Refactor LBT07 layout to work with trim_levels_to_map

[anajens]

Redesign family of functions used to create LBT07 layout to work with rtables::trim_levels_to_map. Background issue (insightsengineering/rtables#203)

Rationale: improved speed, allow users to precisely control which combinations of levels among several categorical variables used in the layout should be displayed.

Note note all of the step above may be relevant for the layout. If an update is necessary, please add the PR link below.

• tern design doc
• investigate if design doc can be improved by using .spl_context as shown in nested .ungroup_stats in make_afun rtables#240 (comment)
• tern production code for s_, a_, layout function. Also please update relevant unit tests.
• teal.modules.clinical update. Note: Modules should aways split on the empirical map (observed combinations in the data) to make processing fast.
• TLG-C update: https://code.roche.com/nest/docs/tlg-catalog/devel/-/merge_requests/38
• nsdl.tests update: https://code.roche.com/nest/nsdl.tests/-/merge_requests/18

[anajens]

Moving to sprint 29

[anajens]

Not ready for October release so moving to SME backlog.

[imazubi]

@anajens here should we start the work from the beginning so that we can make use of .spl_context and hence split_rows_by(var = ANRIND) with a map dataframe that only accepts high/lowvalues but not normal?

[anajens]

@imazubi yes I think it would be nice to see if you can update the design doc to take advantage of .spl_context for this layout.

[anajens]

@imazubi just a heads up I added the nsdl.tests checkbox to the task list above.