Structured references

[kaitejohnson]

Goal
Other epinowcast packages have structured references, essentially grouping functions into families. We could consider doing this for this package.

Propose categories

Simple workflow: single reporting triangle, no "borrowing"

One layer more complex: estimate delay from a separate triangle than the ones used to estimate uncertainty and for nowcasting

All components separate

Generally follow the above descriptors, so:

• estimate delay
• generate retrospective data
• generate point nowcasts
• estimate observation error
• generate probabilistic nowcasts

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
#22 (review)

With example in primarycensored here

@seabbs Also added these to the analysis plan. Will make accompanying function flow diagrams. I couldn't figure out a nice way to show all 3 due to lack of imagination, but think we can combine into one figure easily with dashed/dotted lines or reformatting with inputs on the LHS