check last contextual variable value for nans and raise error

[roussel-ryan]

This pull request improves the handling and validation of contextual variables in the Bayesian generator by ensuring that no NaN values are present in the latest row of training data for contextual variables. It also introduces a corresponding test to verify this behavior.

Validation improvements for contextual variables:

• Added a new method _validate_contextual_variables_no_nan to the BayesianGenerator class, which checks that the most recent row of training data does not contain NaN values in any contextual variable columns and raises a ValueError if this condition is violated.
• Integrated the validation step into the candidate generation process by calling _validate_contextual_variables_no_nan before model training in the generate method.

Testing enhancements:

• Added a test test_contextual_variable_nan_in_last_row_raises to ensure that a ValueError is raised if the latest row contains NaN in a contextual variable column during candidate generation.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!