SHIPP tool integration

[rtesra]

To do:

• [x] Add tests
• [x] Add function to update excel workbook with outputs from agyw_format_naomi()
• [x] Add scale to KP consensus estimates from KP workbook
• [x] Add scale to new infections estimates from KP workbook or Goals ASM database
• [x] Test workbook on multiple countries

[vimc-robot]

Thanks. Corresponding hintr PR at https://github.com/mrc-ide/hintr/pull/478

[r-ash]

On warnings in general, Ashton, Robert we should add a clearer error for if the workbook fails to generate. So far we have the following warnings/errors in place: Error: If area hierarchy if Naomi inputs does not match any of external data inputs. Flags which area_ids are inconsistent and tells user to contact support. Warning: (I'll add this in now) If KP consensus estimate is larger than 5% of matched population denominator. Tells user that default proportions will be used instead. Error: If sum of disaggregated infections does not equal total infections - tells user to contact support. This shouldn't happen now that we flag the mismatched area_ids at the start but left it for edge cases (who knows).

Errors sounds good, I will add some specific error handling into the excel file generation to return a nicer message should this fail.

Note about the warnings, these are not surfaced at the moment for downloads. So fine to add these here if you want but users won't see them until we add some specific handling for this in the front end. It will require a new endpoint in hintr to get these surfaced.

In general also after seeing the other warnings I am reluctant to use them. I think people have been trained to ignore the warnings.

That said, see the note above about a few floating calls to t_ which I think you want to wrap in a warning?