make library_column_order a shared parent class as a move toward consistency

[gerrycampion]

made library_column_order a shared parent class for expected_variables, permissible_variables, required_variables, and get_dataset_filtered_variables. Achieved by adding optional filter params for library_column_order.
This is to reduce inconsistent handling of variable_metadata and wildcard replacement.

[github-actions]

Updated schema has not been merged with markdown descriptions. Please run the "Merge Schema with Markdown Descriptions" workflow to update the merged schema files.

[RamilCDISC]

Previously for here is core was missing the code would fall back to "PERMISSIBLE" but here it would become None. Is this behavior change intended?

[gerrycampion]

This is the reason for the updates in sdtm_utilities. Does that resolve your concern?

[RamilCDISC]

This might not be a big issue for our case but if any chance key-value be 0 or False then this comparison will not work as intended. Could the better approach be: if self.params.key_name is not None and self.params.key_value: is not None ?

If you think these values are not possible here we can skip the change.

[gerrycampion]

The values can be an empty string. I changed to only check truthiness of key_name. This way, if key_name is a non-empty string, we can do the filter. If key_value is missing, it will be treated as a filter for empty string.

[RamilCDISC]

The PR refactors the code to put shared code in parent class "library_column_order" to make code more consistent.
The PR was validated by:

1. Review the PR for any unwanted code or comments.
2. Review the PR in accordance with AC.
3. Ensure all unit and regression testing pass.
4. Ensure all relevant testing is updated.
5. Ensure CORE_test_suite pass as this validates no rules are affected
6. Manually run validation using dev editor.
7. Ensure all downstream files are updated to use parent class.