Create data comparing aggregated Naomi and national Spectrum programme data

[rtesra]

To do:

• Change group in ANC data to females 15-49
• Update warning message to reference correct name of table /plot
• Add wrapper function to output both ART and ANC and only reads in pjnz and shape file once
• Add endpoint for warning to be displayed at upload inputs
• Fix tests

[r-ash]

This looks great, a few comments. Mainly about need to potentially run this with one of anc or art missing.

I also wonder if we should move this out of input-time-series.R that is getting quite long and this is data aggregation for a different plot.

[r-ash]

Could also do

spectrum_level <- as.integer(length(spectrum_region_code) > 1))

[rtesra]

Yup much better - thanks!

[r-ash]

I think we need to handle the case when either anc or art is missing? Users could upload 1 but not the other.

I'd change signature to

prepare_spectrum_naomi_comparison(shape, pjnz, art = NULL, anc = NULL)

Then have a check,

if (!is.null(art)) {
  if(!inherits(art, c("spec_tbl_df","tbl_df","tbl","data.frame" ))) {
    art <- read_art_number(art, all_columns = TRUE)
  }
  art_comparison <- prepare_art_spectrum_comparison(art, shape, pjnz)
}

and similar for ART. Then only need to bind if both present.

[r-ash]

Can you add a quick test for this too please? You can mock our prepare_art_spectrum_comparison and prepare_anc_spectrum_comparison to just test that it reads the data and that handles case above if ANC or ART are missing

[rtesra]

Have added both and to produce an empt table when user provides no programme data. Kept rbind in to be able to bind empty df.

[r-ash]

Possibly worth using dplyr::bind_rows here. I think it should generally be faster for large data frames (makes a big difference with a list of data frames) but seems to even with test data here

Browse[1]> bench::mark(rbind(art_comparison, anc_comparison), dplyr::bind_rows(art_comparison, anc_comparison))
# A tibble: 2 × 13
  expression           min median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result   memory     time       gc      
  <bch:expr>        <bch:> <bch:>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>   <list>     <list>     <list>  
1 rbind(art_compar… 68.9µs 89.9µs    10047.    4.31KB     4.32  4648     2      463ms <tibble> <Rprofmem> <bench_tm> <tibble>
2 dplyr::bind_rows… 47.5µs 63.3µs    13903.    4.01KB     4.74  5867     2      422ms <tibble> <Rprofmem> <bench_tm> <tibble>