Make function dependency graph

[fredjaya]

Fred to have a look at DependenciesGraphs.

This should help identify internal functions (non-user facing, helpers) and help with:

• avoiding doubling up on warning messages Throw warnings for some input variable ranges #3 (comment)
• avoid having to rename the input parameter names for internal/helper functions Standardise input variable names and local aliases #1

[AngusMcLure]

Looks like a useful tool and a good plan. The dependencies should be pretty straightforward currently I think, but this will be a useful tool to be across as the package gets more complicated (and for PoolTestR)

[fredjaya]

That was easy!

What are your thoughts on changing the input parameter names for the user-facing functions only? e.g. optimise_sN_prevalence, optimise_s_prevalence (any others?), and keep the parameter names for the internal functions as the greek notation?

[AngusMcLure]

This is a great tool, and the figure is really clear!

In terms of user-facing functions, I think design_effect_cluster_fisher should also be on the list.

However, though I can see the logic behind keeping the greek/math notation for the back end, I think I would prefer consistent notation throughout, with local aliasing as summarised in #1

[AngusMcLure]

Another point this highlights is that perhaps fi_pool and fi_ratio are superfluous as they aren't really called by any user-facing functions. In particular fi_ratio is kind of just another way of viewing design_effect_cluster_fisher. fi_pool is just a special case of fi_pool_imperfect. Could we perhaps get rid of the two superfluous functions and then rename fi_pool_imperfect to fi_pool?

[AngusMcLure]

We could also then rename fi_pool_imperfect_cluster to fi_pool_cluster. Makes the name a bit shorter...