLMs these days. :)

I already did it in a follow up PR. Please approve so I can move on